F1-SP2e All

Special Issue Hondas Third Formula One Era
On December 5, 2008, Honda announced its withdrawal from Formula One racing. For me personally, as for all the
members of the Honda team directly involved in the Formula One projects and for everyone related to Honda racing,
this was a great disappointment and represented a very difficult decision. Honda was not isolated from the global
economic turmoil occurring at the time, and it was a period in which threats to the continued existence of the company
were developing rapidly.
It is no exaggeration to say that Formula One was a symbol of the Honda brand, and the companys withdrawal
from competition increased the sense of crisis throughout the Honda organization. This fact contributed significantly to
the rapid implementation of a variety of crisis measures, and now, 10 months later, Honda is recovering somewhat
more quickly than its competitors.
Hondas challenge during its third Formula One era was to compete and to win at the pinnacle of the worlds motor
races, and in doing so, to increase its technological capabilities, foster the development of its engineers, and boost its
brand value. Above all, in our third era we set ourselves the tremendous goal of focusing not merely on the engine, but
of achieving improvements in technologies throughout the vehicle.
While withdrawal from Formula One with the title within our grasp remains a great regret, I will be delighted if the
discussions of the technologies that Honda developed during its period of competition, published in this special issue
of the Technical Review, are read by a wide variety of specialists and prove useful in stimulating future advances in
related areas.
Having achieved incredible development as the leading industry of the 20th century, the automotive industry is now
confronted by a number of deep-rooted issues, including the reduction of CO2 emissions, the implementation of measures to deal with the depletion of fossil fuel resources, and the determination of ways to improve the balance of supply
and demand for resources on a global level. What is now demanded of Honda is for the company to demonstrate even
greater energy and passion than it poured into Formula One in taking up the challenge of these global issues and
carving out a new future for the automotive industry.
Takeo Fukui
Director and Advisor
Honda R&D Technical Review 2009
F1 Special (The Third Era Activities)
CONTENTS
Development Overview
Looking back on Hondas Third-Era Formula One Activities .................................................... Takeo KIUCHI ...................2
Transition of Regulation and Technology in Formula One ................................................ Hiromasa TANAKA ...................5
Summary of Honda Formula One Engine in Third-Era Activities ................................ Kazuo SAKURAHARA .................14
Overview of Gearbox Development for Formula One ............................................................... Atsushi MANO .................19
Summary of Honda Third-Era Formula One Chassis Development
......................................Akio TONOMURA, Shun KAWABE, Nobuhisa NAKAYOKU .................25
Development of Materials during Third Formula One Era ........................................................ Masami HOSHI .................34
Descriptions of Engine Technologies

Explanation of Hondas Third Era Formula One Engine Development
................................... Kiyoshi KONDO, Yasuhiro MOTOHASHI, Kazushi OGIYAMA
Hajime ENDO, Satoshi NAKAMURA, Nagao YANAGISAWA .................38
Technical Description of Formula One Engine Structural Design
.......................................................... Kiyoshi KONDO, Hajime ENDO, Tetsuo GOTOU .................44
Development of Reciprocating Parts and Crankshaft in Hondas Third Formula One Era
......................................................................... Tetsuo GOTOU, Nagao YANAGISAWA .................54
Development of Valvetrain for Formula One Engine
....................................... Shuichi HAYAKAWA, Kazushi OGIYAMA, Masanori TATE .................64
Development of Induction and Exhaust Systems for Third-Era Honda Formula One Engines
..................................................................... Ken NISHIMORI, Yasuhiro MOTOHASHI
Masayoshi TAKAHASHI, Ryuichi FURUKAWA .................72
CFD Technology for Formula One Engine
.................................... Naoki HANADA, Atsushi HIRAIDE, Masayoshi TAKAHASHI .................82
Combustion Diagnosis of Formula One Engine Using Micro-Cassegrain Sensor
................................................. Naoki HANADA, Atsushi HIRAIDE, Kohei YAMADA
Yuji IKEDA, Atsushi NISHIYAMA .................89
Development of Wind Simulator Equipment for Analysis of Intake Phenomena in Formula One Engines
....................... Satoshi NAKAMURA, Yasuhiro MOTOHASHI, Shuichi HAYAKAWA .................95
Measurement Technologies for Formula One Engines
................................ Nagao YANAGISAWA, Naoki HANADA, Takeru HAMAKAWA
Kazushi OGIYAMA, Tetsuo GOTO ............... 101
Development of High-Pressure Fuel Supply System for Formula One Engine
........................................ Tetsuya TANAHASHI, Kazuji ONO, Masanori HAYAFUNE
Yosuke SAWADA, Atsushi SHIMIZU ............... 109
Descriptions of Gearbox Technologies

Development of Seamless Shift for Formula One Car
.................................................................... Takashi YOSHIOKA, Takeshi UCHIYAMA
Katsumi KUBO, Ryo MATSUI ............... 120
Development of Honda Gears for Formula One Gearbox
................................................................................. Koichi KONISHI, Hiroaki NOJIMA
Teruoki NAGANUMA, Hiroshi KIMURA ............... 126
Development of Lightweight and Compact Differential for Formula One Car

.............................................................................. Toshio HIYOSHI, Yuichi SUENAGA
Haruki YOKOYAMA, Yoshikazu KATSUMASA ............... 131
Development of Clutch System for Formula One Vehicle
............................... Koichiro INUKAI, Yasuo NAKAGAWA, Hiroyuki KUSHIYAMA
Ryuhei KATAOKA, Yoshiki NAGATOSHI, Keiji SHIOTA ............... 136
Descriptions of Chassis Technologies

Development Methodologies for Formula One Aerodynamics
..................................................... Atsushi OGAWA, Shujiro YANO, Susumu MASHIO
Takashi TAKIGUCHI, Shinsuke NAKAMURA, Mitsuru SHINGAI ............... 142
Aerodynamics Analysis of Formula One Vehicles
........................................ Atsushi OGAWA, Susumu MASHIO, Daisuke NAKAMURA
Yasutaka MASUMITSU, Masayuki MINAGAWA, Yusuke NAKAI ............... 152
Technologies for Enhancement of Dynamic Performance of Formula One Vehicle
.............................................. Hideaki SHIBUE, Kazuhiro TANEDA, Yasutaka KITAKI ............... 163
Weight Reduction and Stiffness Enhancement Technology in Formula One Chassis Development
.................................................... Hajime WATANABE, Jun IWAI, Naoki HIROMASA
Terukazu EBISAWA, Kazuhiro TANEDA ............... 172
Measurement and Analysis Techniques for Formula One Chassis Development
................................................ Hideaki SHIBUE, Tomokazu SUZUKI, Mamoru URAKI
Akihiro IDA, Atsushi TSUBOUCHI, Yasutaka KITAKI ............... 179
Descriptions of Vehicle Control Technologies

Development of Traction Control Systems for Formula One
....................................................... Kazuharu KIDERA, Yoichiro FUKAO, Tatsuya ITO ............... 190
Development of Gearbox Control during Honda Formula One Third Era
.......................................................................... Takayuki KISHI, Yoshiki NAGATOSHI
Hiromasa NAKAMURA, Yoichiro FUKAO ............... 201
Development of Direct Push Clutch Control during Honda Formula One Third Era
.......................................................................... Takayuki KISHI, Yoshiki NAGATOSHI
Hiromasa NAKAMURA, Yoichiro FUKAO ............... 207
Descriptions of Electronic Equipment Technologies

Development of Electronic Control System for Formula One
............................................... Kenichiro ISHII, Toshiyuki NISHIDA, Kohei TOSHIMA
Masaki NEGORO, Masataka YOSHIDA, Yutaka MARUYAMA ............... 212
Development of Hybrid System for Formula One
........................................... Masataka YOSHIDA, Masato KITA, Hirofumi ATARASHI ............... 225
Descriptions of Material Technologies

High-performance Shell Bearing from New Material
......................................................... Kiyoshi ITO, Makoto ASAMI, Hiroshi KOINUMA ............... 240
Development of Hollow Crankshaft .............. Kiyonobu MIZOUE, Yasushi KAWAHITO, Ken MIZOGAWA ............... 243
Development of Metal Matrix Composite Piston

........................................ Kazuhisa ISHIZUKA, Katsuhito KUSAKAI, Nobuyuki IMAI ............... 246
Development of Lightweight Titanium-aluminide Piston Pin ................ Sadami MINATO, Yasunori ONAHA ............... 248
Development of Hollow Titanium Connecting Rod ............................... Sadami MINATO, Ken MIZOGAWA ............... 251
Development of Titanium Aluminide Valve
......................... Masayuki TSUCHIYA, Sakae TSUNASHIMA, Hisashi HASHIMOTO ............... 254
Development of Titanium Hollow Valve and Study of Sodium-potassium Valve .................. Taihei SANADA ............... 257
Development of DLC Coating on Camshaft and Rocker Arm
........................................... Naoaki ITO, Masaomi YONEHARA, Kazushige YAKUBO ............... 260
Development of Laser Clad Welded Valve Seat ........................................ Nobuyuki IMAI, Hiroaki NISHIDA ............... 263
Development of Titanium Exhaust Pipe
................................. Takeshi MUNEMURA, Hiroshi YAMADA, Takayuki OHNUMA ............... 266
Development of Lightweight and Low-friction Resin Materials for Oil Pump Rotor
.................................................................................. Yu MURAI, Masayuki TSUCHIYA ............... 269
Research on Coolant Using Nano Additives and Heavy Water ........................................................ Yu MURAI ............... 271
Development of High-performance Gear Materials
.................................. Daiki KONAGAYA, Tsutomu TANAHASHI, Takashi TANAKA ............... 273
Development of DLC for Transmission Gears .............................. Tsutomu TANAHASHI, Yasunori ONAHA ............... 276
Development of High Thermal Conductivity Material for Heatspreader
................................................................................. Yu MURAI, Masatoshi OKUMURA ............... 278
Development of High Induction Stator Core ....................................................................... Takehiro MIYOSHI ............... 280
Weight Reduction Study for Side Panel of Formula One Monocoque .............................. Yasuhiro YAMADA ............... 282
Development of Brake Caliper Production Process with High Strength Al-Li Material
............................................................................. Akihiro YANASE, Hiroshi YAMADA ............... 286
Descriptions of Race Management Technologies

Development of Lifecycle Management System for Racing Parts
.................................... Hidetada TANIGUCHI, Katsunori TAGUCHI, Masataka NODA
Shigeo MIYAJIMA, Osamu WATANABE ............... 290
Activity Reports
Looking Back on Assembly Activities Racing Spirit
...................................................................... Atsushi NORIHARA, Motoaki ISHIZAKA
Tsuyoshi ISHIHARA, Naoto SUNAKO ............... 300
Activities of Honda Racing Development
.................................... Yusuke HASEGAWA, Shiro HISATSUNE, Tatsuya KODAMA ............... 312
Concluding Remarks for Hondas Third-Era Formula One Activities ................................ Katsuhiko SUZUKI ............... 317
Race Records
Specification Summary for Hondas third-Era Formula One ................................................................................................320
2000-2008 Honda Formula One Racing Data Tables ...........................................................................................................330
List of Companies Participating in Hondas Third-Era Formula One Activities

List of Companies Participating in Hondas Third-Era Formula One Activities ..................................................................350
Comments from Project Members of Hondas Third-Era Formula One Activities

Comments from Project Members of Hondas Third-Era Formula One Activities ..............................................................354
< Prior Reports >

Fujita, H., Horiuchi, Y., Kubonoya, H., Baba, J., Sakai, K.: Development of Electric Power Steering System for F-1 Race Car,
Honda R&D Technical Review, Vol. 14, No. 2, p. 45-50
Shimasaki, Y., Sakaguchi, J., Maki, H., Suzuki, H., Kato, A., Nishizawa, K.: Study on Combustion Monitoring System for
Formula One Engines Using Ionic Current Measurement, Honda R&D Technical Review, Vol. 17, No. 1, p. 56-64
Development Overview
Looking back on Hondas Third-Era

Formula One Activities
Takeo KIUCHI*
In attempting to define the scope of the challenge that

we were setting ourselves with regard to Hondas thirdera Formula One activities, we began from the standpoint
of staking out a new area of challenge that would take
us beyond the achievements of our second-era Formula
One. We sought for a way of approaching racing that
would differ from our second era, when we had focused
on supplying engines. Our third-era activities got
underway with us assuming a more comprehensive
involvement not limited to engine supply, but also
incorporating the development of vehicle chassis
technologies and team management.
Viewed from this perspective, by contrast with our
second-era activities our challenges extended beyond
simply producing results in races. Basing ourselves in the
UK and importing our own particular technologies,
organizational management methods, and style of
activities into a European framework can also be
considered to be part of our new challenge, and
overcoming the many, many issues that this generated
might be said to be still another challenge that we set
ourselves.
The main point of difference with our second era,
however, was our full-fledged involvement in the area
of developing chassis technologies. While Formula One
vehicles are certainly cars, they are very different to
mass-production vehicles, and the development of openwheeled vehicles in which the major focus was on
aerodynamic performance presented us with new
technological challenges. In the area of total vehicle
packaging, in addition to the basic issue how to create
a stiff but lightweight framework consisting of a
monocoque, an engine, and a gearbox it was also
necessary to resolve other practical issues, namely how
to lower the center of gravity and reduce the moment
of inertia, and increase the degree of freedom of frontrear weight distribution. However, it is extremely
difficult to reconcile these demands at a high level, and
in reality development programs must balance them
while achieving yearly performance advances. In
proceeding in these areas in the initial stages, we
formulated numerical targets indicating our goals for
specific elements, but we could not predict how these
would affect lap times.
No doubt our rival teams, with their long histories
of development, understood how the changes made each
year would affect lap times, and established priorities in
moving ahead with the development of each element. A

new team can only obtain this type of know-how from
scratch, and we realized that this would be the area that
would take the longest time.
Team management was another area in which Honda
was inexperienced. In addition, the collaboration we
established with a new team formed by a commercial
production formula car constructor who had achieved
success primarily in the US, which had been the result
of our extended deliberations regarding the system that
Honda should use in competing, had, while representing
an ambitious endeavor, involved us in competitions with
an management style preferred by European teams that
had abundant Formula One experience and was hungry
for victory. This was also a major issue to be addressed.
During our second era of Formula One, we had often
been made aware of the differences between Japanese
and US and European management styles - specifically,
the difference between participation by every member of
the team and an individual-focused, top-down approach
that deployed specialists in each technological element.
However, we had managed the issue successfully by
completely separating areas of responsibility: the engines
were Hondas domain; and the vehicle chassis was the
territory of constructors such as McLaren.
Bringing in the new element of joint development
necessitated an entirely new and unprecedented
management style: an equal fusion of styles based on
different sets of values. There is a fundamental
difference between personnel fostered in the survival
strategies and cultures of companies working on contract,
and those fostered in a culture in which outcomes are
shared by everyone. This is perhaps a difference between
responsibility that is shared among the group, and
responsibility that is invested in the individual, and it is
not possible to state categorically which is superior.
When organizations of a similar scale come together on
an equal basis, these differences will cause friction to
arise at the points of contact between those
organizations. In retrospect, this fact, and the effects of
this friction, should have been more deeply considered
in advance.
It may appear that it would have been best for us to
simply proceed in our own way, but most companies, in
the US and Europe at least, operate on a contract basis,
and it was necessary for Honda, having for some time
been expanding and engaging in global business
* Automobile R&D Center

2
activities, to also engage with this issue in the area of

racing. However, we were unable to respond effectively,
and as a result significant burdens were placed on the
shoulders of those at the management levels of the
practical development before development commenced,
and it was not possible to organize a system that
functioned at maximum efficiency.
On the technological front, our belief was that if the
technology was excellent, it should be used regardless
of whether it emerged from UK or Japanese research.
However, for the UK side, if the technology was not
based on an idea emerging from their group, it could not
be applied. This was due to their contractual obligations,
and they believed that they were not satisfactorily
performing their duties if other ideas were used. Given
that the engines were supplied by Honda, the effects of
this were felt most strongly in the vehicle chassis
domain, where we had proceeded on the basis of joint
development, rather than in the area of engine
development. Frequent conflicts arose between the
Honda technical team, who were eager to engage in new
challenges to taking victory, and the personnel on the
UK side, who felt that the area of the vehicle chassis
was their responsibility and sought to proceed on the
basis of the knowledge they had attained through
experience. Numerous adjustments were made in roles
and modes of organization, but ultimately the only
solution was to allow a culture of mutual understanding
and respect to develop from a long involvement.
Control systems have advanced beyond conventional
engine control, and we are now in an era of control
systems that use vehicle parameters, such as traction
control. Here also, a variety of efforts were made to
achieve integrated engine and vehicle control, but some
years had to elapse before we could overcome
differences in concepts and measurement system
environments, and bring together the efforts of both
Japanese and UK sides.
We had gained experience of engine development in
our second Formula One era, but in the ten years that
had elapsed since Hondas withdrawal from factory
works racing, the gap between our projected performance
advancements and the technological level reached by the
top manufactures had grown significantly. As a result,
there was a considerable gulf in terms of performance
between our team and our competitors when we first reentered Formula One, and in particular the materials and
design methods that we used seemed to be from another
era in comparison with those employed by other teams.
Nevertheless, during our third era activities spanned
a period of ten years, including the period of preparation,
we overcame these issues and generated a significant
number of fruits.
In the area of the engine, after we were able to
analyze our competitors high level of performance from
a comparative perspective, within a short period we had
reduced engine weight by 30 kg at the same time as
boosting power by 50 kW. Our ideas and the results that
they produced, as exemplified by a unified-head
configuration, a single-piston-ring configuration, and IN
EX camshafts geared at the front and rear ends, were

generated by the coming together of the aspirations (the
vision) of our engineers and technologies that made full
use of recent 3D design technologies and precision
machining equipment.
Our most sensational technology, and one which
generated some astonishment in the industry, was a
seamless shift that resolved the issue of shift-lag. This
technological achievement had its beginning in the
creation of an upshift mechanism, and was completed
when we overcame issues in the area of downshift.
Aerodynamics-related development is, in certain
senses, an area of race technology with which massproduction manufacturers are quite unfamiliar. In the
development of standard mass-production vehicles,
styling appearance is a major priority in the exterior
design, and there is also, of course, a significant
difference in the speeds at which production vehicles and
Formula One vehicles are used. Because of this, we had
accumulated virtually no experience in the development
of vehicle forms to maximize Formula One aerodynamic
performance. This was the area in which it took us the
greatest amount of time to hone our skills, teasing out
the intricacies of Formula One aerodynamics from basic
theory until we reached the point at which we could
confidently produce new vehicle shapes. The
collaboration between engineers from Hondas
aeronautics and styling design departments in a certain
sense, rather dissimilar fields generated new designs
and new methods of analysis, leading finally to the
application of aero parts that would take us beyond the
top teams.
Aerodynamic performance was a high priority, and
even as we were constrained by the necessity of
generating results in this area, the achievement of weight
reductions and the development of chassis resulted in the
creation of numerous analytic and measurement
technologies, producing superior configurations and
superior performance.
The development of the KERS (a hybrid system),
scheduled to be introduced from 2009, succeeded in
producing a compact, high-performance system within
extremely restricted space and weight allowances in
comparison to a mass-production vehicle. Regrettably,
due to Hondas withdrawal from Formula One, the
system could not be used in racing. However, we were
able to test the system and demonstrate its effectiveness
before our competitors, and we were able to construct a
safe system before many of the other teams also
noteworthy accomplishments.
As a result of these efforts, we achieved 2nd place in
the Constructors Championship in 2004, and were
victorious in the Hungarian Grand Prix in 2006.
However, while the element technologies described
above individually produced results, as a team Honda
unfortunately did not achieve results that put us on par
with the top teams.
In the world of racing, victory will not be achieved
unless all of the many elements essential to competition
are functioning at their top levels. As I have indicated
Looking back on Hondas Third-Era Formula One Activities
above, we developed individual element technologies

that boosted performance when applied, but were unable
to coordinate the timing of these performance increases.
In an arena in which, no matter how superb the element
technology that is developed, competitors catch up
instantly when the technology is employed in the field,
what is necessary is to simultaneously develop and
continually evolve an array of outstanding technologies.
Our inability to supply technologies in this way, as well
as our failure to maintain the lead position of superior
element technologies until other elements could be
supplied, remain as significant issues for us. However,
the vehicle formed from the sum of these technologies
displayed excellent performance in the 2009 season for
the team that took over Honda racing team after its
withdrawal. I can well imagine how this makes the
members of the development team feel, and I hope that
they are confident that the goals they formulated for
themselves were the right ones, and that they had
succeeded in expressing those goals in the form of
technologies.
In the course of Hondas third Formula One era,
numerous staff members left Japan to base themselves
in the UK, separating themselves from the wellorganized development environment that had, in a certain
sense, been formed in our R&D centers. In particular,
many young engineers, relying on their own skills, went
head-to-head on an equal footing with engineers from
other countries, and in this process, their levels of skill
were recognized, and accordingly developed. A simple
comparison is impossible, of course, but I feel that the
burdens shouldered by the members of the development
team were greatest in our overseas activities during our
third Formula One era, when we failed to achieve our
desired outcome. But I think that for exactly that reason,
the growth that they experienced was also that much
greater.
The outcomes of the efforts of the team have now
been left behind, and it is my hope that all of our next
generation of engineers, and in particular those I have
discussed above, will mount new challenges in the near
future in which they will reclaim those outcomes.
In conclusion, I would like to applaud and to thank
all members of the Honda team involved in Honda
Formula One projects, their families, and the staff
members of cooperating makers for their tremendous
efforts and their support over an extended period. Thank
you.
Author
Takeo KIUCHI
Transition of Regulation and Technology in

Formula One
Hiromasa TANAKA*
ABSTRACT
The Formula One regulations are established with consideration of fairness, competition, safety, sustainability, and
entertainment value.
In order to ensure free technological competition, the technical rules of Formula One, under which constructors
are obliged to compete using original vehicles, are essentially free, other than specifying minimum basic items.
From the beginning of the 1990s to 2009, despite continuous changes in the regulations to increase vehicle
performance control and safety, lap times improved almost every year due to the development of technologies that
exceeded regulation stipulations. As a result of these battles, the letter count of the Formula One technical regulations
has increased more than three-fold over 19 years.
Detailed technological development spanning a broad range of elements and involving the use of computers, in
addition to a comprehensive management approach that brings these elements together, are essential to the Formula
One of the 2000s. In recent years in particular, Formula One has responded to rapid social change, for example with
measures to cut costs in line with worsening global economic conditions, and the development of environmental
technologies, as exemplified by responses to global warming, in addition to measures to highlight competition the
essence of Formula One racing and to increase the spectacular nature of the races. This is a period in which Formula
One is reflecting upon the very meaning of its existence.
1. Introduction
(1) First era (1964-1968): Raced under the Honda name as

a full works team, including engine construction, chassis
construction, and team management
(2) Second era (1983-1992): Supplied works engines as an
engine constructor to chassis constructors including
Spirit, Williams, Lotus, McLaren, and Tyrell
(3) Third era (2000-2008):
2000-2005: Supplied works engines to and conducted
joint chassis development with the chassis constructor
British American Racing (BAR); raced as BAR Honda
2006-2008: Raced as Honda Racing F1 Team (HRF1)
as a full works team, including engine construction,
chassis construction, and team management
In parallel with these activities, Honda also supplied
engines in 2001 and 2002 to Jordan Grand Prix, and
from 2006 to 2008 to SUPER AGURI F1 TEAM
(SAF1).
For almost the entire period between Hondas second
and third Formula One eras, from 1994-2000, the
company supported technical development efforts by
MUGEN (now M-TEC), a company that supplied
engines as an engine constructor, and was involved in
Formula One as part of MUGEN-Honda.
This paper will first discuss the characteristics of the
Teams entered in the Formula One World

Championships are basically required to race their own
original vehicles, and are termed constructors. Formula
One racing, which commenced in 1950, encompasses
both a drivers championship, in which drivers compete
to determine who is the worlds fastest, and since 1958
a constructors championship, in which constructors
compete to determine which of their vehicles is the
worlds fastest. However, the production of engines
presents a considerable challenge to any participant other
than an automaker or a specialized engine manufacturer
that possess specialized technologies. In the case of
engines, Formula One regulations permit acquisition or
purchase from an external engine constructor. Therefore,
in the naming of teams, the name of the chassis comes
before the name of the engine, and the constructors title
is awarded only to the chassis.
In the racing world, the term works refers to the
fact that a maker itself manufactures the racing cars,
racing engines, and the like, or manages the team.
Divided into first, second, and third eras, Hondas
Formula One activities were organized as follows:
5
Transition of Regulation and Technology in Formula One
Formula One regulations, and will then go on to consider

the main changes in Formula One regulations and
technologies employed from the beginning of the 1990s,
the last phase of Hondas second Formula One era, to
2008, the last year of the companys third Formula One
era. It will also look at the changes in Formula One
regulations and technologies up to 2009, for which
Honda engaged in vehicle development, but ultimately
did not compete due to the companys withdrawal from
Formula One.
a slight deviation from the narrow scope they define

enables developers to attain a freedom in their
development just because it is not subject to the
regulations. As a result of teams pursuing development
of this type that attempts to sidestep the regulations, and
of the continuing formulation of increasingly detailed
regulations in an attempt to control it, the Formula One
technical regulations increased in volume from
approximately 40000 letters covering 65 items in 16
chapters in 1990 to approximately 86000 letters covering
140 items in 21 chapters in 2000, the first year of
Hondas third Formula One era. By 2008, the final year
of the companys third Formula One era, this had risen
to approximately 122000 letters covering 157 items in
21 chapters.
A description of the detailed mechanism of
formulation and amendment of Formula One regulations
can be found in another paper(1), but in terms of Formula
One history, when a regulation has been clearly
deficient, the principle has been that the relevant item
will only become prohibited following the publication of
a formal codicil and the completion of the amendment
procedures, and until that time teams are able to use it
to their advantage. However, the right to specify
interpretations within a scope in which no modification
of the text of the regulations is necessary belongs to the
Federation Internationale de lAutomobile (FIA) at any
time, and the organization issues technical instructions
called Technical Directives (TD) to all Formula One
teams as necessary. Other than this, the FIA also has the
authority to permit the Stewards of the Meeting to
exclude vehicles that are considered dangerous from race
events. Objections to a judgment at a race event may be
presented before the International Court of Appeal
(I.C.A.), the final court of appeal for motor sports.
2. Characteristics of Formula One

Regulations
Formula One regulations are based on the FIA
International Sporting Code, which provides shared rules
for all motor sports, and incorporate sporting regulations
and technical regulations unique to Formula One
championship(1). The sections below will discuss the
characteristics of these regulations, with a focus on
technical regulations that directly affect technical
development for Formula One.
2.1. Basic Principles and Structural Conflicts
Formula One involves competition between vehicles
built from scratch by constructors. To help ensure
competition, absolute prohibitions and mandatory
observances have been minimized in the technical
regulations, and constructors are free to approach any
point not explicitly covered in the regulations as they
choose. For example, there are no rules concerning the
wheel base, and constructors may employ a wheel base
of any length. One can associate the Formula in
Formula One with the minimum standards in the
regulations.
This basic concept is actually the antithesis of the
concept of production car racing. In production car
racing, a homologated mass production vehicle that has
satisfied specific requirements regarding the number of
units produced is used as a base, and only modifications
that are specified in the regulations can be made to the
vehicle.
Formula One racing is a contest of speed, and the
performance of the vehicles therefore tends to improve
year by year. To help ensure that the performance of the
vehicles does not exceed the limits of the capacity of the
circuit safety facilities such as run-off areas and guard
rails, the rules of technical regulations are continuously
enhanced in order to increase the safety of the vehicles
themselves and to impose restrictions on vehicle
performance. However, in order to take victory in a
fierce competition, each team pursues technical
development to the very limits of the regulations. Clear
standards and accurate methods of judgment are
therefore necessary for regulations designed to help
enable fair and consistent judgment of the outcomes of
this competition.
To this end, the regulations themselves should be
finely subdivided and highly detailed. However, once
highly detailed regulations have been formulated, even
2.2. Example of a Conflicting Interpretation

The 2005 regulations specified the following points:
First, the regulations defined vehicle weight as the
weight of the vehicle including the weight of the driver
in full racing apparel, but did not indicate whether or not
the fuel was included in this figure. The regulations
further specified that the weight of the vehicle was to
be no less than 605 kg during the qualifying practice
session, and no less than 600 kg at all other times during
a race event. Ballasts could be employed to reach the
minimum weight, but the regulations specified that a
ballast should be secured in a manner that necessitated
the use of tools for its removal, and that it should be
possible to affix seals if deemed necessary. Further,
during the race, no substance may be added to the
vehicle other than fuel and compressed gases.
Since 1994, it has been customary in Formula One
to verify the weight of the vehicle with the fuel removed
in the car inspection following the completion of the
race, in order to help stop teams from attaining an
advantage by running under the minimum weight during
the race and bringing the vehicle up to the minimum
weight after the race by adding fuel in the closing stages
of the race. However, there was insufficient consensus
between the FIA and the teams regarding the level to

which fuel should be removed. In actual fact, to
completely remove all fuel from the fuel system would
necessitate dismantling part of the engine, which was
unrealistic on the circuit. The teams therefore considered
that under normal circumstances it was acceptable to
include the remaining unusable fuel in the vehicle
weight.
However, an issue arose concerning two BAR Honda
vehicles that took 3rd and 5th place in the San Marino
Grand Prix, the 4 th round of the 2005 season. At the
time, BAR Honda vehicles employed a sub fuel tank to
help maintain a stable fuel supply pressure. The
configuration was such that the fuel in the sub fuel tank
could not be easily removed using ordinary methods. It
was reported that when one of the vehicles was
reweighed during the car inspection following the race
after all fuel from the sub fuel tank and all remaining
fuel from the main fuel tank had been drained, it was
found to be 5.4 kg under the regulation minimum weight
of 600 kg. Based on the report of the outcome of the
car inspection, the Stewards of the Meeting interviewed
the people concerned, finally upholding the original race
result. However, the FIA brought an objection to the
I.C.A.
The I.C.A. considered that it could not be physically
proven that the vehicle had not been run during the race
below the 600 kg minimum vehicle weight, and that
insufficient evidence was provided by the combination
of data and theory. In addition, it upheld the
interpretation that because the use of fuel as ballast did
not correspond to the rules concerning ballast, it could
not be allowed. However, if this interpretation of ballast
was followed, then oils and cooling water would also not
be included in the minimum weight of the vehicle. In
actual fact, however, vehicles are weighed following
races containing oils and cooling water. Because addition
of oils or cooling water during the race is not allowed,
they cannot be used to bring the vehicle up to the
minimum weight, and they are therefore customarily
considered to be included in the vehicle weight.
The final verdict of the I.C.A. was that while it could
not be proven that the team had deliberately set out to
commit fraud, the teams actions at the time that the fuel
tank was drained for inspection, and their failure to
request a clarification of the interpretation of the relevant
regulation in advance, demonstrated negligence on their
part. The two BAR Honda vehicles were disqualified
from the San Marino event and the team was suspended
from the next two events in the championship. The team
was also suspended for a further period of six months,
but this penalty was suspended for one year.
In Formula One considered as a sport, whether or not
an inspection is conducted, competitors are regarded as
having sworn that their vehicles are in compliance with
regulations when they participate in a race. The onus of
proving this rests on the teams. The authority to interpret
the regulations is stipulated as resting with the FIA, the
entity that is authorized to conduct the competition, and
this authority should be respected. The author therefore
has no intention of arguing with the I.C.A. ruling.

However, a clause was added to Section 6.6. of the
technical regulations in 2008 stipulating that all
competitors should provide a means of removing all fuel
from the vehicle, and this can be seen as a means of
more certainly heading off any disagreement over the
interpretation of the regulation.
3. Changes in Formula One Regulations

This section will discuss changes in Formula One
regulations and technologies (Table 1), while providing
some examples of development within each area of
technology. For details of specific technologies, please
refer to the relevant paper in this special issue of the
Technical Review.
3.1. Engine and Powerplant
Regulations concerning the engine have been
continuously revised in order to limit the performance
of Formula One vehicles. In 1995, the previous
maximum displacement of 3500 cm 3 was reduced to
3000 cm3, and this figure was further reduced to 2400
cm 3 from 2006. Until 1999, the maximum number of
cylinders that could be employed was twelve. However,
as a result of a technical trend that valued the overall
balance of the vehicle over the performance of engine
unit teams gradually moved towards the use of ten
cylinders, and by the time the regulations specified ten
as the maximum number of cylinders in 2000, all
Formula One engines already employed this number of
cylinders. In 2006, together with the reduction in
maximum engine displacement, the maximum number of
cylinders was reduced to eight.
While it cannot necessarily be considered ideal from
the perspective of seeking increased engine power, from
the mid-1990s teams proceeded to increase the
temperature of the cooling water in their engines, in
order to reduce the aerodynamic drag generated at the
openings for the cooling system. In order to put a stop
to this trend, from 1999 the regulations specified a
maximum cooling water pressure of 3.75 bar.
Maximum engine speed reached about 20000 rpm in
2006. While it had previously not been treated in the
regulations, from 2007 maximum engine speed was
limited to 19000 rpm, and from 2009 to no more than
18000 rpm. To help enable the vehicles to be run faster
at a restricted engine speed, there was a trend towards
an increased frequency of use of the high engine speed
range, and this made developments aimed at increasing
durability even more necessary.
In order to reduce costs by limiting changes to engine
specifications and the number of engines used in a year,
a (long mileage) regulation was introduced stipulating a
minimum use distance for the engine. Previously,
constant engine changes had been allowed during a
three-day race event, but in 2004, regulations obliged
teams to use one engine per race event. The following
year, more stringent conditions were applied, with
regulations stipulating that one engine should be used per

Table 1
1990
1991
Supercharging
Capacity
Number of cylinders
3500
1992
cm3
1993
1994
Transition of F1 regulations
1995
1996
1997
maximum
Free
Hybrid systems
Free
KERS
permitted
Prohibited
Commercial fuel or fuel in development
for future commercial use
Type used by general public
Free
Prohibited
7 maximum
Free
Prohibited
Twin clutch prohibited
Free
Free
Mileage
Free
Rear wing
maximum width
Rear maximum
600 mm
overhang
Rear center diffuser
maximum width
Limitations on
fundamental
dimensions
and materials
1 gear box for 4
race events
(Excluding FP1
& FP2)
595 kg
600 kg (with driver)
(with
driver)
Free
2000 mm maximum
505 kg (without driver)
2150 mm maximum
Flat bottom
Commercial fuel
5.75% (m/m) biofuel
Commercial fuel
Sampling at circuit
Approval before use
Free
Minimum: 4; maximum: 7
Prohibited
1000 mm above
ground
Homologated
3.75 bar maximum
Clutch
605 kg for qualifying

(with driver)
600 kg for others
(with driver)
Prohibited
Moving ballast prohibited and area of skid block fasteners limited
1800 mm maximum
Stepped bottom
950 mm
above ground
950 mm
above
reference
plane
800 mm above reference plane
1000 mm
750 mm
500 mm
600 mm
1000 mm
300 mm
Rear diffuser
maximum height
1000 mm
125 mm
maximum 175 mm
(without
center maximum
area)
Free
Bodywork
Front wing
and
maximum overhang 1200 mm 1000 mm
dimensions
Front wing
maximum width 1500 mm
Up to height of front wheel rims (end plates)
Front wing height
2400 cm maximum
8
19000 rpm 18000 rpm

Variable geometry prohibited
Specification
limitations
Rear wing
maximum height
Free
2009
1 engine for 1 engine for 1 engine for 2 race events 8 engines

1 race event 2 race events (Excluding FP1 & FP2) for 1 season
500 kg
(without
driver)
2008
Free
Specific
gravity
controlled
Sampling
2007
Limitations
on
fundamental
dimensions,
weight,
and
materials
Prohibited
Automatic gear changing

Forward gear ratios
Continuously variable
transmission
EPS
Ballast
Overall width
Facing ground
2006
Free
Free
Fuel
Chassis
2005
Free
Rev limit
Inlet systems
Cooling system
water pressure
Machine weight
2004
10
Mileage
Properties
Minimum
weight
of car
2003
Free
Refueling during race
Gear box
2002
12 maximum
Homologation or
limitation of
specifications
Engine
and
powerplant
1998 1999 2000 2001

Prohibited from 1989
3000 cm3 maximum
900 mm
1000 mm
1400 mm
1800 mm
Between
75 mm
At least
At least 25 mm above 40 mm
above
reference plane
reference
plane
Between 100 mm and 300 mm Between 150 mm and 350 mm and

275 mm
above reference plane
above
Between 50 mm and 250 mm

reference
plane
Side wing
prohibited
No body Height
work reductions
just
area
around front
front of rear
wheel wheel
Enforcement of
other bodywork
limitation and
deflection test
Tire width
Front
wing test
(downward)
Rear
wing
assembly
test
(rearward)
Tires
Tire supplier
Electronically
controlled
devices
Traction control
Launch control
Chassis devices
Materials
Safety
Prohibited
Software validation in advance
Prohibited
Software validation in advance
Free
Free
Suspension tether
Grooved
Front-4
Rear-4
Slick
Single supplier
Prohibited
(Single ECU)
Free
Prohibited
(FIA logger observation)
Free
Prohibited
40 GPa/
(g/cm3) limit
for engine
moving
parts
Specific modulus
Material designation
Frontal crash test
speed
Side impact test
Side panel
penetration test
Designated
front
center
wing
Front 305-355 mm
Rear 365-380 mm
Single supplier
Free
Rear
wing
element
bridge
mandated
test
test
(rearward) (downward)
Grooved
Front-3
Rear-4
Slick
Tire tread
Rear
wing
Splitter wing
top
front
test
(upward) element element
Front 305380 mm
Rear 365380 mm
15 inch maximum
18 inch maximum
Rear
Test just
front of
rear
wheel
(downward)
40 GPa/(g/cm3) limit for all parts
11 m/s
12 m/s
Designated for engine parts

13 m/s
14 m/s
15 m/s
Mandated
Specific
designation
Single tether
Mandated
Double tether
3.2. Gearbox
A semi-automatic gearbox based on a sequential shift
mechanism developed in the early 1990s, by means of
which gears are shifted using a paddle attached to the
steering wheel, has been continuously used as the
standard Formula One gearbox type. As will be
discussed below, from mid-2001, a significant relaxation
in the regulations concerning electronically controlled
systems temporarily enabled the use of automatic shifting
systems that were not operated by the driver, but
regulations were tightened to once again prohibit these
systems in 2004 (as part of driver aids ban) over
concerns that they would obscure differences in levels
of skill between the drivers.
With regard to the number of forward gear ratio,
there has basically been no change since the regulations
stipulated a maximum of seven in 1994, but a prohibition
on the use of continuously variable transmissions was
added from the fifth round of the 2001 season. Following
this, in 2002, regulations prohibited dual-clutch
transmissions, which enable uninterrupted switching to
the next pre-engaged gear stage. It was considered that
a de facto prohibition existed on shift mechanisms that
produced no discontinuity in torque, the ideal form of
gearbox, but in 2005, BAR-Honda introduced a seamless
shift using a one-way mechanism that the team had
developed based on a close study of the regulations.
Other teams sought to keep pace, and this has now
become the standard technology in use in Formula One.
Since the 1980s, every team has employed original
designs and methods of manufacturing gearbox case, which
is a structural element of the chassis to which the
suspension is mounted. In the first half of the 1990s, most
gearbox cases were cast from magnesium, but after this a
variety of methods came into use in attempts to produce
lightweight, high-rigidity cases, including the manufacture
of the cases from thin-walled titanium castings and welded
plates, and the use of CFRP reinforcing. In 2004, BARHonda became the first team to produce a practicallyapplicable all-CFRP gearbox case.
With regard to the gears and shafts inside the
gearbox, even in the early 2000s it was still standard
practice, with some exceptions, to employ existing parts
manufactured by gearbox makers. However, teams began
to develop their own original gears as development
competition seeking weight savings and greater
compactness intensified. Based on concerns over
spiraling costs, from 2008 the regulations have stipulated
that forward gears should be manufactured from iron,
and have also specified minimum gear width, gear
weight, distance between shaft centers, and the like.
In 2008, the long-mileage concept was also applied
to gearboxes, and a single gearbox is now required to
be used for four race events. However, the adjustment
of the gear ratios to match the characteristics of specific
circuits and the replacement of the dog rings is allowed
conditionally.
two race events (from 2007, the Friday free practice (FP)
session was excluded from the regulation). In 2009, the
total number of engines that a team could use per year
was limited to eight.
From 2006, in order to avoid cost increases resulting
from excessive development competition, common basic
dimensions for engines were established, and minimum
weight and center of gravity of an engine were specified.
Then, from the next year, 2007, a system of engine
authorization (homologation), under which major engine
specifications could not be modified once they had been
registered, was introduced one year earlier than initially
scheduled. At the same time, the use of variable
geometry intake manifolds, previously allowed, was
prohibited. The introduction of this homologation system
represented a significant change in orientation for the
Formula One philosophy, which as indicated above is
based on free technical competition. Broad-ranging
development efforts, seeking even minor enhancements
of basic performance, were concentrated in the period
immediately before the submission deadline for engine
homologation applications, but following this, engine
development was largely limited to subtle modifications
within the allowable scope or tuning of characteristics.
In order to more thoroughly embody the way of
thinking that sources of drive power other than engines
should not be basically allowed, and to prohibit the
hybrid systems that some teams were said to be using,
in 1999, power sources other than 3000 cm3 engines
were prohibited, and the maximum recoverable stored
energy was restricted to 300 kJ (the maximum stored
energy recoverable at a rate greater than 2 kW was
restricted to 20 kJ). From 2009, however, the use of a
type of hybrid system called a Kinetic Energy Recovery
System (KERS) has been allowed with the intention of
promoting the development of environmental
technologies through the medium of motor sports. These
are the only systems excluded from the regulations
discussed above. The maximum energy that can be input
to or output from a KERS is 60 kW, and a maximum
of 400 kJ can be recovered per lap.
Refueling during a race was prohibited for some
time, but this prohibition was lifted from 1994, focusing
attention on pit stop timing and the amount of refueling
as aspects of race strategy. In line with this, technologies
associated with aircraft refueling systems were applied,
and from 1995 regulations stipulated the use of identical
refueling systems able to safely and rapidly supply at a
fixed refueling speed.
With regard to the composition of the fuel employed,
while the interpretation that Formula One fuels should
fundamentally be based on commercial fuels was upheld
in the 11 th round of the 1992 season(2), from 1999 to
2002, detailed rules concerning the ratio of various
hydrocarbons in the fuel made it possible to use fuels
that had been developed for the purpose of enhancing
commercial fuels in the future. In addition, as a response
to worldwide efforts to ameliorate global warming, from
2008, the regulations have stipulated the use of fuels
containing a 5.75% ratio of biofuels.
3.3. Chassis
In line with advances in electronic chassis control
technologies in production vehicles, at the beginning of

the 1990s the application of a variety of technologies
was tested, including active suspension, traction control,
four-wheel steering, and anti-lock brakes. Semi-active
suspension had a particularly significant effect, enabling
the realization of ride heights control that produced
stable aerodynamic performance. However, based on
considerations of performance control, and to control
cost increases and ban driver aids, electronic chassis
control systems were prohibited in 1994.
Power steering had been used in Formula One since
the early 1990s, in order to respond to the increased
steering force resulting from advances in aerodynamics
and increased tire grip. From 1994, the use of this
system was limited to the provision of assistance to the
drivers physical effort. Initially, electronic power
steering (EPS) systems were allowed if they satisfied this
condition, and from the 12th round of the 2000 season
through 2001, Honda led other teams in using a Formula
One EPS system in racing(3). However, because of the
challenge represented by the verification of and judgment
on the details of the control, making it possible that the
systems incorporated control that functioned as a driver
aid, the use of EPS was entirely prohibited from 2002,
and hydraulic power steering systems are now used
exclusively in Formula One.
The minimum weight of the vehicle was initially
specified in regulations without the driver, and was 500
kg in 1990, increasing to 505 kg in 1993. From 1995
the driver was included in rules regarding minimum
weight. In 1995 the figure was 595 kg, increasing to 600
kg in 1996. In 2004 the minimum weight was increased
to 605 kg for the qualifying practice only, and in 2007
this was made the minimum weight for the entire race
event. Compared to the previous circumstances, the
minimum weight figures have not changed significantly
in themselves.
In 1993, the maximum width of the vehicle, including
the tires, was reduced from the previous 2150 mm to
2000 mm. The figure was further reduced to 1800 mm
in 1998. In the case of formula cars, this directly means
a reduction in track, and this regulation was applied in
order to reduce the cornering limit speed of the vehicles
by increasing load transfer on the left and right tires.
While reducing the weight of their vehicles, teams began
to make efforts to lower the center of gravity by even a
small amount and thus increase the cornering limit speed,
positioning corresponding ballasts, manufactured from a
tungsten alloy with a high specific gravity, as low as
possible in the vehicle. As a result of these efforts, the
top teams came to employ ballasts weighing 70 kg or
more. This enabled the front-rear weight distribution to
be maintained towards the front of the vehicle while
helping to ensure the dimensions of the front half of the
vehicle (the distance between the front wheels and the
side pontoons) necessary from an aerodynamic
perspective, which was also effective in increasing
warming of the tires and boosting the dynamic
performance of the vehicle. In 2000, from considerations
of safety, the regulations banned the use of movable
ballasts and limited the area of the skid block fastener,

which is used as the lowest ballast on the vehicle.
In an attempt to significantly lower downforce, which
had previously been continuously increasing, and thus to
reduce the cornering limit speed, in 1995, flat bottom
rule was replaced by a stepped bottom rule that would
increase the distance between the bodywork facing the
ground and the road surface, and the use of skid plates
that would increase the ground clearance was also
stipulated. However, a high nose configuration that
increased downforce by encouraging air flow beneath the
vehicle had been developed in 1990, and this
configuration was applied by all teams in 1996. Efforts
were also made to increase the contribution of the shape
of the front suspension to vehicle aerodynamics. A twin
keel mount was introduced in 2002 and a zero keel
configuration was introduced in 2005. Nevertheless,
since the front suspension stroke is extremely small in
Formula One vehicles, the tendency has been to
emphasize rigidity over geometry.
In parallel with this, from about 1991, the regulations
concerning the dimensions of the front and rear wings
and the diffusers positioned in the rear of the bodywork
facing the ground, which were making an increased
contribution to aerodynamic performance, began to be
reviewed. Since 2000, the regulations have been made
more stringent on an almost annual basis. In addition,
given various advancements in aerodynamic components
such as vortex generators, bargeboards, and winglets,
designed to increase the overall downforce of the vehicle
by adjusting the flow of air around different parts of the
body, a large number of additions have been made to
regulations concerning the dimensions of parts of the
vehicle body.
In addition, as a result of the use of flexible wings
to reduce aerodynamic drag at high speed, since 1999
regulations have stipulated measurement of the level of
deformation of each part of the vehicle body and the
fitting of stiffening elements, among other requirements.
The previously used slick tires, which had no
grooves, were replaced by grooved tires from 1998, in
order to reduce the absolute level of grip. In 1999, one
extra groove was added at the front of the tires for a total
of four front grooves and four rear grooves. This
configuration was employed until 2008. The stiffness of
the tread rubber of grooved tires is low; they are prone
to delay in response and present issues of controllability.
In addition, they are also challenging to handle, with the
blocks between the grooves being deformed to incline
and abrasion termed graining tending to occur
commencing from the edge. Appropriate temperature
management of the tires is necessary for the effective use
of these grooved tires. At the same time, it is also
necessary to increase controllability, maintaining the
overall vehicle balance and a minute transient response
characteristic. In 2009, there was an extensive review of
aerodynamic issues, and aerodynamic regulations were
introduced based on the concept of lowering cornering
performance and reducing the effect of the wake of the
vehicle in front, in order to enable active overtaking. At
10
the same time, high-controllability slick tires were

reintroduced.
by developing surface treatments and conducting

research at the molecular level, in order to realize
desirable materials that are able to achieve high
performance levels while remaining compliant with
regulation rules concerning composition and physical
values.
Since 2006, detailed regulations have been in effect
concerning the types of materials that can be used for
each engine part, and development has continued within
the scope defined by the regulations.
3.4. Control
It is the fate of control systems in Formula One to
be constantly subject to regulations from the perspective
of driver aids ban. It is more challenging to regulate
engine-based torque control systems such as traction
control than chassis-based control systems, in the case
of which it is comparatively simple to make a judgment
based on whether a device is fitted or not. Since 1997,
a software validation in advance has been mandatory, but
even if the FIA inspections were to confine themselves
to a hotel room with the team members for several days
for each team in order to check source code, it would
still not be possible to achieve complete restriction.
Perhaps because regulations relating to this area had
reached their limit of operability, from the 5th round of
the 2001 season, regulations were abolished for all forms
of software excluding safety-related software, and engine
torque control became effectively free.
This did not only affect traction control. Development
has also proceeded in other areas, including engine brake
control and over-run control, which achieves a type of
ABS effect by using engine torque during braking to
help keep the rear wheel brakes from locking.
Launch control, which increases acceleration by
making maximal use of the tire grip at race start, which
changes moment by moment, also advanced
considerably. However, from 2004, launch control was
prohibited through detailed regulation of the method of
using the engine and the clutch, and the introduction of
a method of using standardized FIA data loggers which
should be fitted to vehicle to monitor the way of using
them.
Following this, from 2008, torque control, including
traction control, was once again prohibited because the
requirement to fit a standardized FIA ECU to vehicles
made it possible for the first time to accurately manage
the regulation.
3.6. Safety
The deaths of Roland Ratzenberger and Ayrton Senna
in accidents in 1994 resulted in a large-scale revision of
the regulations seeking a rapid enhancement in passive
safety. Continuous changes to the regulations have also
been made since then, and as a result no Formula One
driver has been killed in an accident in the intervening
15-year period. More details can be found in two
previously published papers(1),(4).
In the opening round of the 2001 season, the BARHonda vehicle being driven by Jacques Villeneuve ran
head-on into another vehicle and became airborne before
suffering a severe crash. Protected by a robust carbon
monocoque, Villeneuve escaped with almost no injuries.
Tragically, however, a wheel torn off the vehicle struck
and killed a track official. Since 1999, regulations had
already stipulated the use of suspension tethers between
the uprights and the chassis in order to help reduce this
type of accident, and standards have been raised each
year. However, given the fact that the achievement of
complete safety is not a realizable goal, the regulations
have continued to be enhanced since the following year.
4. Trends in Formula One Regulations and

Their Effects
Formula One regulations are formulated based on
considerations of fairness, competitiveness, safety,
sustainability, and entertainment value(1). If the changes
in the regulations discussed above are summed up as a
whole, certain trends can be observed from the 1990s to
2009, and these trends have also had a significant impact
on the direction of technical development. These trends
will be discussed below.
In 1994, the first driver death in eight years made the
1990s a period in which reconsideration of Formula One
safety was pushed ahead rapidly. From 2000, when
Honda reentered Formula One racing, until today, the
application of limitations to vehicle performance and the
achievement of increased vehicle safety have been
constantly promoted. The limitation of vehicle
performance has mainly focused on restriction of driving
performance, primarily by means of engine regulations,
and the application of regulations concerning chassis
dimensions, aerodynamics, and tires, focusing on the
reduction of cornering speed.
In order to seek enhanced performance and help to
ensure competitiveness against the background of these
regulations, teams have pushed ahead with the
3.5. Materials
There had previously been few rules in the
regulations concerning materials of formula one vehicles,
but attention was directed to them when the use of
beryllium alloys, which are harmful to human health,
became an issue, ultimately leading to restrictions on the
use of expensive lightweight and high-rigidity materials.
In 2000, materials with a specific modulus of elasticity
of greater than 40 GPa (g/cm3) were prohibited from use
in any part other than the moving parts of the engine,
and in 2001 this prohibition was extended to the entire
vehicle. In 2003, methods of testing compliance were
detailed.
While materials development programs had
previously been focused exclusively on the achievement
of low weight and high rigidity, in response to these
changes, fine control began to be applied to the
composition and process of production of materials, and
more advanced technologies were pursued, for example
11
development of new high-efficiency mechanisms that

reduce loss and modifications based on heat and flow
analyses conducted using computers. Developments
focusing on the achievement of weight savings, increased
compactness, and higher rigidity universal and
important elements of development for racing have not
rested with simple substitution of materials, but have
entered a new stage, with limit design using structural
analysis technologies and the development of
sophisticated technologies to engineer characteristics and
functions into materials, among other initiatives.
In addition, a change in technical orientation that
seeks to increase competitiveness by enhancing not only
the performance of each technical element but also the
overall balance of the vehicle as a package, which could
already be observed in the 1990s, became particularly
important in the 2000s. The reduction of the maximum
width of the vehicles discussed above reduced the
clearance between the tires and the bodywork, and the
efforts to realize a low, slim rear cowl with consideration
of the flow to the rear wings was related to the reduction
in the size of the engine, the use of a narrow longitudinal
shaft gearbox, and, further, the length and arrangement
of the exhaust pipes.
A variety of measurement, analysis, and simulation
technologies have come to be employed to help ensure
the competitiveness of the vehicles, but for the
evaluation of the results of developments, it has become
necessary to once again consider what is after all the
fundamental predicate of Formula One, i.e., having
actual drivers run the vehicles at high speeds. The rapid
advancement of computer technologies has enabled
Formula One developers in the 2000s to process large
volumes of data at high speeds, and to push ahead with
sophisticated technical developments spanning a broad

range; at the same time, Formula One in this era is a
world in which systems of comprehensive technical
management that bring these technologies together and
technical philosophy themselves are under scrutiny.
Figure 1 shows how lap times in Formula One
qualifying practice sessions have changed since 2000 as
a result of these changes in the regulations and
subsequent technical developments. For the base data, a
track was selected on which there had been no change
in the length of a lap in nine years; results for rainy
weather were excluded. The figure shows the transition
in the rate of reduction in the fastest qualifying lap times
with the time for 2000 as the benchmark. As the results
make clear, despite ongoing changes in the rules aimed
at restricting vehicle performance, such as the 20%
reduction in engine displacement introduced in 2006,
there has been almost no increase in lap times. Times
have either been reduced, or have at least been held
steady.
This indicates that Formula One is technically
incomplete, and there is still margin for enhancement.
In particular, the importance of the effect of tires on
performance in contemporary Formula One is
demonstrated by the rapid reduction of lap times in the
period when multiple tire suppliers were in competition,
and the deterioration in lap times in 2005, when tire
changes were prohibited from the qualifying practice to
the completion of the final race.
It has long been feared that the fierce competition in
development for Formula One will lead to spiraling
costs, and specific regulations have been rapidly put in
place to reduce costs against the background of the
recent deterioration in global economic conditions. These
102%
No tire
change
Enforcement of
parc ferme rule
Enforcement of
2 engines for 1
race event rule
Mel
Sep
Sil
Mza
Int
Average
100%
Lap time proprotion
1 engine
for 1 race
event
Introduction
of 2.4L V8
engines
98%
96%
94%
Introduction of
engine
homologation
92%
Prohibition
of traction
control
90%
2000
2001
2002
2003
2004
year
2005
Competition between
multiple suppliers
Single tire
supplier
Fig. 1
Transitioin of lap time
12
2006
2007
2008
Single tire supplier
References
regulations do not restrict themselves to regulations

concerning the materials that can be employed and the
like; a variety of new regulations have been added,
including regulations extending the life of engines and
gearboxes, regulations concerning a homologation system
that will limit development itself, regulations restricting
the distance and the amount of days for running tests,
and a parc ferme rule designed to reduce the amount of
operations at the circuit by prohibiting repair or
adjustment of vehicles from the qualifying practice until
the completion of the final race. However, if for example
running tests are restricted, developers will turn to bench
tests and simulations, and the development of these
methods themselves will increase in importance. In
response, the FIA is seriously considering the
introduction of measures, including restrictions on the
amount of time that development facilities can be used
and a cost cap system that will control total development
costs.
Environmental measures are also being emphasized
in response to global warming, and the recent
recommendation of technologies enabling energy to be
reused and the introduction of obligation to control total
CO2 emissions from fuels are noteworthy points.
Efforts are also once again being made to highlight
the competition that is fundamental to racing, and to
increase the excitement and spectacle of Formula One.
The ability to make the spectators feel the differences
in skills and the individuality of the drivers, and to
transmit the thrill of a race fought out at white heat will
be a necessary element in helping to ensure the
sustainability of Formula One into the future.
As indicated by the discussion above, we can point
to present day as an era in which Formula One is
reflecting upon the very meaning of its existence as it
responds to rapid social changes. A vehicle that is
responsive to any driver input will run fast, and
eventually be safe, non-fatiguing, and enjoyable to drive.
While the environments in which and the speed at which
they are operated are very different, Formula One
vehicles and production vehicles share this fundamental
characteristic. The author can only express his hope that
technical development for Formula One will continue to
contribute to fundamental advances in vehicles.
(1) Tanaka, H.: The Transition of Regulation and

Technology in Formula One (20074177), JSAE
SYMPOSIUM, No. 22-06, p. 32-41 (2007)
(2) Shirai, H.: Evolution of F-1 Regulation, Honda R&D
Technical Review, F-1 Special, p. 6-9 (1993)
(3) Fujita, H., Horiuchi, Y., Kubonoya, H., Baba, J., Sasaki,
K: Development of Electric Power Steering System for
F-1 Race Car, Honda R&D Technical Review, Vol. 14,
No. 2, p. 45-50
(4) Tanaka, H.: Safety Technology in Formula One Racing,
Journal of Japan Society of Mechanical Engineers, Vol.
109, No. 1048, p. 148-149 (2006)
5. Conclusion
Formula One encompasses a variety of aspects,
including competition, technical rivalry, and
entertainment, and the like. Having failed to achieve a
good record of victories, it would be difficult to say that
Hondas activities during its third Formula One era
achieved their purpose in increasing recognition of the
company or enhancing its image. However, we applied
ourselves with a respect for sportsmanship, a dedication
to technical development, and an indomitable spirit. The
various technical outcomes and primary experiences that
were obtained are valuable assets to the Honda, and the
process itself can be said to be emblematic of the Honda
approach.
Author
Hiromasa TANAKA
13
Summary of Honda Formula One Engine

in Third-Era Activities
Kazuo SAKURAHARA*
ABSTRACT
Honda has entered six models of V-type engines with 10 cylinders (below, V10 engine) and three models of Vtype engines with eight cylinders (below, V8 engine) as third-era Honda Formula One engines. The goals of
development were to achieve output approaching that of the second era, the turbocharged engine era, with natural
aspiration, and to realize a smaller, lighter engine with a low center of gravity, focusing on the vehicles height of
gravitational center, weight distribution and aerodynamics. Revising the structures of different parts, modifying materials
and surfacing processes, and making design engineering and evaluation technique progress among other advancements
resulted in per-liter power that is 100 kW/L greater than a second-era naturally aspirated V10 engine, as well as having
approximately 6500 rpm higher engine speed for peak output, being more than 50 mm shorter in length and about 70
kg lighter, having about 50 mm lower crank center height and more than three times the mileage. Honda has proceeded
with development of V8 engines with the goal of high engine speed and won a third-era victory. However, because
of regulations restricting maximum engine speed and an engine development freeze due to homologation regulations,
issues relating to drivability (below, DR) have been left unsolved.
1. Introduction
than 740 kW (1000 hp), and as Table 1 shows, the

transition to 3.5 L naturally aspirated engines began in
1987, while from 1989 all engines became naturally
aspirated with a maximum of 12 cylinders. In 1995, after
Honda had taken a break from being a works engine
supplier, engine displacement was diminished to 3.0 L
and the V10 engine became the mainstream. From that
point on, changes in regulations became less significant,
and the era of the 3.0 L V10 engine lasted 11 years until
2005. The third era began in 2000 and lasted for six
years under these regulations. Subsequently, the engine
regulation changed to a 2.4 L V8 in 2006, and in the
inaugural year of the V8 engine, our victorious third-era
engine was realized.
Development of the third-era Honda Formula One

engine began in the autumn of 1998 with the program
of returning to racing in 2000. Honda has launched a
total of nine engines: six V10 engines starting with the
RA000E in 2000 and continuing through the RA005E;
and three V8 engines from the RA806E to the RA808E.
This article gives an overview of Hondas third-era
activities, looking back on second-era engines and
recounting the movements through the third era, then
reviewing the progress made by comparing second-era
and third-era V10 engines.
2. Trend from Second Era to Third Era

2.2. Output
The key requirements of second-era racing engines
were output, drivability, reliability, lightness of weight,
compactness, and fuel consumption, which are the same
elements needed today. A different era and different
regulations, however, mean different priorities, so
second-era engines, built when vehicle aerodynamics
were still simple and slick tires were allowed, were
developed strongly focused on output.
As Fig. 1 shows, the Honda Formula One engine
2.1. Regulations
Development of second-era engines continued for 10
years, 1983-1992, with Honda as an engine supplier.
During that period as well, engine regulations were
changing greatly, and in 1983, when Honda restarted
Formula One racing, the 1.5 L turbocharged engine was
the most common type. Subsequently, regulations
limiting boost pressure and total amount of fuel
consumption were put into place to limit outputs of more
14

Table 1
Year
1984
1986
1987
1988
1989
1991
1992
1993
1995
1999
2000
2001
2002
2003
2004
2006
2007
2009
F1 engine regulations
engine era). This shows that the technology of the engine

itself had advanced at high speeds, and some of the
content relating to this will be mentioned below in the
comparison of second-era and third-era engines.
Amendments to F1 Engine Regulations

Engine capacity (with supercharging) : 1.5 L maximum
(without supercharging) : 3.0 L maximum
Number of cylinders : 12 maximum
Fuel storage capacity : 220 L maximum
Fuel RON : 102 maximum
(without supercharging) : 3.5 L maximum
Manifold pressure (turbocharged engine) : 4.0 bar maximum
Manifold pressure (turbocharged engine) : 2.5 bar maximum
Fuel storage capacity (turbocharged engine only) : 150 L maximum
Supercharging prohibited
Engine capacity : 3.5 L maximum
Fuel RON : 102 maximum , Fuel MON : 92 maximum
Fuel RON : 100 maximum , Fuel MON : 90 maximum
Fuels of a kind used by general public mandatory
Engine capacity : 3.0 L maximum
Fuel sampling at circuit
Fuel approval before use
Throttle and pedal relationship fixed whilst car is in motion
Restriction for engine and clutch control (traction control prevention)
Cooling system pressure 3.75 bar maximum
Fuel incorporate 2000 EU limit
Number of cylinders : 10 maximum
Fuel incorporate 2006 EU limit
High specific modulus of elasticity material (Be-Al, etc.) completely prohibited
Traction and launch control permitted
Spraying substances other than fuel into engine prohibited
Parc ferme rule (no engine change between qualifying and race)
1 engine for 1 race event
Launch control and fully automatic gear shift prohibited
90 degree V8 2.4 L engine
1 engine for 2 race events
Engine homologation (freezing of engine main component development)
Rev limit : 19000 rpm
8 engines throughout 1 race season
Rev limit : 18000 rpm
KERS (hybrid system) permitted
2.3. Weight
This part discusses changes in the weight of engines.
As Fig. 2 shows, the weight of the engine alone was
120-130 kg even during the V6 era, and there have even
been some naturally aspirated V10 or V12 3.5 L engines
weighing 155-160 kg. Magnesium and titanium were
used at the time as materials to make engines lighter, but
even so engines could hardly have been called very light.
In 1992 and beyond, although this was not a works
activity for Honda, engine output seemed to stagnate, as
Fig. 1 shows, in part because of the impact of regulation
change and, the lack of competition among works engine
suppliers. For that reason, the chassis constructor teams
started emphasizing weight distribution and
aerodynamics, and making the engine lightweight with
a low center of gravity became more important than
before. As a result, engine development advanced year
by year during the era when Honda was racing as
Mugen-Honda, and engine weight reached as low as 122
kg by 1998.
Coming into the third era, there was even greater
need for lightness of weight, and development has
continued until V10 engines finally weighed just 89 kg.
780
740
already had output of 440 kW (600 hp) during the 1983

races, and the era of the 1.5 L V-type engine with 6
cylinders equipped with turbochargers (below, V6
engine) further enhanced output by increasing boost
pressure. By 1986, the engine had achieved 770 kW
(1050 hp), actually increasing output by more than 300
kW in three years, with the result that regulations were
changed to prohibit supercharged engines.
Another factor that increased output in this era was
the development of fuels. As Table 1 shows, under the
regulations of this era, there were few limitations on fuel,
and in particular no restrictions on energy density, so
fuels came into use that deviated substantially from
commonly available products. Starting with the 1992
Hungarian Grand Prix, teams were obliged to use
premium levels of market fuels and the use of special
hydrocarbons was prohibited, so output actually dropped
by 32 kW (43 hp) as compared to the German Grand
Prix that preceded it. This example demonstrates how
fuel contributed to engine output.
In the third era, even stricter regulations came into
effect on premium levels of market fuels in 1999 and
thereafter to prevent excessive competition, and there
were no major increases in output as a result of fuel as
in the second era. In the V10 engine era, however, when
the free development of engines was allowed, the
performance of naturally aspirated engines advanced to
rival the output of the second era (the turbocharged
Power (kW)
700
660
1.5L V6 T/C
620
3.0L V10
3.5L V10 or V12
580
2.4L V8
540
500
460
From Hungary GP
Mugen-Honda
420
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
Year
Fig. 1
Honda F1 engine power
170
3.5L V10 or V12
160
Weight (kg)
150
140
130
1.5L V6 T/C
3.0L V10
120
With out T/C system
110
2.4L V8
100
90
Mugen-Honda
80
1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008
Year
Fig. 2
15
Honda F1 engine weight
Summary of Honda Formula One Engine in Third-Era Activities

Table 2
RA100E versus RA005E

RA100E
RA005E
V-Bank angle (degree)
72
90
Power (kW/L)
143
240
Ps peak Ne (rpm)
12250
18700
Engine length (mm)
633.0
581.5
Engine weight (kg)
160
89
Crank center height (mm)
109.0
58.5
Mileage (km)
400
1400
regulation; there was no choice in the matter. With V10

engines, all teams were already using 90, so there was
no problem with this. However, it means that no unique
engines such as Renaults would come into existence
thereafter.
3.2. Comparison of Output
A comparison of output per liter shows that the
RA005Es power is 240 kW/L, or about 100 kW/L
greater than the RA100Es 143 kW/L. The elements of
enhancing performance are to increase volumetric
efficiency and combustion efficiency and to reduce
friction; these are fundamental to engine development
and there is nothing special about them. Since this
concerns naturally aspirated engines, engine speed is the
dominant factor in volumetric efficiency, and the
RA005Es peak output was achieved at 18700 rpm, 6500
rpm higher than the RA100Es 12250 rpm. Necessary
conditions for achieving high engine speed are to
stabilize valve behavior by reducing valve train
equivalent mass and to increase durability and reliability
of the reciprocating system. The Honda Formula One
engine reduced equivalent mass specifically by changing
from a direct-driven tappet system to a rocker arm
system, and furthermore switching materials from
titanium to a titanium/aluminum alloy for the valves and
making the stems more slender. In the reciprocating
system, Honda was able to reduce the weight of moving
components by using aluminum matrix composite
(AMC) pistons and box-structure connecting rods
(below, conrods) made with an intermetallic bonding
production method. By additionally using an alloy with
good heat conduction for the conrod bearing (plain
metal), the engine reduced the temperature of sliding
against the crankshaft pin and achieved reliability at high
engine speed.
3. Comparison of RA100E and RA005E

3.1. Comparison of External View
The following compares the RA100E in 1990 and
RA005E in 2005, two engines of the same V10
configuration from the second and third era, respectively,
to discuss in broad terms the progress that was made
over 15 years. Table 2 shows comparative values.
As the photos of the external view shown in Fig. 3
indicate, the RA005E is not so high. This difference has
to do with the change in V-bank angle from 72 to 90;
naturally, an engine with a large V-bank angle will have
a low center of gravity. In a V10 engine, when the left
and right banks have a common crank shaft pin, a Vbank angle of 72 with even firing intervals would
theoretically be advantageous, considering the load
placed on the pin. However, in the third era, engines
were already using an 80 angle as early as 2000, the
inaugural year of that era, daring to choose uneven firing
intervals. In 2001, there was an angle of 80, in 2002,
94, and since 2003, 90, so priority was not given
completely to even firing intervals. The reason is that
enhancement of aerodynamics and a lower center of
gravity took priority as elements determining vehicle
speed in Formula One cars in the third era; the basic
structure prioritized the total car packaging layout instead
of chasing some ideal for the engine itself, which is after
all just one component of the vehicle. Renault went so
far as to use V-bank angles of greater than 100 for a
time in its try for a lower center of gravity. The V8
engines of the time had to have an angle of 90 by
RA100E
Fig. 3
3.3. Comparison of Overall Length

A comparison of engine dimensions shows that,
whereas the RA100E is 633 mm long, the RA005E is
581.5 mm, or 51.5 mm shorter. Cylinder bores are 93
mm and 97 mm, respectively; the RA005Es bore
diameter being 4 mm larger means that its length would
be 20 mm longer, or 653 mm. Converted from that basis,
it is 71.5 mm shorter.
Broadly speaking, there are three technical elements
that made the shorter engine possible. The first was that
it became possible to join the cylinder block (below,
block) and cylinder liner (below, liner) into a single
piece. In the earlier Formula One engine, the block and
liner were separate, and they were generally built into
wet liner structure so that cooling water could make
direct contact. The separate structure has commonly been
used because the liner undergoes special surface
finishing where it slides against the piston, which makes
it more convenient if it is separate from the block. The
separate structure has had the additional advantages of
offering a high degree of design freedom for the water
jacket of the block itself, and during rebuilding, only the
liner needs to be replaced. However, recent
RA005E
Engine photo
16
advancements in block casting technology and piston

sliding surface-processing technology have made it
possible to integrate these into a single piece. Moreover,
higher engine speed has decreased block life, whereas
engine assembly life has increased due to regulation
requirements. Thus, engine assembly life and block life
are now about the same, so there is less meaning in
replacing only the liner; in the third era, the movement
to a single piece has advanced, and this is more or less
the case with other engine suppliers. In addition, the
structure of the gasket that seals between block and
cylinder head has changed greatly, so the gasket is now
an O-ring type, sealing just the bore perimeter, instead
of the old sheet form. As a result, the dimension between
cylinder axes is shorter, making it possible to reduce
engine overall length.
The second technical element is that advancements
in technology to make gear materials stronger and reduce
gear vibration have enabled a change of valve gear train
with a reduction system to one with a serial system, thus
making it thinner.
The third element is that advances in conrod bearings
have allowed them to become narrower, which has
reduced the amount of offset of the left and right banks.
they prohibit the use of magnesium, titanium, resin, and

CFRP as materials. If these restrictions were not in place,
calculations suggest that designs weighing about 78 kg
would be possible, which indicates how much the
technology to reduce engine weight has evolved.
3.5. Comparison of Crank Center Height
Crank center height is a factor that determines the
engines center of gravity and affects the vehicles
overall height of gravitational center, and as such is an
important dimension determining vehicle dynamics.
Current regulations thus stipulate that it must be at least
58 mm high to prevent excessive competition, and
furthermore that the engines center of gravity itself must
be at least 165 mm from the bottom of the engine.
Whereas the RA100E has a crank center height of 109
mm, in the RA005E it is 58.5 mm, or more than 50 mm
lower. Broadly speaking, there are three technical
elements that made this possible. The first is stronger
ferrous materials, which allow for more slender
crankshafts, as well as more slender conrod bolts. As a
result of combining these newly developed parts, smaller
conrod locuses are realized. The second element is
advances in crankshaft counterweight design engineering.
Materials with high specific gravity, such as tungsten,
that are used as balance weights can now be bolted down
directly under even high levels of centrifugal force such
as at 20000 rpm, which has reduced the crankshafts
radius of gyration. The third element is modifications in
clutch friction material and construction, so that even
with a smaller diameter it is still possible to get
sufficient torque transmitting capacity.
3.4. Comparison of Weight

The weight of the RA100E is 160 kg while that of
the RA005E is 89 kg. This reduction of about 70 kg in
15 years represents a weight reduction of more than
40%. Even just looking at the third era, the RA000E in
2000 was 112 kg, so engines reduced 23 kg in five
years: an average yearly reduction of 4 - 5 kg. Although
the push to lower weight stagnated for a time in 2002,
the engine resumed reducing weight in 2003. Magnesium
and titanium, useful for reducing weight, were already
being used in the second era, so in the third era it has
been thought that there are few advantages of using these
materials. Advances in design technology can be
mentioned as a reason for even further weight reduction.
From the second era to sometime in the third era, design
methods were mostly 2D, but starting around 2003, 3D
gradually came into use, and with the implementation of
CATIA V5 shown in Fig. 4, completely 3D design and
CAE became easy to use at the designer level, and
engines became correspondingly lighter. In order to
prevent excessive competition, current regulations
stipulate that V8 engines must weigh at least 95 kg, and
3.6. Comparison of Mileage

In 1990, when the RA100E engine was in use, a
single driver used at least three engines during one
racing event, because they were allowed to use one for
the Friday free practice, one for the Saturday qualifying
practice, and one for the Sunday final race. Since the
mileage of the final race is about 300 km, the engine
only needed to have reliability of 400 km, including the
Sunday morning warm-up session. The RA005E engine,
on the other hand, had to have durability of about 1400
km, because it arrived in the second year of regulations,
begun in 2004, which intended to reduce high engine
costs by requiring that a smaller number of engines be
used. After that, each driver had to use just one engine
during two race events. This restriction remained in force
until 2008; beginning in 2009, each driver was permitted
to use a maximum of eight engines in a year. As noted
above, advances in the engine itself and in evaluation
methods can be mentioned as technical elements that
more than tripled the durability and reliability of engines
as compared to the second era. In addition to the
previously mentioned advances in the engine itself, there
have also been advances in surface modification
technology, as exemplified by diamond-like carbon
(below, DLC) coating. The evolution of DLC has been
particularly remarkable, so that, in almost all of the
major sliding areas of a Formula One engine, DLC of
3D design
CAE
Fig. 4
CATIA V5
17
Summary of Honda Formula One Engine in Third-Era Activities
one specification or another is currently used. If this

were to be prohibited, it is likely that both reliability and
performance would diminish greatly. The next
advancement is in evaluation methods, the biggest factor
in which is the use of low-inertia transient dynamos
(dynamos with the same inertia as wheels and tires),
which allows personnel, using the engine alone, to
perfectly simulate driving a circuit. Engine durability and
reliability are affected when using traction control or by
irregular revolution resulting from ignition cuts when
engine speed is restricted, but in the past it was only
possible to reproduce this by actual driving, so durability
and reliability could not be definitively guaranteed in
advance for all 16 or more circuits.
a lot of reflecting on the fact that there has been little

development emphasizing this.
5. Conclusion
This paper has looked back on Hondas Formula One
engines, using numerical values to compare the second
and third eras, which reminds us anew of how amazing
the technical advances have been over the course of 15
years. Because there are presently so many regulations,
there is little room for further advances such as these,
but it is still supposed that the technology will advance
bit by bit hereafter. Regardless of whether Honda returns
to Formula One racing, the company feels it must
endeavor to preserve the information network it has
created and pursue further technical advances (concepts).
Otherwise, in such a rapidly advancing world, it may
take a great deal of time to catch up if we again get the
chance to take part in Formula One engine development.
4. V8 Engines
Development of V8 engines, the use of which has
been obligatory since 2006, began in November 2004.
In May 2005, before any other team did so, Honda
conducted test runs on the Jerez circuit in Spain with a
prototype engine, a V8 adapted from the 2004 V10
model. Having a goal of maximum engine speed of
20000 rpm in Hondas 2006 racing engine, development
was carried on at a quick pitch, with first firing test on
dynamometer in August 2005. The result was Hondas
only third-era victory, at the summer Hungarian Grand
Prix, which proved the excellence of Hondas V8 engine.
Development was proceeding that sought an even higher
engine speed for 2007, but regulations limited engines
to a maximum of 19000 rpm. Meanwhile, homologation
regulations limited annual development of engines
themselves, and engine specifications have been
determined that have become the basis for the
development freeze since 2008, specifications that Honda
was unable to deal with satisfactorily. As a result,
transient combustion has not stabilized, and the 2008
prohibition on traction control has additionally caused
DR issues to emerge. Honda has taken measures as far
as development is possible, but has not achieved any
fundamental solutions. Since DR was still insufficient,
the season was a disappointing one. Talks with the FIA
resulted in a partial lifting of the development freeze for
2009, and Honda began development in October 2008
to solve DR issues in anticipation of the opening of the
next season, and confirmed in December that
performance was as targeted. Unfortunately, Honda has
withdrawn from Formula One activities, so it has not
been possible to prove the results of enhancements. The
emphasis in current Formula One engine development is
not to put the foremost priority on maximum output,
which has always been considered Hondas strength, but
rather on improving output characteristics, particularly in
the range where they affect DR, and on somehow
eliminating irregular combustion. The reason is because
even if output were increased by 10 kW, it would only
improve lap times by less than 0.1 seconds, whereas in
comparison poor DR makes the driver lift off the throttle
pedal while coming out of the corners, resulting in a loss
of more than 0.5 seconds. Honda engineers need to do
Author
Kazuo SAKURAHARA
18
Overview of Gearbox Development for Formula One
Atsushi MANO*
ABSTRACT
In the racing world, the transmission is ordinarily referred to as the gearbox. Honda first tried developing an Formula
One racing gearbox with its third-era Formula One activities. To do this, it was necessary not only to develop the
technology, but also to solve several problems, including production, supply and operation. This article recounts how
Honda overcame these problems to bring a number of technological firsts to the racing world and gives an overview
of the advancement of Formula One gearbox technology.
1. Introduction
Other than the parts newly established to avoid

excessive technical competition, these restrictions are
only a general framework, and there is overall a very
high degree of freedom in the design of gearboxes.
Because of that, each Formula One team is aggressively
developing technology with the major objectives being
lightness, compactness and high efficiency.
Generally, when people talk about Formula One

technology, engine power and aerodynamic performance
are often mentioned, but there are few opportunities to
bring up gearbox technology. This article, therefore, will
start by discussing the basic elements essential to
gearboxes and recount how Hondas development
advanced these elements.
2.2. Required Performance

The following lists the performance required of a
Formula One gearbox.
(1) Durability and reliability
(2) Lightness and compactness
(3) Lower center of gravity and low yaw inertia moment
(4) Quicker gear shifting
(5) High transmission efficiency
(6) Stiffness of casing
Unlike mass-produced vehicles, the rear suspension
is directly attached to the casing (Fig. 1), so sufficient
2. Formula One Gearboxes

2.1. Regulations
The following is a brief excerption (not a verbatim
quotation) from the Formula One technical regulations
on transmissions as stipulated by the Federation
Internationale de lAutomobile (FIA).
(1) Only two-wheel drive is allowed
(2) It must be possible to cut the clutch manually when the
vehicle is stopped
In the event that a vehicle stops along the course
because of vehicle trouble or any other reason, course
attendants must be able to push it out of the way, so
vehicles have a manual clutch cutoff switch that even
works when the engine is shut off.
(3) The minimum number of forward gear ratios is four and
the maximum is seven, and CVT is prohibited
(4) Vehicles shall have a reverse gear
(5) Left/right torque transfer is prohibited
In addition to the above, the following rules were
added in 2008 to control costs.
(6) A single gearbox must be used in four consecutive races
(7) Gear ratio pairs must have a minimum thickness of 12
mm wide with at least 85 mm between centers, and each
set of gears must weigh at least 600 g
Fig. 1

19
Appearance of F1 gearbox
stiffness and strength are required, and this has an impact

on vehicle behavior.
(7) Must contribute to aerodynamic performance
The gearbox must be as small as possible so that it
does not impede airflow to the rear wing and diffuser
at the back of the vehicle. In particular, the rear part
must be narrow.
(8) Easy to maintain
It must be possible to complete ratio gear changes in
the interval between racing sessions.
In addition, the internal mechanism transmits engine
torque input through the clutch through seven sets of
gears, a bevel gear and final gear, as shown in Fig. 2,
and from there through the drive shaft to the tires.
Ratio gears
Final gear
the two sides agreed that Hondas Tochigi Automobile

R&D Center (HGT), a development base in Japan, would
oversee development of technology, especially for gears
and shafts, and full-scale development began in HGT in
2002. Subsequently, development continued with the
same division of responsibilities even after the team
reorganized into the Honda Racing F1 Team (HRF1) in
2006.
3.2. Learning Formula One Gearbox Development
Technology
Before development could proceed, Honda first had
to learn the level of technology in Formula One gearbox.
The following four major issues had to be addressed in
order to do this.
(1) Design engineering
As for the type of gear, mass-produced automobiles
use helical gears, where the emphasis is put on gear
noise. In contrast, racing cars use spur gears, with more
emphasis on efficiency and strength. Although there are
differences, the design tools are the same, and the basic
design techniques of the former could be used in the
latter. The conditions under which racing cars are used,
however, made it impossible to use the historical data
from mass-produced automobiles as reference, and
Honda learned that it is necessary to deal with early
onset gear surface pitting and to make revisions to gear
teeth tips, taking into account elastic deformity of the
gear teeth caused by impact load input from such forces
as shift shock.
(2) Production technology
In order to ensure gear strength and maximize
transmission efficiency, all gears underwent heat
treatment and subsequent gear grinding (that is, grinding
of the entire tooth down to the bottom), and then for the
surface finish underwent a treatment to reduce surface
roughness of the tooth (including barrel grinding) and
shot peening to enhance fatigue strength. In particular,
the grinding process required numerous rounds of trial
and error and new arrangements in order to get the same
level of surface roughness as the specialist
manufacturers gears, which were used as reference. In
addition, during shot peening, in order to give the
material areas with higher surface residual stress and
deeper residual stress, Honda had to spend much time
catching up to the specialist manufacturers performance,
not only selecting media but using multistage shot and
choosing the right process pressure and times.
(3) Materials technology
The gear materials that were being used in the racing
industry were the same materials used in mass-produced
gears but with extra strengthening, and to be competitive
it was necessary to develop specialized high-strength
gear material. Honda worked with steel manufacturers
and had to spend many hours until it could get
satisfactory material performance.
(4) Preparing a cooperative prototype production system
To develop competitive technology in timely fashion,
one absolutely must have production support in the form
of fast prototype-production times by a number of
Differential
Bevel gear
Clutch
Fig. 2
Structure of F1 gearbox internals
3. Preparation for Original Development

3.1. Establishment of Joint Development Partnership
with Racing Team
Hondas third-era activities began in the form of joint
development of racing cars with British American
Racing (BAR). Because development of the gearbox was
mainly done by the race team BAR, the first issue to
resolve was how to proceed with joint development. The
development conducted by BAR at the time was on the
part concerning the overall package, including the casing,
while detailed design and production of internal
components like the gear shaft were contracted out to
manufacturers specializing in racing transmissions.
Therefore, although BAR was able to present its
requirements relating to gear and shaft layout, it was
difficult to make these parts lightweight and develop
innovative mechanisms faster than other teams, and even
if it came up with good ideas, it was difficult for BAR
to use them exclusively. On the other hand, Hondas
strengths were its ability to manufacture casings, gears
and shafts internally and the fact that it had bench testing
equipment to evaluate the strength and reliability as well
as performance of gearboxes, something which other
racing teams at the time did not have. If Honda therefore
could provide its competitive technology in place of the
specialized manufacturers, the advantages to BAR would
be significant, such as being able to develop pioneering
technologies and use them exclusively. For that reason,
20
manufacturers specializing in the area of prototype

production. Honda received support from many proven
prototype producers of gears and was able to
manufacture gears and shafts with relative ease, but it
faced more difficulty in prototype-producing bearings.
When developing new gearboxes for racing purposes, the
use of bearings of a specialized form contributes to
performance enhancement. Bearings, however, are highly
specialized functional components, and there are no
manufacturers who prototype-produce them. As a result,
Honda had to depend on manufacturers of massproduced bearings to prototype-produce some, but the
manufacturers had difficulty meeting Hondas need for
small quantities of many types, so the prototypeproduction period grew longer and there were frequent
situations where the manufacturers could not meet
Hondas development needs. However, thanks to
development support from NTN Corporation, Honda was
able to build a cooperative development system for
Formula One bearings. As a result, the time from
examination of specifications to prototype production and
supply of components was shortened and it was possible
to work in a timely manner.
(2) Faster gear shifting time

Beginning in the early 1990s, Formula One autos
used a sequential shift mechanism consisting of a dog
clutch and a shift barrel driven by a hydraulic actuator,
as shown in Fig. 3, a system that works on the same
principle as that used in motorcycles. Particularly when
upshifting during acceleration, the vehicle can decelerate
with a force of as much as 1.0 G from air resistance
because of interruption of torque delivery, so reducing
shifting time is a crucial objective directly linked to
better lap times.
3.4. Establishment of Racing Component Supply
System
In order to bring competitive, unique technologies,
one must not only develop those technologies but also
produce and supply the number of components needed
for operations with stable quality. Prior to this
development, Honda had supplied assembled complete
engines to a race team, but this was its first experiment
supplying individual components. For that reason, it was
necessary to create a system, including the organization,
in order to achieve the two following points.
(1) Production quality assurance
For those important components associated with the
driving function, the history of each component is
controlled (this includes control of production lot and
driving lifecycle). For this purpose, the part number,
serial number and production lot are controlled.
Infrastructure needs to be prepared so that this
information can be placed directly on components by
laser marking. Honda has not only introduced a laser
marker, but also established rules, routes and a system
for conducting this work.
In addition, it was necessary to attach an inspection
report card to each component as evidence that it
satisfied blueprint quality. To do this, Honda arranged
inspection systems for all components and decided on
serial number engraving and key point control
dimensions. Honda additionally established rules for
writing concession reports on rescue measures for
components slightly outside the tolerance which were
judged to be no issue functionally, and a system was set
up at both HGT and HRF1 to determine whether
components presented any functional issues, and if so,
to rescue them. This effort sought to prevent unnecessary
cost increases and stabilize component supply.
(2) Component shipping system
Since HGT did not previously have a shipping
function for individual components, Honda prepared the
processes from packaging to shipping of finished
individual components. To make sure that ratio gears,
final gears and so on could be efficiently stored and
assembled at the factory or circuit, Honda created special
packing boxes designed to prevent rust, cushion impact
and simplify component identification.
3.3. Technical Development Goals

To address the issues mentioned in the preceding
sections and ensure competitiveness in the area of
gearboxes, the major development goals were narrowed
down to the following two, which are big factors in lap
times.
(1) Lightness and compactness
Regulations dictate that racing vehicles must weigh
no less than 605 kg, including the driver, so a lighter
gearbox does not lead directly to a lighter vehicle.
However, by increasing ballast weight to adjust overall
weight, one can try to increase the degree of freedom
for vehicle weight distribution and lower the center of
gravity. Also, greater compactness not only leads to
lighter weight but also enhances aerodynamic
performance by giving the gearbox a slim form. These
contribute to the vehicles competitiveness.
Ratio gear
Dog ring
Shift barrel
Shift fork
4. Details of Technical Advancement
Rotary actuator
Fig. 3
Competition among racing teams to develop
Appearance of F1 gear-change system
21
ranking teams procured gears from specialty

manufacturers. Internal mechanisms standardized for
Formula One use were customized into packages for
each teams vehicle. As stated above, greater
compactness of the gearbox itself contributes
significantly to the vehicles competitiveness, and
therefore active initiatives have been taken to reduce the
size of gears. Techniques for doing this include using
high-strength gear material, optimizing the safety factors
by using FEM, and advancing the production methods.
HGT pursued independent development starting with the
gear material, and by enhancing root bending strength
and tooth surface pitting strength, it was able to reduce
gear width by 1 mm for first through third gears and by
2 mm for fourth through seventh gears, as compared to
the gears produced by a specialty manufacturer which
had been the base up to that point. This also reduced
total gear weight by approximately 1 kg. The new gears
were released with the final stage of the 2003 season and
thereafter were the standard. At first there was some
unexplained gear tooth damage (breakage from the root),
but analysis revealed the mechanism: shock loads more
than anticipated while gear shifting and sudden changes
in vehicle behavior caused some gear teeth to undergo
plastic deformation, lowering fatigue strength, and
causing breakage in a short span of time. As a result,
design standards were established and applied to prevent
tooth deformity in ordinary use. In terms of design,
however, increasing strength and reducing weight are
contrary to each other when dealing with greater than
anticipated inputs, and this reduces competitiveness.
Therefore, the engineers developed controls to keep
excessive input from occurring, and by integrating torque
sensors they put a system in place to monitor input
values on gears at all times, so that if a gear underwent
more than the allowed torque, that gear would not be
used again. This effort helped to reduce gear weight in
balance with gear strength. The engineers used a gear
tester to identify the torque at which tooth bending
occurred and found the correlation to design stress,
which made it possible to set a highly precise design
allowable stress.
technology has caused gearbox technology to evolve

during the third-era Formula One activities. This section
introduces and reconsiders the details of this
advancement, looking at changes in the trends at other
teams and specific technologies worked on at HGT.
4.1. Casing
Since the casing also functions as part of the chassis,
it must be highly stiff as well as lightweight. Techniques
used by Honda to achieve this were research into a
variety of materials and development of production
methods. In the past, sand casting of aluminum alloys
and magnesium alloys was standard procedure, but in
recent years the paths that racing teams have chosen
have gone in two different directions(1), (2). The first path
is to advance the casting production methods and save
weight by casting with thinner walls. This approach,
which used rapid prototype technology, makes the walls
thin and at the same time, since there are no restrictions
on draft or undercut, it is possible to eliminate useless
material. Aluminum or titanium is chosen as the casting
material. The second path is to use carbon fiber
reinforced plastic (CFRP) as shown in Fig. 4. CFRP can
offer the advantage of much greater specific stiffness
than metal materials and is also superior in terms of both
lightness and stiffness. It takes a long time to
manufacture, however, and the production cost is
disadvantageous compared to casting. As of 2008, the
various racing teams production methods have diverged
into the two methods mentioned, and there is no evident
trend of unification to one or the other. At HGT, the
engineers worked to develop a lightweight magnesium
casing using a CAE program to optimize stiffness, and
this was released at the end of the 2002 season. BAR,
however, developed a CFRP casing that could be made
even lighter, which was put into use starting in 2004,
so HGT stopped developing casings.
4.2. Gears
To procure gears, the top teams set up their own
technology development environments in which gears
were designed and produced internally. In contrast, lower
Fig. 4
4.3. Gear Shifting Mechanism

The sequential semi-automatic gearbox that appeared
in the early 1990s became established as the basic
structure, with no change until recent years. In the
meantime, attempts to reduce gear shifting time focused
on reducing operating time by reducing component
weight and friction, and on torque optimization control
by means of cooperative control with the engine, but no
great progress was made. To achieve faster gear shifting
times, one of their key development goals, the HGT
engineers developed a seamless gear-change mechanism
which eliminated interruption of torque delivery while
shifting. Theoretically, this mechanism allows the change
in gears to take place instantly by letting the next gear
engage while still driving at the previous gear. Broadly
speaking, there are two major issues involved as
described below to achieve this, and the Honda engineers
Appearance of CFRP gearbox casing
22
dealt with each of them.

(1) Avoiding interlock from double engagement of gears
Interlock generally results when engaging two gears
simultaneously, which causes gear or shaft damage. To
make double engagement of gears tolerable, therefore,
a one-way clutch structure was incorporated, enabling the
one-way clutch to idle even when the former gear
switches to a deceleration tooth surface after selecting
the next gear, and making it possible to avoid interlock.
However, if a one-way clutch is always in operation,
issues occur, such as being unable to use the engine
brake during deceleration, so the one-way clutch is
locked except when shifting, which prevents it from
idling.
(2) Increase in shock while gear shifting
Instant gear shifting causes instant change in torque
equivalent to the change in inertia based on the gear
ratio, so that spike torque increases, in addition to the
engine torque. In response, since a major portion of the
change in inertia is engine inertia, the engineers reduced
spike torque by developing optimization control over the
amount of clutch during gear shifting. Combining this
control with the mechanism described above made it
possible to achieve both this objective and gear strength.
Results of repeated bench tests and circuit tests
indicated that these issues had been resolved, so the
technology was used in races starting with the 2005
season. Lap times dropped by 0.4 seconds per lap, a very
significant advantage for a single technology. Moreover,
with no interruption of torque delivery while upshifting
and with reduced spike torque, there was less drive
fluctuation while shifting and it became possible to shift
even while cornering or when there was low tire grip,
such as in rainy conditions. Honda was the first to use
this technology for eliminating interruption of torque
delivery when shifting in a race, but since it was so
effective, other teams began eagerly developing the
similar technology, so that by 2008 all teams were using
it as seamless shifting. It is said that the structure used
by other teams, however, had two shift barrels, with the
even-numbered gears and odd-numbered gears controlled
separately. In contrast, the structure used by Honda
consisted of just one shift barrel and was unique in that
it prevented double engagement of gears, giving it
superior reliability and lightness.
The engineers furthermore undertook to develop an
innovative gear shifting mechanism with the goal of
producing a technology to reduce the overall length of
the gearbox. As Fig. 2 shows, the length of the gearbox
can be broadly divided into the clutch, ratio gear, bevel
gear, final gear and differential portions. Here, the
engineers focused on shortening the ratio gear portion
and began a series of investigations based on the idea
of eliminating the dog ring and shift fork located
between gears. The result they hoped to achieve was a
structure that, without losing the advantage of seamless
shifting, would reduce overall length by 19% and weight
by 12% by putting the shifting components inside the
main shaft and aligning the ratio gears without any gap.
Although it was difficult to ensure durability of an
internal shifting mechanism, the given target

performance was achieved as a result. It had been
decided to use this technology in circuit tests at the end
of 2008, but this plan was canceled when Honda decided
to withdraw from Formula One racing.
4.4. Clutch
In order to maximize launch performance at the start
of races, development was begun with the goal of
stabilizing clutch transmission torque. Conventional
mechanisms controlled clamp load against clutch friction
material by adjusting the piston force that hydraulically
operates the preload set by the diaphragm spring. This
structure made it difficult to control launch torque with
much precision because individual assembly statuses
caused variance in the clamp load, and wear of the
friction material caused diaphragm spring characteristics
to change. Figure 5 shows the Direct Push Clutch (DPC)
that was developed to address this. By using this
mechanism, the clamp force of the friction material was
acquired directly from the hydraulic force, eliminating
the element of variance and enabling high-precision
clutch-torque control. The greater part of load hysteresis
during operation, a factor that can reduce control
characteristics, is seal friction. To reduce this, the type
of hydraulic piston used was one that placed smalldiameter plungers at two points, minimizing the impact
of seal friction. This was first used as original
technology in 2006 and subsequently became standard
in HRF1.
4.5. Differential
The differential is one of the components showing the
most progress toward compactness in the past 10 years,
and according to data from a specialty manufacturer,
differential assembly weights were just 6.1 kg in the
2004 specifications, a reduction of more than 50% from
the 12.5 kg in 1998 (2). This progress is primarily the
result of advancements in package technology and
optimized safety factors. Since the differential is located
higher than the vehicles center of gravity, making it
lighter is a very effective way of lowering the height of
Clutch
Fig. 5
23
Twin
hydraulic
plungers
Appearance of Direct Push Clutch (DPC)

Table 1
the gravitational center. Reducing differential width,

moreover, helps enhance aerodynamic performance and
makes longer drive shafts possible, so the sweepback
angle of the drive shaft is smaller, which enhances
transmission efficiency. Differential mechanisms include
a bevel gear set-type and planetary gear-type, but the
latter is superior in terms of weight and total width. The
HGT engineers also actively engaged in an effort to
make differentials more compact, and developed a
number of original technologies. This section introduces
the Ultra Short Diff (USD), which used a unique
mechanism to try to achieve a more compact differential.
F i g u r e 6 compares USD to RA108, the 2008
specification, in cross-section. Using a full-engagement
double-pinion planetary has resulted in 20% less distance
between centers and 14% lighter weight. By the end of
2008, the engineers had finished ascertaining
performance, durability and reliability and were aiming
to introduce the technology to racing in 2009.
Classification of transmission losses (loaded)
Bearings
20%
Churning
14%
Pump
21%
5. Conclusion
Although the third-era Formula One activities
resulted in only one race victory, they also gave birth
to seamless shifting and a number of other industry-first
technologies, so Hondas engineers could at least take
pride in such developments as they faced the closing of
this era. As development began, it was marked by one
failure after another, but the engineers learned many
things and grew along the way. This is the prize won
from participation in Formula One racing and will
certainly be an asset for Honda into the future.
References
(1) Mano, A.: Development of Gearbox Technology for
Formula 1, Journal of Society of Automotive Engineers
of Japan, Vol. 59, No. 9, p. 8-11 (2005)
(2) McBeath, S.: F1 transmission trends, Racecar
Engineering, p. 34-42 (2005)
Author
Fig. 6
Seals
2%
through the 2007 racing season and on all gears starting

with the 2008 season opening race.
The engineers additionally addressed how to reduce
transmission losses through gearbox oil development.
Generally speaking, lowering the oils viscosity reduces
churn resistance, but simultaneously reduces oil film
thickness, which increases friction on tooth surfaces
when in a boundary lubrication status and, as a result,
lowers transmission efficiency. By reviewing various
base oils, the engineers were able to address both churn
resistance and friction loss on tooth surfaces, cutting
transmission losses by 0.4 kW.
4.6. Enhancement of Transmission Efficiency

The original key development goals were lighter
weight, greater compactness and faster gear shifting, and
Hondas unique development yielded results on these
goals and enhanced competitiveness, so in 2006 the
enhancement of transmission efficiency was addressed as
a key issue. Table 1 gives the results of measuring
transmission efficiency under maximum gearbox load at
the time.
As is evident, most of the transmission loss came
from the meshing of the gears. For that reason, the
engineers added surface finishing to their research with
the goal of reducing friction losses in gear engagement.
Diamond-like carbon (DLC) coating yielded good
results, and adding DLC coating to all ratio gears, the
bevel gear and final gear made it possible to reduce
transmission loss by 3.4 kW. On the subject of coating
durability, at first there was an issue with the coating
peeling off the bevel gear, which is subject to a
particularly high level of contact surface pressure. This
was solved by revising the materials and production
methods, namely enhancing adhesiveness by reducing the
underlying surface roughness and making the film finer,
so this technology was used in part starting midway
USD
Gears
62%
RA108
Comparison between USD and RA108
Atsushi MANO
24
Summary of Honda Third-Era Formula One

Chassis Development
Akio TONOMURA*
Shun KAWABE*
Nobuhisa NAKAYOKU*
ABSTRACT
The third era of Hondas Formula One participation started in 1998 with a stage of preparation for the return to
actual racing, which began in 2000 and extended up to their withdrawal in 2008. A summary view of the Honda
Formula One cars that were developed every year during this period of participation in competition is presented from
the perspective of chassis technology. An overview will be given of each years aims regarding the chassis and the
technology that was required for third-era Formula One racing, accompanied by photographs of the chassis. Each years
regulations and the aims of chassis development will also be surveyed from the separate perspectives of aerodynamics,
the chassis, and the like. The points on which development was successful will be discussed on the basis of racing
results, and points calling for rethinking will also be examined.
1. Introduction
Formula One vehicle testing is carried on with a

variety of different limitations, including the funding for
team operations as well as limitations on the distance of
test drives, limitations on the number of tires, and so on,
in addition to which the circuit conditions are constantly
changing due to weather, dirt, and tire rubber deposits.
When determining the sensitivity contribution of a single
factor on vehicle performance, it is necessary to test it
with all the other factors held constant. Under these
conditions, however, it always seems as though either
nothing more than a very rough evaluation or a mistaken
evaluation can be done. Therefore, every team competes
fiercely on the development of evaluation methods, using
bench tests in a stable environment that also includes
inexpensive simulations and wind tunnel testing. One
might say that the determining factor lies in these
evaluation methods. However, it is in the nature of
simulations and bench tests that the results may diverge
from what is obtained using the actual machine. In order
to compensate more or less correctly for this divergence
and to decide the direction for development, it is
essential to obtain correlation with the results of repeated
testing with the actual machine. This means that tests run
with the actual machine are necessary, and so the
development is caught in a dilemma. This is what makes
Formula One development such a challenge: it is a race
to overcome dilemmas like these, to find reliable
evaluation methods and evaluation criteria, and to
quickly discover the cars balance points.
This article will relate how the Honda team dealt
with this development race.
The Formula One cars of recent years cannot succeed

just by one prominent performance capability, but have
to achieve a high-level balance of all factors in order to
win. Even if one car has the most powerful engine, if it
is heavy it will damage the rear tires. If efforts to perfect
an aerodynamic advantage limit the space available for
the suspension, the chassis will be lacking in stiffness
and will become unstable. Going too far in reducing the
weight of parts will detract from rigidity and reliability.
The question of how to discern the points of optimal
balance as a whole for conflicting approaches like these
is the key to chassis development.
The team conducts actual driving tests in order to
search out the optimal balance points. However, testing
on a slow car will only yield low-level balance points.
For example, when a car has unstable braking, the brakes
cannot be applied hard, and since the brakes cannot be
applied, the tire temperature will not rise to the working
range, and the cornering grip will be reduced. Low grip
on corners means that the corner exit speed will be
lower, which reduces the top speed on the straight-end.
In this way, a sequence of various different factors can
fall into a vicious cycle. A team that has a fast car faces
the opposite situation, and the difference cannot readily
be overcome. These are the reasons that a new team
finds it so challenging to come into its own, and it takes
a considerable number of years to raise all the factors
to a high level. This is because there are various
constraints placed on the development environment.
25
2. The RA099: 1998-1999
is designed for a gearbox made as narrow as possible.

At that time, however, the mainstream type had exhaust
pipes that passed between the top and bottom of the
suspension to vent to the rear, and this was the form
followed in the RA099.
The only adventuresome step taken amid this
orthodox fabrication was in the top wishbone in front.
The arms were made of Carbon Fiber Reinforced Plastic
(CFRP), the fittings that attached them to the body were
made of titanium, and they were bonded together with
adhesive. Since that was the time when the use of CFRP
for suspension arms was just getting started in the
Formula One world, one arm was selected as an example
to begin development. The other suspension arms were
all made of steel.
The weight distribution on the front axle of the
chassis (hereafter referred to as the weight distribution)
was designed using calculations based on tire
characteristics data. The objective was 45%, and the
design faithfully achieved this. The wheel-base was 3020
mm. The height of the center of gravity was 250 mm
above the floor, which was lower than that of the
following years car, the BAR002. A crucial issue in race
car design is how to assure traction when cornering. This
was addressed by an effort to bring the jack-up power
that arises from the suspension link mechanism down
toward zero by locating the rear roll center more or less
on the rear floor. The RA099 was built in an orthodox
manner without any remarkable measures to reduce its
weight, but even so it turned in a series of good lap
times in 1999 tests. The car looked promising for the
next years races, but Dr. Postlethwaite suddenly passed
away in April. That May, the project was put back on
the drawing board, and RA099 development was brought
to a halt.
The RA099 came into being in late 1998 when Dr.

Harvey Postlethwaite, who had been working with
Tyrrell until 1998, came to Honda Racing Development
as technical director. The chassis fabrication and
simulation analyses were carried out on contract by the
Dallara Automobili company in Italy.
Honda was aiming toward entry into Formula One
competition from 2000, and the RA099 was intended as
a prototype to be test-driven the year before that. The
engine was a Mugen-Honda V10. The concept was true
to the basics: to build up a proven system by driving it
steadily as a dedicated test car and gathering definitive
tire data, aerodynamic characteristics, and directions to
take in vehicle setup so as to develop a car for actual
racing in 2000. The primary focus in this was the
aerodynamic concept that there was little variation in the
downforce (hereafter referred to as the aerodynamic pitch
sensitivity) due to changes in the chassis pitch angle. If
this sensitivity was excessive, then changes in the chassis
attitude when braking and accelerating would result in
excessive aerodynamic changes, causing instability. This
effect could not be mitigated through other tuning
elements such as the suspension springs or the weight
distribution, and the result was that the tire
characteristics could not be employed to full advantage.
Ascertaining the tire characteristics required raising the
tire temperature to the working range of approximately
70C and maintaining a speed that would keep the
temperature at that level. It was therefore crucial to
assure the cars aerodynamic stability.
The front suspension was largely like those of the
other Formula One vehicles of recent years, in a double
wishbone configuration that has torsion springs located
in front of and behind the rocker arm shafts. The rear
suspension employed a double wishbone with a coil
spring coaxial damper positioned perpendicularly to the
chassis and connected by a large rocker arm protruding
from the aluminum cast gearbox. This layout of the rear
suspension opened up space on the top surface of the
gearbox, allowing a third damper to be added easily in
later testing. The third damper consists of a spring and
damper that expands and contracts when the chassis
moves vertically with the wheels on either side in phase
(i.e., heave). Its purpose is to control pitching.
The layout around the gearbox today, when the Cokebottle shape and top exhaust represent the mainstream,
3. The BAR002: 2000
In 1998, Tyrrell was bought up and British American

Racing (hereafter BAR) was founded. The BAR001 was
built starting in 1999, and was entered in Formula One
competition. The BAR002 was the model from the
second year of BAR.
Honda contracted with BAR in 1999 to supply
engines and conduct joint development of the chassis
from 2000. The work on the BAR002 therefore
represented Hondas debut in the third era. The BAR001
from the previous fiscal year had repeatedly experienced
26
trouble every time it raced, and with two cars in a total

of 16 races, it had 22 retires. The reliability was
therefore upgraded, but even though the BAR002 placed
fourth and sixth in the opening competition, it
experienced mechanical trouble that resulted in Jacques
Villeneuve retiring from the race five times and Ricardo
Zonta two times out of 17 races. There was no choice
for the development effort throughout 2000, therefore,
but to focus on upgrading reliability, and weight
reduction was sacrificed accordingly. The fact is that the
car was able to load 27.5 kg of ballast in the opening
race, but the amount was down to 3 kg in the final race.
The Formula One rule is that the car must be driven at
a weight of 600 kg or more, including the driver, and
considering the weight of the drivers equipments, the
maximum-target weight for the chassis would be
approximately 500 kg. The chassis is made as light as
possible, and then ballast is loaded to bring the total
weight up to 600 kg. As much ballast as possible is
loaded, and it is placed in as low a position as possible.
This lowers the center of gravity and allows adjustment
of the weight distribution, and explains why the amount
of ballast is directly tied to chassis performance. By the
end of the season, the BAR002 had almost no ballast
loaded.
Another issue was the 42% weight distribution. It is
not known why this weight distribution became a starting
specification. Considering that the RA099 had a figure
of 45% using the same Bridgestone tires, this means
there was a difference of as much as 3%. A weight
distribution that is skewed to the rear is advantageous
for acceleration but becomes unstable in the deceleration
range. To control that, the rear wing is turned up to
increase the downforce. The higher up the rear wing is
turned, the more the aerodynamic efficiency is degraded
and the drag increases. It is also necessary to suppress
fluctuations in the aerodynamics in order to limit the
vehicles instability, and there was no choice but to
stiffen the springs in order to limit fluctuations in vehicle
height. The actual spring wheel rate was 25 kg/mm,
which was stiffer than the vertical spring rate (24 kg/
mm) of the tires themselves. In the lower speed range,
therefore, the vehicle would bounce and lose its grip.
Having the weight distribution skewed to the rear also
placed a burden on the driving wheels and caused the
tires to rapidly deteriorate. When the situation
progressed, the tires would generate excessive heat,
causing blisters to form.
The way that races developed for the BAR002 was
to do a rocket start and pass two or three cars before
the first corner. After that, the rear tires would
degenerate, and the car would drag back and lose its
position. This pattern was repeated. The weight
distribution was gradually enhanced toward the final
races of the season, reaching 44%. In conjunction with
this, the distribution of the aerodynamic downforce
exerted on the front wheels (hereafter the aero balance)
changed from 33% to 38%, but the aerodynamic
efficiency ended up decreasing. The reason for this was
that the aero balance target value served as a reference
in promoting the development of all the aerodynamic

parts so as to raise the aerodynamic efficiency to a
maximum at the balance point, and making mid-season
changes to the target value therefore would cause a fatal
delay in aerodynamic development. Similarly, the weight
distribution target values that are calculated from the
gripping force of the front and rear tires should also be
decided initially when beginning chassis development.
The reason is that the weight distribution target values
end up determining the dimensions, weight, and
positioning of the monocoque, fuel tank, engine,
transmission, and suspension, so that trying to make
changes later causes the yaw inertia moment and the
height of the center of gravity to increase.
Another notable development topic was electric
power steering (EPS), which was used starting with the
Hungary Grand Prix in August, ahead of the other teams.
No one doubts its effect any more, but when it was
initially developed, people wondered whether power
steering really would reduce the lap time. This was
because the increase in lap time due to the addition of
2.6 kg to the system weight was clear, but the enhanced
stability caused by reduction of the burden on the driver
could not be quantified. At that time, however, a selfalignment effort of 40-50 Nm was considered to be
necessary for the steering on the high-speed corners in
Barcelona and the USA. This means that drivers would
have to exert fine control on a 10-inch diameter steering
wheel while supporting a 20-kg load on it with their
arms extended. The drivers were in a situation where
keeping the steering aligned was all they could manage
without exerting further control, and it was therefore
decided to adopt the EPS. The size of the motor was a
constraint on the EPS, and it was developed to be
capable of a maximum assist of 15 Nm.
The issues faced with the car during that year were
the inability to optimize the height of the center of
gravity and the weight distribution because of the weight
increase caused by reliability upgrades. With the
following year in mind, therefore, weight reduction of
the various parts was given priority in order to properly
establish those parameters.
4. The BAR003: 2001
Making use of lessons learned in the previous fiscal

year, a weight reduction of 33 kg was accomplished and
a weight distribution of 45% achieved. Typical weight
27
5. The BAR004: 2002
reductions are as follow:

(1) Use of CFRP in suspension: 13 kg
(2) Overhaul of gearbox structure: 6.5 kg
(3) Hydraulic systems: 1.5 kg
(4) Thinner walls for tubes in aluminum radiator and
enhancements to fins: 3.0 kg
(5) Optimization of orientation and number of monocoque
laminates as well as of honeycomb thickness using
computer-aided engineering (CAE): 3.0 kg
(6) Achieving a balance of collision performance and weight
reduction by using CAE to optimize nose cone, front
wings, and rear wings: 4.4 kg
(7) EPS motor enhancement to increase assist torque from
15 Nm to 27 Nm while reducing weight: 0.2 kg
(8) Use of titanium for seatbelt buckle: 0.5 kg
These are some of the ways in which the parts were
overhauled and refined.
The aerodynamic performance achieved the lift/drag
ratio (L/D) of 3.0 under the aerodynamic conditions of
a downforce of 12500 N at a speed of 250 km/h, which
was the initial objective when the BAR003 was first
raced. Development was pursued with the aero balance
objective of 42%.
The wheel spring rate came out at 14 kg/mm, finally
bringing the cars basic characteristics and
specifications to appropriate values, and making this a
year in which development could raise the car to a
whole new level. The weaknesses of the BAR003 were
its instability from braking up to the point when the
steering wheel started to be turned, understeering (US)
at the cornering apex, and the kind of sudden
oversteering (OS) called snap-oversteering that occurs
when the accelerator is pressed and the car is building
up speed coming out of a corner. In other words, the
steering characteristics during cornering were
inconsistent, requiring constant corrective steering. This
was addressed by altering the control in the
electronically controlled differential to increase stability
by directly linking the left and right wheels when the
brakes were applied, then freeing the wheels after that,
which was effective in resolving US in the middle of
a curve as well as snap-OS when the car was coming
back up to speed. It was necessary to develop optimal
control, and even before that it was necessary to
enhance the basic characteristics of the suspension. In
other words, the camber angle-change characteristic
(hereafter suspension compliance steer) and the toe
angle obtained by input from the tires became the toe
out when lateral force was input to the rear suspension.
Measurements made after the season showed that these
were in the direction of instability. Carrying out
measurements with a suspension tester reveals these
clearly, and it is taken as the accepted wisdom that new
cars should ordinarily be tested once they have been
built and before they are driven. Under pressure from
the testing and racing schedules, however, there was no
choice but to give priority to circuit testing over bench
testing, so that practice could not be implemented. As
result, however, the discovery of an essential issue was
delayed, and the development ended up taking a detour.
This car represented a further evolution with a weight

reduction of as much as 14 kg, as well as increased
stability achieved by having control play a larger role.
While traction control during acceleration was a matter
of course, over-run control was adopted to transfer
torque from the engine when the rear wheels were
detected as locked during braking. Having the rear
wheels lock during braking is ordinarily very likely to
result in a spin. In order to avoid this, the brake force
distribution to the front and rear brakes must allow an
extra margin of distribution to the front, but this overrun control now allows the rear-wheel brakes to be
applied even closer to their limit.
Efforts were also made in this year to enhance the
analysis of maneuverability. One rather basic matter was
to enable quantitative evaluation of driver comments
about the degree of OS and US from the chassis data.
Tools for evaluating stability by means of simulations
that take suspension compliance steer characteristics into
account were also adopted. All of these originated in
chassis performance evaluation technology for massproduction vehicles.
The weight reduction measures included the
development of materials for disks and pads in carbon
disk brakes. Friction material made with carbon is
subject to progressive wear when they are oxidized due
to heat, so they are made thick enough to last through
the running distance of a race. Curbing that oxidation
and thereby reducing the amount of wear enabled
reductions from 24 mm to 20 mm in the thickness of
disks and from 18 mm to 15 mm in pads made to
Suzuka Grand Prix specifications. This also led to
reduction in the total width of the calipers themselves,
achieving a total weight reduction of 1.6 kg in the
system. Weight reduction was pursued by going so far
as to implement proprietary development of parts that
ordinarily must be purchased from the brake
manufacturers. Efforts were also begun to reduce the
weight of the materials employed in many different
CFRP applications in the chassis. Carbon laminates are
made up of carbon fiber and resin, and in order to raise
the fiber volume fraction (VF), extra-thin prepregs were
applied. These were made from carbon fibers that had
been opened up and crushed and then lined up side-byside without any gaps between them. Their use in the
engine cover achieved a weight reduction of 3.1 kg.
28
New attempts were also made regarding chassis

stiffness. Air intake to the engine passes through holes
in the roll hoop above the drivers head to the air box
where it is conducted to each cylinder. The air box was
treated as more than just an air vent, however, by
making it a structural member that fastens the top of the
engine to the monocoque in order to provide a dramatic
increase in chassis stiffness. This measure enhances
torsional stiffness 8%, vertical bending stiffness 55%,
and lateral bending stiffness 6%. When tested in actual
driving for comparison with and without this air box,
however, the results showed no difference, and the
regular air box was returned to use partway through the
season. This led to the formulation of a rule of thumb
that driving stability is insensitive to chassis bending
stiffness in the vertical direction, and this provided an
excellent reference point for implementation of weight
reduction the following year. Advances were made in
weight reduction and control development during this
year. Although the car was balanced, however, it was
slow. The balance that had formed was still at too low
a level, and it would be necessary to upgrade
performance in all directions.
making the BAR005 the lowest of all the third-era

Formula One cars. Although further measures to reduce
weight were also pursued in subsequent models, changes
in the collision safety regulations turned the focus toward
how to reduce the resulting weight increases.
Consequently, the year-to-year changes in the height of
the center of gravity tended to show either leveling off
or slight increases.
With this machine, the basic elements of chassis
performance, such as the height of the center of gravity,
the weight distribution, and the moment of inertia, as
well as the aerodynamic performance, reached the
highest level standard. Even though the BAR005 showed
signs of being fast during winter testing before the
season, weight reduction had resulted in diminished
reliability in the gearbox and hydraulic systems, so that
Jacques Villeneuve retired eight times out of 16 races
while Jenson Button retired five times. The car ended
up with fifth place in the constructors standings. This
was the result of pressing hard in development, and
taking drastically aggressive measures, in order to catch
up with and overtake the top team, and the reliability of
the points in question was steadily upgraded in the
following year.
6. The BAR005: 2003

7. The BAR006: 2004
During the previous fiscal year, every element had

been reviewed from the ground up under the guidance
of technical director Geoffrey Willis, who had moved
over from Williams. Dramatic progress was made in the
quality control of carbon parts and the reliability of
bonded parts in particular. Faults in carbon parts
occurred much less frequently after that. Rigorous quality
control of other parts was implemented, as well.
Measures to reduce weight also extended beyond
simple weight reduction. The effort to reduce weight
above the rear axle, to take a particular example,
expedited selective weight reduction in the chassis rear
end in order to lower the rear center of gravity, making
the gearbox lighter and more compact and the engine 16
kg lighter, which was expected to promote stability.
These measures had a significant effect in reducing the
yaw inertia moment, and even though the wheel-base
was extended by 61 mm, from 3133 mm to 3194 mm,
the yaw inertia moment remained largely the same at 688
kgm2. The overall height of the center of gravity was
lowered from 243.1 mm above the floor of the previous
fiscal years BAR004 to 226.6 mm above the floor,
Of all the third-era Formula One cars, this was the

most competitive machine throughout its season. Of the
18 races, Jenson Button finished with points in 15 races
and went to the victory stand 10 times, while Takuma
Sato finished with points in 10 races and went to the
victory stand once, with the BAR racing team placing
second in the constructors standings. The reasons for
this breakthrough were the way in which the BAR005,
which had great potential, matured through evolution
along normal lines to secure its reliability, and the
change in evaluation methods used in aerodynamic
development. Target values were set for the variation
range so that the aero balance could be maintained
constant as much as possible during cornering. The aero
balance change from braking to corner exit can increase
by as much as 10% in some cars, and these measures
were part of an attempt to bring that figure down to zero.
The conventional approach was to evaluate aerodynamic
pitch sensitivity by mimicking the changes in vehicle
height at the front and rear during cornering. This was
now augmented with a crosswind mode that subjected
29
the car to a diagonal wind from the front, taking the

actual vehicle slip angle during cornering into
consideration, and also adding a mode in which the tires
are steered. These efforts placed the wind tunnel
evaluation modes on a well-established basis. The
establishment of these aerodynamic evaluation methods
held an important meaning: making the aero balance
stable meant that drivers could attack corners with
confidence, and this is a key to building fast racing cars.
It would be ideal, essentially, if the degree of confidence
with which a driver can attack a corner could be
expressed in quantitative terms, and if it could then be
predicted how many seconds of lap time that would be
equivalent to. This quantification, however, is no easy
matter, so it is necessary to ascertain and evaluate the
causal relationship between the aero balance change and
the drivers confidence.
Advances in methods for analyzing tire characteristics
also enabled a better understanding of how to use
Michelin tires with the weight distribution configured at
48%. This was another major factor in the upgraded
competitiveness of the BAR006.
Another gain in competitiveness came from the front
clutch pack. When the driver begins turning the steering
wheel while braking, the normal load on the inside front
wheel is reduced, and that tire tends to lock up. Once it
has locked, the braking distance increases, and not only
is lap time lost, but flat spots form on the tire surface,
vibrations are generated, and subsequent driving is
impeded. The front clutch pack is a system that avoids
the above by transferring torque from the outside wheel
to the inside wheel in order to prevent locking. It
connects the left and right front wheels by means of the
clutch. The results for this year were achieved because
the basic elements of chassis performance reached their
highest level, reliability was upgraded, and on top of
that, the car possessed another powerful advantage.
Unfortunately, the front clutch pack ended up being
prohibited in the following years regulations.
the rear while keeping the location of the rear axle

unchanged and placing the differential center 5 mm
farther to the front. The wheel-base was 3085 mm, and
the gearbox was changed from a 3-speed to a 7-speed
with seamless shift to reduce the time loss when shifting
gears to zero. As the speed of the cars increased year
after year, the FIA had adopted a policy of placing limits
on it, so that the regulations this year aimed at reducing
the downforce by raising the height of the front wings
50 mm and moving the rear upper wing 150 mm
forward, new regulations were introduced for the diffuser
area, and so on. The FIA was eying a 20% reduction in
the downforce, whereas the Honda team decided to limit
the reduction target to 10%. Although the target was met,
the other cars achieved smaller reductions than
anticipated, so there was a considerable gap at the start
of the season. Every effort was made in the subsequent
development, and a 13% increase in the downforce was
achieved in the course of the season, while the yearround figure had usually been 7%. However, the other
cars achieved similar advances, and the gap was not
closed. The drag had also increased to a maximum, and
enhancement of the L/D remained an issue. The lack of
downforce resulted in excessive deterioration of the rear
tires, which also had an impact on traction and braking
performance. In combination with insufficient warm-up
of the front tires, this caused conspicuous delays due to
OS in high-speed corners and US in low-speed corners.
In the later competitions, push rod on upright was
introduced as a front suspension measure to achieve
better weight-shift distribution, and this provided a better
grip in low-speed corners. In push rod on upright, the
push rod is attached to the upright instead of the lower
arm, where it is conventionally, and also offset from the
kingpin, so as to control the vehicle height and weight
shift during steering. Although this measure yielded its
effect, it did not increase performance enough to
overtake the top team, and this was a year that left
unresolved issues in tire handling.
The fourth competition at the San Marino Grand Prix
would have been a double victory, the first time in the
season for both drivers to earn championship points
(Jenson Button in third place and Takuma Sato in fifth).
The car inspection after the race, however, found a
violation: when all the fuel in the fuel collector tank was
removed, the cars came in under the regulation minimum
weight (600 kg). The team responded by disclosing their
data and insisting as follows: the fuel tank is not a
special item, and the car did not go under the regulation
minimum weight of 600 kg under any conditions
whatsoever while running in the race. The FIA
responded that data can be altered and cannot be
recognized as evidence. They took the view that,
structurally, there was a possibility that the cars fell
below the minimum weight, and they deemed the cars
to have infringed the regulations. The race results were
forfeited, the team was suspended from entering the fifth
and sixth competitions the Spain and Monaco Grand
Prix and the team was suspended from racing for six
months, with a probation period of one year. This year,
8. The BAR007: 2005
Measures were taken with the aim of further

increasing the range of adjustment of the weight
distribution toward the front by 1% (to a maximum of
48.9%) compared to the previous years BAR006, and
to reduce the yaw inertia moment (-9%). The result was
a packaging layout that moved the front axle 46 mm to
30
therefore, was also one in which severe rulings were

imposed.
turning performance that exceeded the other cars, but

delays in developing technology to reduce rear wing drag
at high speeds meant that this machine was far
outstripped by the top teams at maximum speed. In the
area of tires, the excessive deterioration of the rear tires
was resolved. However, warm-up became an issue on
certain circuits, making it necessary to further upgrade
tire temperature management. Although it was helped by
the dropout of the top teams in the 13th race, this year
the RA106 did achieve Hondas first win (with Jenson
Button) of the third era of Formula One racing. Various
measures were taken to upgrade the cars
competitiveness even from the 14th race, such as
extending the wheel-base by 50 mm at the front axle to
enhance its braking stability, and these yielded positive
effects, but not enough to put its performance on a par
with that of the top teams.
9. The RA106 (BAR008): 2006
Honda bought up all the BAR shares during this year,

so that it was 100% Honda-owned. The chassis
designated BAR008 was renamed during the
development stage as RA106. Since regulations required
that the engine be changed to a 2.4 L V8, the total
engine length was reduced by 90.5 mm, affecting the
chassis package. In order to maintain stability, the
monocoque and gearbox case alone were lengthened, and
a driveshaft with lightweight constant velocity joints was
adopted. The wheel-base retained the same 3085 mm
length as in the previous year.
With the new engine, the lateral mounting points
were spaced with a larger pitch and the chassis lateral
stiffness was increased 33% to heighten the stability. In
the gearbox, the seamless shift that had been introduced
the previous year was expanded to all the gear shifting
stages. A direct-push clutch was adopted with the aim
of stabilizing the clutch friction performance necessary
to control tire slipping at the start, which came with the
reduced engine displacement, and this measure also
increased acceleration from a standing start.
Aerodynamic measures were taken in light of the
reduced power with the change to a V8 engine, and a
policy change from emphasis on downforce to emphasis
on efficiency aimed to enhance efficiency by 13%. In
order to obtain downforce without increasing the drag,
steps were implemented to increase the rate of flow
across the bottom surface of the floor, and the nose was
raised higher and the keel was eliminated in order to
heighten performance. This setup sought to overcome the
excessive wear of the rear tires that had been an issue
the previous year by taking steps to maximize the
traction performance of the tires, and the front-rear ratio
of the weight-shift distribution, which was caused by the
lateral force generated when building up speed coming
out of a corner, was changed to give it a forward shift.
This allowed the temperature of the front tires to be
brought up to the working range of 70C or higher.
The RA106 (BAR008) produced top-level times in
the early qualification rounds when the season was
beginning, but the car was not producing winning times
in the actual races, and this became a conspicuous issue.
The effects of aerodynamic enhancement brought out
10. The RA107: 2007
Chassis development up to the RA106 had continuity,

and consisted basically of carrying on with the previous
years development while adding refinements. In June
2006, however, technical director Geoffrey Willis left the
team, and for this among other reasons, the RA107
represented a very challenging attempt to discard the
preceding base and build instead from a new concept on
up.
Development of the RA107 was begun with top
priority to aerodynamics as its slogan. The wheel-base,
made longer than in the RA106 by shifting the front axle
30 mm forward, was 3165 mm. In order to minimize the
chassis inertia, the gearbox interior was shortened and
the final gear center was moved approximately 50 mm
forward. As a result, the sweepback angle of the
driveshaft also became larger than in the RA106.
Significant changes were also made in the layout of
the interior where they are not visible from the outside.
For example, the oil and water radiators located on either
side of the driver inside the engine cover were reshaped
to be longer and more slender, slanted forward, and
contained inside the side pontoons. A layout was also
attempted that led the exhaust pipes from the engine
forward from the left and right banks of four cylinders,
gathered them together, then turned them outside at a
point close behind the driver. While the purpose was to
reduce the moment of inertia of the chassis by any
amount possible, the increased vibration and noise from
31
the exhaust were not welcomed by the drivers.

Another major change made this year was the return
to the use of Bridgestone tires, because Michelin, whose
tires had been used for the preceding three years,
withdrew from Formula One racing. The differences in
the deformation of the Michelin and Bridgestone front
tires when running had a greater effect on the
aerodynamics than had been anticipated, and the time it
took to understand it caused delays in development.
The aerodynamics of the RA107 were the first to be
developed making use of the full-scale wind tunnel that
entered operation in 2006. At the start of the design
process, the objective was development to reduce
movement of the aerodynamic center as much as
possible, from the point of braking just before cornering
up to the zone of acceleration at the exit. When testing
began with the actual car, however, the amount of
movement was greater than anticipated, counter to the
plan, so that the chassis behavior showed unstable
characteristics from the point of braking throughout the
entire corner. Furthermore, the response emphasized
downforce at the sacrifice of aerodynamic resistance,
with the result that the cars straight-line speed fell short
of the top teams by 10 km/h or more. A number of
changes were made partway through the season in order
to enhance the aerodynamic characteristics, prominent
among them a change in the shape of the side pods, and
the aerodynamic performance did gradually enhance. The
other teams, however, also advanced with similar
rapidity, and the RA107 did not achieve superiority in
performance.
The result was that the team did not mount the
victory stand even once throughout the entire season,
they earned points three times, and were eighth in the
constructors standings. Even in the Drivers Titles,
Button was 15th and Barrichello was 20th. These were
the poorest results since the start of the third era.
introduction of the concept of usable downforce as an

index for aerodynamic development. In this procedure,
wind-tunnel test results conducted under various
conditions, including front and rear vehicle heights, are
multiplied by weighting factors according to importance.
These are then integrated to obtain a single scalar value
by which analysis is carried out. Setting up the factors
requires expertise, but the method expresses performance
using a single numerical value, and it has the advantage
that, when further refined, it can be used to keep the
development orientation in clear focus. Target usable
downforce values were set for each update from the
opening race and the major Grand Prix competitions, and
development was carried out so as to realize those
targets.
The monocoque was made more slender and longer
(by 45 mm) to the external appearance, and the side pod
locations were also moved toward the rear, largely due
to aerodynamic requirements. As a result, the rear axle
was shifted rearward while the sweepback angle of the
driveshaft was maintained unchanged, and the wheelbase was increased 45 mm over the previous years to
3210 mm. The nose was also raised and the bottom of
the side pods narrowed, resulting in a chassis shape that
takes the flow over the chassis bottom into consideration.
However, development that placed top priority on
aerodynamics also had negative effects in some areas.
For example, the overall width of the gearbox was
narrowed out of concern for the flow along the side to
the rear wing at the rear of the car. In the rear
suspension, this caused a diminishment in the mounting
stiffness and deterioration of the damper ratio. Traction
control was forbidden starting in this year, and it
emerged as an issue of driveability that the engines lowspeed torque did not increase smoothly when the throttle
was applied. The RA108 was troubled throughout the
season with unstable cornering and inadequate traction.
During the last half of the season, in particular, graining
of the tires (in which just the tire surface peels off as
rubber grains, and these are rounded into roller-like
shapes that cause the tire to lose its grip) caused by lowtemperature road surfaces also became an issue.
In order to enhance the damper ratio and the stiffness
of the rear suspension, the damper was changed to a
rotary type. This enhanced suspension was adopted
starting with the 11th race the Hungary Grand Prix
but it did not amount to a fundamental enhancement.
Another factor was that Ross Brawn, who joined as
Team Principal in November 2007, adopted the policy
of concentrating development resources on the 2009 car.
Mid-season upgrades and enhancements to the RA108,
therefore, tended to be omitted if they were not items
that could be treated as advance tests for 2009.
As a result, the car scored points only five times, and
the team came in eighth in the constructors standings
(McLaren was divested of its title due to a spying
incident, so the actual ranking would have been ninth)
with Rubens Barrichello the only third-place winner to
mount the victory stand in the ninth competition, at
Silverstone in the rain, when the pit strategy was
11. The RA108: 2008
One of the major causes of the lukewarm results

achieved by the RA107 was its aerodynamic
performance. Therefore, the RA108 also adopted the
slogan of top priority to aerodynamics and
development was begun with the goal of pursuing
enhancement of L/D (20% better than the previous year)
as well as yaw and pitch sensitivity reduction.
One major change in the development was the
32
successful. The driver rankings had Rubens Barrichello

at 14th and Jenson Button at 18th place, making this
another year of low rankings to follow those of 2007.
by each individual factor, as well as the relative merits

of different combinations of factors in terms of total
performance. In the latter part of the third era, efforts
were therefore concentrated on understanding of the gaps
that exist between bench test results and simulations on
the one hand and the actual machine on the other, as
well as establishing methods of evaluation. Step-by-step
tests and simulations had to be constantly repeated in
order to obtain accurate tools and judgment criteria.
Once these are in hand, a team will not lose sight of the
direction of development, no matter how the regulations
may change. Furthermore, any reduction in the necessity
for tests using the actual machines enables a reduction
in development costs, as well. As people increasingly
call for the cost of Formula One development to be
lowered, this style of development will no doubt
continue to accelerate.
The movement to reduce development costs and to
perform evaluation of cars as much as possible on the
desktop and using bench tests is by no means limited to
Formula One development. This is clearly a major trend
in the development of mass-production vehicles, as well.
It is a matter of firm belief, therefore, that the knowledge
acquired during this 10-year period of development can
definitely be used to advantage in the development of
mass-production vehicles. All of that knowledge cannot
be related in the pages of the Technical Review, but it
will be fortunate indeed if some hints of it can be
conveyed herein.
12. Conclusion
The decade-long flow of Formula One chassis
development in the third era can be generally
summarized as follows. The years from 1998 to 1999
were the period of initial involvement with Formula One
racing in recent years, with the fundamentals being
learned from Dr. Postlethwaite and the Dallara engineers.
The years from 2000 to 2001 were a time of working
to get the cars at least up to a certain level by reducing
the weight in every part in order to reduce the height of
the center of gravity. The other main focus was assuring
reliability. In 2002, an attempt was made to take a step
forward from the fundamentals and address such issues
as suspension stiffness and body stiffness. However, the
vehicles that were the base for this effort had individual
unit variations because of the manufacturing quality, the
driving data evaluation methods and related matters did
not evolve to match the effort, and the development
stalled. The year 2003 brought advances in the
manufacturing quality of CFRP and all other parts, as
well as a new ability to engage in design while making
limited predictions of how it would affect the
maneuverability of the overall chassis package, including
the engine. This became the year of transition from
development of constituent technologies to development
with a view to the balance of a single whole vehicle. The
year 2004 represented a maturation and an evolution
along normal lines from 2003, with new aerodynamic
evaluation methods proving their worth. The car reached
the top level. The second place in the constructors
ranking earned in this year, however, turned out to be a
peak. In 2005 and thereafter, efforts were made to
overcome the differences between phenomena in the
actual machine and testing in a wind tunnel, and to
determine techniques for obtaining the optimal
performance from the tires, but these core technologies
proved elusive, and the car slipped back in the rankings.
It became apparent in 2007 that the regulations on
aerodynamics would be changed significantly in 2009.
This was taken as an opportunity to recover from a
moribund situation, and work was begun on development
of 2009 specification aerodynamics. The year 2008
turned out to be the final year of the third era. Although
the car finished eighth in the rankings, the preparations
for 2009 were steadily taking shape, and after Hondas
withdrawal, that technology was inherited by the Brawn
GP Formula One team and contributed to their finish in
the two top places in the opening race.
Formula One development in recent years has been
a puzzle about how to bring mutually antagonistic factors
that govern car performance into balance with each other
so that the level of the car as a whole can be raised. In
order to achieve a high-level balance among complexly
interrelated factors, it is necessary to evolve technology
that can accurately grasp the degree of contribution made
Author
Akio TONOMURA
33
Shun KAWABE
Nobuhisa NAKAYOKU
Development of Materials
during Third Formula One Era
Masami HOSHI*
ABSTRACT
The realization of high-power and high-efficiency powertrains, lightweight and compact vehicle frameworks, and
a high level of durability and reliability are necessary to increasing the competitiveness of Formula One race vehicles,
and material technologies are an essential factor in this. Honda has developed high-strength, high-stiffness, and lowdensity materials such as titanium-aluminum and aluminum metal matrix composite (MMC) materials in order to achieve
weight-savings in structural parts. The company has also developed low-friction and high-durability surface modification
technologies, in particular DLC coatings, for friction materials. For use in hybrid systems, Honda has developed
magnetic materials for use in motors and highly functional materials such as high-thermal-conductivity materials. These
material technologies have enhanced vehicle performance and increased race competitiveness.
1. Development of Materials during Honda

Formula One Third Era
out and introduced not only domestic but also European

and U.S. technologies, and proceeded with development
programs aiming to create materials technologies
surpassing those of other teams. While a later change in
circumstances reduced the need for precision casting
technologies, in the area of structural steels the company
developed high-strength gear materials and hightoughness steels for use in cranks, and began to employ
them in races. Later, the practical use of high-strength,
high-rigidity materials such as titanium and aluminum
metal matrix composite (MMC) materials became
feasible, and Formula One teams entered a stage of
fierce development efforts. Although research efforts in
these areas had diminished in Japan, these were
fundamental technologies for the aerospace industry, and
research was ongoing in Europe and the U.S. Honda also
made concerted efforts at development, and established
its own processes for manufacturing high-strength
materials. This resulted in the development of titaniumaluminum valves and piston pins and aluminum MMC
pistons, which were subsequently employed in race
vehicles.
In 2002, Honda made the transition from
conventional bucket tappet valvetrains to rocker arm
valvetrains, which enabled the achievement of increased
engine speed and enhanced performance. Tribological
technologies for the friction components formed an
important aspect of these development efforts, and
surface modification technologies were an essential
research focus which would affect the feasibility of the
new mechanisms. As in the case of structural materials,
Honda sought out technologies, including technologies
developed overseas, and developed a diamond-like
The role of Hondas Materials Division in the

companys Formula One activities during the 1980s
centered on the development of technologies for the
enhancement of materials that would enhance engine
performance. In the process of development of race
vehicles for the companys third era activities
commencing in 2000, that role evolved to the
development of advanced materials of the type used in
the aerospace industry and their application throughout
the vehicle, from the powertrain onwards. This process
involved rapid changes of a type not previously
experienced, and the Formula One third era became a
demanding period in which the Formula One team faced
a variety of technological development challenges.
At the beginning stages of Hondas third era
activities, the European and U.S. teams had already
established a technological foundation in advanced
technologies developed in the aerospace industry, and
were striving to apply these technologies in Formula
One. Having ceased its Formula One activities, Honda
was delayed in its own efforts in this direction, and the
company fell behind its competitors in terms of
technological advancement. Precision casting
technologies for titanium and steel materials for gears
and cranks can be pointed to as representative examples.
Because of this technological lag, the enhancement of
Hondas development program to enable the company to
quickly catch up with and overtake the leaders in the
field, the European teams, was an important issue at the
time. Given this, the Materials Division actively sought
34
carbon (DLC) coating as a new surface modification

technology. This technology significantly increased the
durability of the friction components of the camshaft and
rocker arms of the new mechanism, and contributed to
making it technologically feasible.
Moving in another direction to the development
activities outlined above, discussions intensified among
the Federation Internationale de lAutomobile (FIA) and
participating teams regarding the reduction of the
excessive development budgets resulting from
competition for technological development, and from
2006 materials regulations were enhanced.
In order to minimize the possibility for losses from
having technologies that had already been developed and
put to use rendered ineligible for use, Honda continued
development efforts situated within the scope of the
materials regulations. As a result, new ideas emerged,
and Honda proceeded with the development of conrod
bearings using a high-thermal-conductivity material as
the back metal and technologies for the enhancement of
the performance of the DLC coating. These technologies
increased powertrain efficiency. In particular, in the area
of DLC coatings, a technology resistant to high surface
pressures was developed enabling the previous limit in
surface pressure to be exceeded. The application of this
technology from the shift to the bevel gears of the
transmission increased Hondas race competitiveness. In
the period from 2007, the final period of the companys
third era activities, Honda also focused on development
efforts in the area of functional materials for the

development of hybrid systems, the use of which
changes in regulations had made possible from 2009
onwards. Aiming to reconcile hybrid device performance,
the framework structure, and durability, the company
conducted development efforts in the areas of magnetic
materials for use in motors and high-thermal-conductivity
materials to enhance the cooling performance of power
control units, and succeeded in developing new materials
technologies.
2. Direction of Materials Development

This chapter will discuss the direction of development
of materials technologies for race vehicles as a whole,
including the technologies discussed above. The elements
composing race vehicles are engines, transmissions,
chassis, and hybrid systems, and as Fig. 1 shows, there
are four basic directions for the development of materials
enabling the realization of the necessary characteristics
of each element. It is essential to maximize the functions
of these basic materials and thus enhance the
competitiveness of the race vehicle.
High-strength, high-stiffness, and low-specific-gravity
materials, both metallic and non-metallic, are a key
technology in the achievement of improved performance
and durability in race vehicles, from the engine and
transmission components and the framework structures
to the suspension and the brakes.
Performance requirements
Engine
High engine speed

Reciprocating mass reduction
Fine valve train motion
Enhanced tribological durability
High power /
High thermal efficiency
Material development direction

High-strength materials
High-modulus materials
Combustion efficiency
Enhanced intake and exhaust efficiency
Low friction
Mechanical loss reduction
Pumping loss reduction
Oil agitation loss reduction
Lightweight / Compact
Low-density materials
Low-friction materials
High wear resistance materials
Lightweight frame structure

Lightweight & compact cylinder head/block
Lightweight & compact intake and exhaust system
Compact gear train and auxiliary system
Narrow gears
Low pump capacity
Lightweight and
complicated (hollow)
structural technologies
Low friction
Transmission
High transmission efficiency /
Lightweight and compact
Narrow gears
Thermal management materials
Enhanced durability of high-efficiency mechanism
Chassis
Lightweight / High stiffness /
Vehicle dynamics
Lightweight and stiffer body structure

Lightweight and stiffer suspension system
Functional materials
High-efficiency magnetic materials
Hybrid System (KERS*)

High Power /
Lightweight and compact /
Thermal management
Effective cooling technologies

* KERS: Kinetic Energy Recirculation System
High-power motor
High-efficiency, lightweight, small PCU
High-capacity, lightweight, small battery
Fig. 1
F1 material development
35
Battery materials
PDU (Power Delivery Unit)
Electrical assembly technologies
Development of Materials during Third Formula One Era
Low-friction and low-wear materials are an important

technology directly related to the achievement of
enhanced vehicle efficiency by reducing friction
resistance in the engine and in friction components
throughout the vehicle and enabling components to be
used under extremely severe conditions.
Thermal management material technologies, which
enable control of the heat balance, are important in
enhancing the functions of each part of the vehicle,
including the enhancement of thermal efficiency around
the engine and brakes and the cooling of the hybrid
system. This is achieved, for example, by means of
exploitation of the thermal conductivity of base materials
and the heat absorption and dissipation characteristics of
surfaces.
In the area of functional materials, magnetic materials
for use in fuel injectors are necessary in engines, and
magnetic materials, materials with excellent thermal
conductivity and thermal insulation characteristics,
battery materials and others are important technologies
in enhancing the functions of the motors, PCUs, and
energy storage devices used in hybrid systems.
(6) Murai, Y., Tsuchiya, M.: Development of Lightweight

and Low-friction Resin Materials for Oil Pump Rotor,
Honda R&D Technical Review 2009, F1 Special (The
Third Era Activities), p. 269-270
(7) Murai, Y.: Research on Coolant Using Nano Additives
and Heavy Water, Honda R&D Technical Review 2009,
F1 Special (The Third Era Activities), p. 271-272
(8) Murai, Y., Okumura, M.: Development of High
Thermal Conductivity Material for Heatspreader,
(9) Tsuchiya, M., Tsunashima, S., Hashimoto, H.:
Development of Titanium Aluminide Valve, Honda
R&D Technical Review 2009, F1 Special (The Third
Era Activities), p. 254-256
(10) Ito, N., Yonehara, M., Yakubo, K.: Development of
DLC Coating on Camshaft and Rocker Arm, Honda
(11) Miyoshi, T.: Development of High Induction Stator
Core, Honda R&D Technical Review 2009, F1 Special
(The Third Era Activities), p. 280-281
(12) Ishizuka, K., Kusakai, K., Imai, N.: Development of
Metal Matrix Composite Piston, Honda R&D Technical
Review 2009, F1 Special (The Third Era Activities), p.
246-247
(13) Minato, S., Onaha, Y.: Development of Lightweight
Titanium-aluminide Piston Pin, Honda R&D Technical
248-250
(14) Minato, S., Mizogawa, K.: Development of Hollow
Titanium Connecting Rod, Honda R&D Technical
251-253
(15) Imai, N., Nishida, H.: Development of Laser Clad
Welded Valve Seat, Honda R&D Technical Review
2009, F1 Special (The Third Era Activities), p. 263-265
(16) Tanahashi, T., Onaha, Y.: Development of DLC for
Transmission Gears, Honda R&D Technical Review
(17) Yamada, Y.: Weight Reduction Study for Side Panel
of Formula One Monocoque, Honda R&D Technical
282-285
(18) Yanase, A., Yamada, H.: Development of Brake Caliper
Production Process with High Strength Al-Li Material,
3. Conclusion
For examples and more detailed discussion of
Hondas work in the development of materials
technologies, the reader is referred to the special edition
of the Technical Review dealing with materials(1)-(18).
While few of these materials are employed in massproduction vehicles due to cost, the knowhow and
guidelines in the area of element technologies gained
from their development do encompass aspects that can
be extensively used in development for mass production.
In conclusion, the author wishes to express his sincere
gratitude to Hondas suppliers, both domestic and
international, for their generous assistance in these
development projects.
References
(1) Ito, K., Asami, M., Koinuma, H.: High-performance
Shell Bearing from New Material, Honda R&D
Technical Review 2009, F1 Special (The Third Era
Activities), p. 240-242
(2) Mizoue, K., Kawahito, Y., Mizogawa, K.: Development
of Hollow Crankshaft, Honda R&D Technical Review
(3) Munemura, T., Yamada, H., Ohnuma, T.: Development
of Titanium Exhaust Pipe, Honda R&D Technical
266-268
(4) Konagaya, D., Tanahashi, T., Tanaka, T.: Development
of High-performance Gear Materials, Honda R&D
(5) Sanada, T.: Development of Titanium Hollow Valve
and Study of Sodium-potassium Valve, Honda R&D
Author
Masami HOSHI
36
Descriptions of Engine Technologies
Explanation of Hondas Third Era Formula One

Engine Development
Kiyoshi KONDO*
Hajime ENDO*
Yasuhiro MOTOHASHI*
Satoshi NAKAMURA*
Kazushi OGIYAMA*
Nagao YANAGISAWA*
ABSTRACT
During the third era Formula One engine development, as a measure for increasing the engine power through an
increase of engine speed, the enhancement of the induction static efficiency and weight reduction of the moving parts
was focused on. From 2007, restrictions on the maximum engine speed and development of the engine except for the
induction and exhaust system were introduced. Therefore, the focus of the development changed to practical use of
dynamic effect of induction and exhaust pressure. Moreover, the driver-aids including traction control was banned in
2008, so attention moved to the development with the objective of shortening lap times, and including drivability
enhancements.
Also, to comply with new regulations concerning the longer use of engines, the restriction of use of materials, and
to respond to changes in actual usage in circuit running in accordance with the change of regulations, the subdivision
of the endurance mode and results feedback was performed, which achieved durable reliability in parallel with
performance enhancement.
This document describes the evolution of the engine as outlined above, and includes descriptions of technology
such as the lap time evaluation method and tools.
1. Introduction
induction efficiency was focused on.

Because friction and vibration increase quadratically
with the rise in engine speed, attention was also given
to technology that reduces mechanical loss, and to
dealing with a vibration acceleration that reached 300 G.
Engine failure on the circuit may cause immediate
retirement from the race, so high reliability was also
required. The load on the engine was estimated for each
circuit from the throttle opening angles and the duration
of continuous wide open throttle (WOT), a driving mode
suitable for the load was formulated, and then an
endurance test on a test bed was performed.
From this, operating conditions for the race such as
the temperature of oil and water and the individual
ignition timing for each cylinder were established in
advance, and measures were taken to help ensure high
performance and reliability, and make one engine last for
two races.
In 2007, the FIA restricted the maximum engine
speed to 19000 rpm, and at the same time enforced a
freezing of development of the main engine parts
(homologation).
After that, attention moved to output characteristics
that would shorten lap times, concentrating on the
development of induction and exhaust system parts.
Simulation tools for lap times and fuel efficiency were
Hondas third era Formula One activities began in

2000, with Honda competing with the other constructors
to develop engines with higher engine speed and output.
In 2006, immediately before the International
Automobile Federation (FIA) introduced the engine
speed restrictions, some competitors had achieved a
maximum engine speed of 20000 rpm. At that time, the
maximum engine speed that Honda had reached in a race
was 19600 rpm.
In order to competitively increase engine speed,
technology was required for reducing the reciprocating
mass of the valve train system and the reciprocating
system. To achieve this, in addition to developing
materials and manufacturing technology, predictive
computer aided engineering (CAE) was introduced for
each part. In particular, a system had been firmly
established by 2003 for designers themselves to produce
3D models and perform CAE analysis, enabling design
that is optimized for structural strength, resonance, flow
rate, temperature distribution, thermal stress and other
factors to be performed in a short period of time.
Induction potential was also required to increase engine
power. To achieve this, a higher peak power engine
speed was required, and technology that increases
38
also developed for simulating circuit running conditions

on dyno.
In 2008, the use of a common ECU was made
compulsory and traction control was banned. This gave
rise to issues of drivability when coming out of corners.
Work was started on enhancing drivability by modifying
the engine hardware, but as of 2009, this development
issue remains.
The valve diameter size was decided from a good

balance with the cylinder bore diameter, which was
determined by also taking the piston speed into account.
The bore diameter started at 95 mm in 2000, but by
2002 it had been increased to 97 mm. In 2004, laserclad welded valve seats were used in order to increase
the valve diameter.
From 2000 to 2001, the tappet type was used for the
valve train system, but because the valve lift was limited by
the lifter bore, in 2002, the valve train system was changed
to the rocker arm type in order to increase valve lift.
Cylinder head port development continued throughout
the season, and over 10 types were evaluated. However,
just because the induction performance was high in the
steady flow test, this did not mean that the power would
increase, and so development for enhancing the
combustion, including fuel distribution, was required.
Immediately before the engine homologation in 2006,
power had been increased by developing head port shape
that maximized the dynamic effects, using not only the
steady induction performance, but also 3D-2D coupled
v simulation.
2. Power
2.1. Era of High Engine Speed and High Output (2000
to 2006)
2.1.1 Increased engine speed and enhanced induction
performance
If Pse: Shaft output, Pmi: Mean effective pressure,
Vs: Stroke displacement, Ne: Engine speed, Psf: Friction,
then: Pse = (Pmi Vs Ne/2) Psf.
Thus, an effective way to increase the power is to
raise the engine speed, and Honda competed with the
other teams in increasing the engine speed and power.
Figure 1 shows the trend for the maximum engine
speed.
Hondas V12 engines in 1992 had an engine speed
of only 14400 rpm, but by the time the third era started
in 2000, the level was 17000 rpm. Although the
introduction of mileage extension regulations to make
one engine last over one race event, and then two race
events, slowed this growth, by 2006 it had still reached
20000 rpm in tests on dyno. However, for reasons of lap
time contribution and engine durable reliability, an upper
limit of 19600 rpm was set for the actual races.
When the target of the maximum power and engine
speed are fixed, the power peak engine speed is
determined from maximum engine speed, and the output
value can be estimated as shown in Fig. 2. This means
that the target output shortfall should be compensated for
by reducing the friction and enhancing the combustion.
The induction performance target for the power peak
engine speed was decided using a steady flow test, and
development was performed of the cam profile and port
layout, including the valve diameter, that was required
for achieving the desired horsepower by controlling
engine speed.
3.0L V10
2nd era
3rd era
2.1.2. Making moving parts more lightweight for

increasing the engine speed
With the increased engine speed, moving parts
needed to be made more lightweight from the viewpoint
of reducing friction and durable reliability. To make
these parts more lightweight, materials with high specific
stiffness and high-temperature fatigue strength were
used, and their shapes were optimized using CAE.
In 2004, an aluminum Metal Matrix Composite
(MMC) material with increased high-temperature fatigue
strength was used for the piston, reducing it to a weight
of 210 g at a size of 97.
After the engine mileage extension and the banning
of MMC materials in 2006, the piston weight increased,
but through CAE thermal stress analysis and with the
enhancement of piston cooling using oil jets, a weight
of 230 g was achieved using A2618 enhanced materials.
At the same time, reducing the shaft diameter was
studied to minimize heat generation in the connrod
bearing in the reciprocating system, and to achieve a
good balance with stiffness, it was reduced to 34.
Enhancements for durable reliability were also performed
for the bearing itself, with a copper alloy with high
thermal conductivity being used from 2005.
2.4L V8
2600 rpm
Friction loss
and combustion
enhancement
Power (kW)
V12
V10
Mileage guarantee
2008
2007
2006
2005
2004
2003
420 km 800 km 1500 km1350 km
2002
2001
2000
1992
1991
1990
400 km
500 rpm
Power peak engine speed (rpm)
Year
Fig. 1
Friction loss
enhancement
20 kW
Engine
speed
limitation
2000 rpm
Engine speed (rpm)
3.5L
Transition of max engine speed
Fig. 2
39
V10 F1 engine power peak engie speed - power
The factors that were important in reducing the

reciprocating mass of the valve train system were the use
of rocker arms for the mechanism and the use of
Titanium-Aluminide (TiAl) as the valve material. The
reciprocating mass was reduced from 71.4 g in 2000 to
47.2 g. Even after the regulations on intermetallic
compounds that were introduced in 2006, an equivalent
reciprocating mass could be achieved with a titanium
alloy using CAE analysis.
Figure 3 shows the positioning of the specific
stiffness for the materials that were used in races. From
the perspective of cost reduction, the specific stiffness
was reduced to less than 40 GPa/g/cm3 in 2003, and in
2006, the use of MMC materials, intermetallic
compounds, and magnesium alloys was banned.
Regulations were also introduced for the main elements
of the piston materials, allowing only conventional
materials to be used.
was reduced. These measures reduced the valve drive

system friction of 35 kW in the V10 engine to 12.8 kW
by 2005.
Further studies were performed for reducing the
agitation and shear resistance of oil, the sliding friction
of the journal and reciprocating system, and the ancillary
drive loss.
2.2. Power Curve Characteristics for Shortening Lap
Times (from 2007)
2.2.1. Lap time simulation
In 2007, the development of new systems for the
engine except for the induction and exhaust systems was
frozen in order to reduce costs, and the parts that could
be developed were limited to the air box, exhaust, and
fuel systems.
After the maximum engine speed restriction was
introduced, to maximize the teams competitive strength,
the use of the maximum engine speed was extended to
the 7th gear end of straight. Figure 5 shows the changes
in engine speed for each gear level, while Fig. 6 shows
2 deg
Main combustion period (deg)
2.1.3. Combustion
In Formula One engines that require higher engine
speed and power, increasing the bore diameter is an
effective method of enhancing the induction and exhaust
performance.
However, increasing the bore diameter also increases
the combustion duration, which leads to decreased
thermal efficiency, power, resistance to misfiring, and the
like. As a result, measures for making combustion faster
were required. In the third era Formula One engine, the
bore diameter was increased to 97, and combustion
was enhanced to raise the compression ratio to 13.0. In
2003, a compound angle was added to the inlet valve to
enhance the in-cylinder flow and make the combustion
faster. Figure 4 shows the effects of shortening the main
combustion duration.
Compound valve
angle
1000 rpm
Engine speed (rpm)
Fig. 4
Engine speed (rpm)
2nd 3rd 4th 5th
6th
7th
Up to 2006
From 2007
Fig. 5
Bolt
Change of engine speed usage

Ne frequency at Monza
180
Crankshaft
Valve
160
140
600
Conventional IM
25 GPa/g/cm3
2006 regulation
500
120
100
Time (sec)
Modulus of elasticity (GPa)
200
1st
19000
240
40 GPa/g/cm3
2001 regulation
Effect of compound inlet valve angle
Gear
2.1.4. Friction
When reducing the engine friction, a factor that had
a particular effect was the friction in the valve train
system. The behavior of the valves was enhanced by
reducing the reciprocating mass and minimizing the
angular velocity fluctuations, which reduced the spring
load. In addition, the oil agitation resistance in the
cylinder of the Pneumatic Valve Return System (PVRS)
220
Conventional
Connrod
Piston
Piston
80
Piston
Block, Head
60
2007
2006
400
300
200
Throttle valve, Fuel rail
100
40
Cam cover, Fr cover
20
0
13000
0
0
4
5
6
Density (g/cm3)
10
Fig. 6
Fig. 3
Comparative rigidity of race engine materials
40
14000
15000
16000 17000 18000

Engine speed (rpm)
19000
20000
Ne frequency difference due to maximum

engine speed restriction
the engine speed frequency.

Figure 7 shows a comparison of lap times when the
end of straight engine speed is 18500 rpm and 19000
rpm. It can be seen that at every circuit, the higher the
end of straight engine speed, the shorter the lap times.
Figure 8 shows the effect of the power peak output
engine speed on the lap time and maximum speed at the
Shanghai circuit. It can be seen that the best lap time is
at 18200 rpm, and that the time drops sharply with
engine speeds below this level.
The shift-up engine speed, end of straight engine
speed control (EOS) settings and gear ratios were
optimized for each Formula One circuit to shorten the
lap times.
Also, it became possible to theoretically calculate
how the power curve characteristics would affect the lap
times, and the calculation results were reflected in the
development.
For the Monza circuit, with the power curve
characteristics shown in Fig. 9, even though the power
curve characteristics of Curve B is 12 kW less at the
horsepower peak, the lap time was shortened by 0.04
seconds.
In the development, the lap time calculation was
performed before the race was held in order to optimize
the power curve characteristics.
richer from 11000 to 12000 rpm.

It was thought that this decrease in DR would lead
to a drop in the lap time of about 0.2 seconds.
Regarding these DR issues, for issue (a), spit-back
was reduced by the early closure of the inlet valve and
fuel sticking to the port wall was enhanced by changing
the injector form to a beam. For issue (b), the amount
of residual gas in the cylinder was reduced by tuning the
exhaust, and the torque characteristics were enhanced by
the early closure of the inlet valve. These addressed the
issues of torque following and inconsistent combustion,
reduced the disturbance from transient throttle
operations, and enhanced the DR.
2.3. Engine Oil and Fuel Development
2.3.1. Engine oil development
Engine oil development was important in order to
reduce friction and achieve durable reliability. Joint
development was performed with Nippon Oil
Corporation.
The power can be increased by lowering the base oil
viscosity and the HTHS viscosity, but lower viscosity
results in decreased durable reliability. As such, selecting
the base oil and developing oil additives suitable for the
parts and coating materials were performed continuously.
Oil deterioration was not an issue because the
replacement intervals were short.
2.2.2. Drivability
In 2008, the use of a common ECU was made
compulsory, and driver aids were banned. In the engine,
attention was focused on the fact that the banning of
traction control and restrictions on control resulting from
the common ECU affected lap times because of issues
of drivability (abbreviated here to DR) when coming
out of corners. Figure 10 shows two DR issues: (a) the
initial torque following at 8000 rpm, and (b) the torque
hole generated by misfiring as the air-fuel ratio becomes
2.3.2. Fuel development

The number of regulations, including for distillation
characteristics, has increased since 1995, eliminating the
prospect of increasing the output through fuel
development. This situation continued for fuel
Lap time (sec)
-0.02
-0.04
-0.06
Power (kW)
N
o
d
a
t
a
20 kW
12 kW
Effect of end of straight engine speed (18500 19000 rpm) on lap time
0.00
Curve B
Curve A
15000 15500 16000 16500 17000 17500 18000 18500 19000 19500
Engine speed (rpm)
-0.08
-0.10
Mel Sep Bah Bcn Mco Mtl
Ind Mgc
Sil
Nur Bud
Ist
Mza Spa Fuj Sha
Fig. 9
Int
Power curve shape comparison
Circuit
Fig. 7
End of straight Ne - lap time

1 sec
308
Lap time
Max car speed
95.45
95.4
304
95.35
95.3
18000
306
302
18100
18200
18300
18400
18500
Pedal
Engine throttle
Maximum car speed (km/h)
Lap time (sec)
95.5
Engine speed
Target lambda
Lambda
Target
torque
(a) Initial torque following
300
18600
Actual
torque
Power peak engine speed (rpm)
Fig. 8
Pe peak Ne - lap time / Vmax
Fig. 10
41
Drivability issues
(b) Torque hole

Table 1
development in the third era of Formula One. From

2003, a small modification by mapping the fuel
properties in line with the engine characteristics was
performed, within the regulatory boundaries. In addition
to power, specific gravity from the perspective of the
fuel efficiency and fuel weight when the vehicle is filled
with fuel were considered. Also, properties were studied
that would be suitable for the atomization due to
modified fuel pressure and injectors.
From 2008, the use of bio-gasoline (5.75% additive)
became compulsory in Formula One, in consideration of
environmental issues.
Load
classification
Minimum load
Low load
Medium load
High load
Maximum load
3. Reliability
3.1. Guaranteed Endurance Distance
In the third era Formula One, engine operation (use
distance) has changed greatly since 2004 due to changes
in the regulations. For this reason, each year a review
was held of the guaranteed durable reliability distance
for the engine. Descriptions of the specific regulations
for each year are omitted here, but the required
guaranteed distance for each session was determined
from the previous years results and the regulation
changes, and was used in the endurance tests on dyno
(Fig. 11).
3.2. Guaranteed Endurance Mode
In order to guarantee the Formula One engine durable
reliability, an endurance test mode needed to be set.
However, this could not be set only through engine
operation (use distance); the loads of each circuit also
had to be taken into account.
In the third era Formula One, loads were classified
into 5 levels to guarantee durable reliability.
3.0L V10
1350 km
2008
2007
2006
2005
1500 km
2004
2003
2002
2001
2000
1992
1991
1990
800 km
420 km
Time (sec)
Engine
homologa
-tion
20 kW
Power/L (kW/L)
Power (kW)
50 kW
Power
Power/L
400 km
Year
Fig. 11
Typical circuits
Monaco
Magnycours, Budapest
Bahrain, Barcelona
Istanbul, Fuji
Monza, Spa
2.4L V8
23% increase
Mileage guarantee
Less than 10 sec

Less than 13 sec
13 sec and over
Less than 13 sec
16 sec and over
16 sec and over
16 sec and over
3.3. Reflecting the Endurance Test Results

After the endurance test on dyno and the end of the
race, the changes in dimensions of various parts such as
the tappet clearance, the wear amount of the piston clip
groove, gudgeon pin hole diameter and connecting rod
bearing, and the piston temperature were all measured.
These were then progressively reflected in engine
hardware studies and in operating condition settings such
as the ignition timing settings and the engine oil
3rd Era
3.5L
Less than 50%

Less than 70%
70% and over
Less than 70%
70% and over
Less than 70%
70% and over
Table 1 shows the load classifications (5 levels) in

the endurance test. Up to 2004, the guaranteed durable
reliability was evaluated with one endurance test mode
for each load. However, from 2005 the regulation
changed to one engine for two races, so the endurance
test mode combined different loads to determine the
guaranteed durable reliability.
Included in this, verification was performed on dyno
for various operating conditions such as the temperature
of oil and water, the individual ignition timing for each
cylinder, the air-fuel ratio, the ignition cut and retard
amount (T/C and O/C control) settings, shift cut during
blipping, end of straight engine speed control (EOS),
torque-drive, over-revving, idling, and the oil
replacement timing, in order to maximize performance.
As a result of the endurance test described above, the
engine speed frequency (Ne frequency) was calculated,
and this was used as the reference for operating in the
actual race. Figure 12 shows the representative Ne
frequencies for the endurance test results and the race
results.
With the increased engine mileage, the current
method for confirming the guaranteed durable reliability
takes a lot of man-hours and time. In the future, it will
be necessary to take measures to guarantee durable
reliability by using Miners rule.
2.4. Achieved Power

Figure 11 shows the achieved power and power per
liter for every year. The power increase trend became
lower due to the engine mileage extension and engine
development regulations, but even so, the power
increased every year. By 2008, Honda had achieved a
power increase of 23% over its level in its first year of
participation in 2000.
2nd Era
Endurance test loads classification

Condition of wide open throttle
Frequency
Continuous time
F1 engine performance transition
800
750
700
650
600
Guarantee frequency
550
500
Actual frequency in race
450
400
350
300
250
200
150
100
50
0
10000 11000 12000 13000 14000 15000 16000 17000 18000 19000 20000
Engine speed (rpm)
Fig. 12
42
Engine speed frequency (2006 Rd 17-18)
temperature control. Figure 13 shows the differences in

the piston roof temperature in line with the cylinder bias
(retard amount from the set ignition timing). It can be
seen that by advancing the average ignition timing for
all cylinders by 1.2 degrees, the roof temperature rises
by about 10C. Feedback was performed for these
results, and the optimum balance for material strength
was set for each circuit based on the power changes,
cylinder pressure changes, and temperature changes that
occurred due to the cylinder bias.
350
Piston roof temperature (C)
340
330
320
310
300
290
BIAS ave -2.35 deg
280
BIAS ave -3.6 deg
270
#1
Fig. 13
#2
#3
#4
#5
#6
#7
Cylinder
#8
#9
#10
Ave.
Cylinder BIAS - Piston roof temperature
4. Conclusion
In the third era Formula One development from 2000
to 2008, high performance Formula One engines were
developed that achieved high engine speed and high
output, and durable reliability for one engine over two
races, with the following results.
(1) An increase in the specific output of more than 230 kW/
L was achieved by increasing the engine speed, using
the dynamic effects of the induction and exhaust pressure
wave, enhancing the combustion, and reducing friction
in various areas.
(2) Circuit simulations were used to optimize the use of the
power peak engine speed, shift-up engine speed, and end
of straight engine speed for each circuit, and lap times
were shortened. Also, knowledge was gained about the
output characteristics for shortening the lap times.
(3) Verification of various operating conditions was
performed on dyno through the segmentation of the
endurance evaluation mode, maximizing performance
and providing durable reliability.
(4) In response to regulations regarding material restrictions
and extension of the use of engines, durable reliability
was provided by using CAE technology and through the
development of engine oil.
Author
Kiyoshi KONDO
Yasuhiro MOTOHASHI
Kazushi OGIYAMA
Hajime ENDO
Satoshi NAKAMURA
Nagao YANAGISAWA
Acknowledgements
The authors would like to express their deep gratitude
to Nippon Oil Corporation for their cooperation in
engine oil development and in the transportation of oil
for the Grand Prix.
43
Technical Description of Formula One Engine

Structural Design
Kiyoshi KONDO*
Hajime ENDO*
Tetsuo GOTOU*
ABSTRACT
Through the third Era of Honda Formula One challenge, it was the technical trend that the Formula One engines
were considered to be one of the elements of entire race vehicle. The engines were carefully designed to have such
characteristics as compact size, light weight, low center of gravity and enough stiffness. These characteristics of engines
were very important to increase aerodynamic and vehicle dynamics performance in order to reduce lap times on the
race tracks. Moreover, Honda engines were always expected to have the highest power output amongst the all Formula
One engine constructors.
This paper describes how the engines, including the lubricating and cooling systems, were designed from the early
stage of development in order to satisfy all demands above.
1. Introduction
In addition, Formula One regulations state that a race

vehicle must weigh no less than 605 kg with the driver
aboard. The engine represents approximately one-sixth
of the total weight of the vehicle, and the achievement
of engine weight savings was therefore of great
significance. If the engine weight was reduced to the
limit considering reliability, it would be possible to
increase the weight of the ballast which is positioned
around the floor and near the center of gravity of the
vehicle by the same amount as the engine weight
reduction. This would not only reduce the vehicles roll
and yaw moments, but would increase the degree of
freedom in setting the vehicles weight distribution, and
would thus contribute to enhanced dynamic performance.
For the same reasons, a low and forward center of
gravity was demanded in the engine itself.
The stiffness of the engine was also important as
same as compactness, light weight, and low center of
gravity. The dynamic performance as the whole vehicle
is greatly affected by its lateral bending stiffness.
Therefore, increasing stiffness in the engine can
contribute significantly to the dynamic performance of
the vehicle, for the engine itself is one of the vehicle
framework components.
Figure 2 shows an overview of the engine as
installed, with its lubricating and cooling systems.
The cooling water is cooled in a radiator positioned
on the left side of the vehicle, where the wind is
Figure 1 shows the positioning of the engine in

Hondas 2006 Formula One vehicle. The engine formed
one element of the overall vehicle framework, for it was
rigidly mounted on the monocoque and the gearbox by
bolts. Therefore, it was important to reduce engine size
and increase its stiffness for the sake of the aerodynamic
and dynamic performance of the vehicle.
Oil cooler
Monocoque
Fig. 1
Radiator
Gearbox
Engine
F1 chassis and powertrain configuration

44

Table 1
Year
Capacity - configuration
V-angle (deg)
Length (mm)
Weight (kg)
COG (mm)
Bore (mm)
Bore pitch (mm)
Bore to bore (mm)
Crankshaft height (mm)
Basic engine configuration and dimensions from 2000 to 2008
(Mugen)
3.0L-V10
72
620
122
194.7
94.4
106
11.6
70.5
2000
2001
2002
2003
3.0L-V10
94
600
604.5
111
99
177
177.2
80
588
111.7
191.3
108
191.3
95
106
11
68.5
108
68.5
Oil tank
Warm air
Gas bag
Radiator
Engine installation with cooling system
passing through the side pontoon. The engine oil

cooler is positioned on the right side of the vehicle
with the gearbox oil cooler. The engine oil tank is
mounted at the front of the engine, and is fitted in a
recess in the monocoque. The entire layout was
optimized in order to reduce lap times on the circuit,
and almost the same layout is adopted by all teams
in current Formula One racing.
2. Engine Technologies
2.1. Engine Size, Weight, and Height of Center of
Gravity
Figure 3 shows annual changes in the total length,
weight, and height of the center of gravity (COG) of
Honda Formula One engines. Table 1 shows the engine
650
400
Fig. 3
2008
80
2007
450
2006
100
2005
500
2004
120
2003
550
2002
140
2001
600
2000
160
Length (mm)
180
2005
2006
2007
2.4L-V8
2008
90
581.5
90.9
171.6
88.6
163.5
97
103
6
63
490
95.2
165.5
95.1
165.1
106.35
9.35
58.5
2.1.1. Reduction of total engine length

The total length of the engine is basically determined
by the bore, the bore pitch, and the cam gear train
layout.
The bore diameter is determined by the maximum
engine speed, by the diameter of the intake and exhaust
valves which would enable the achievement of the target
power of the engine, and by the valve layout. The bore
diameter was once increased from 95 in 2000 to 97
in 2002 during the third era of Honda Formula One. 97
was an optimal setting for engine speed of 19000 rpm.
All engine constructors appear to have used bore
diameters of between 97 and 98 when engine
homologation was introduced in 2007.
The method of cooling between the cylinders is one
factor that determines the bore pitch, and it is also an
important element affecting the concept of water flow
throughout the engine. Figure 4 shows the evolution of
engine water cooling systems. From 2000 to 2002,
separate thin-walled cylinder sleeves were employed to
form separate cooling water paths for each cylinder. The
700
Weight (kg)
COG (mm)
Length (mm)
Mugen
Weight (kg), COG (mm)
200
106
9
66
2004
specifications that determined these parameters.

The plots at the left of the graph are figures for the
Mugen Honda engine supplied to the Jordan team in
2000, with a total length of 620 mm and a weight of 122
kg. Honda works engines returned to Formula One in the
same year, and were supplied to the British American
Racing team. Almost every engine part, including the
auxiliary parts, was newly developed with a focus on the
achievement of compactness, light weight, and a low
center of gravity. This helped to reduce the engines total
length to 588 mm and its total weight to 112 kg.
However, in 2000, other teams were already using
engines weighing less than 100 kg. Hondas engines of
the time were rather large and heavy.
Variety of technologies were applied to reduce total
length, weight, and COG every year. As a result, in
2005, Honda achieved the engine weight of 88.6 kg, the
lightest at the time. However, in 2006, FIA regulations
were revised to change the engine type from 3.0 L V10
to 2.4 L V8. At the same time, limits of at least 95 kg,
165 mm, and 106.52 mm were set for the engine
weight, COG, and bore pitch, respectively. This put all
the teams on basically the same level, and no further
changes have been made since then. This paper will
discuss individual framework-related technologies for
each item below.
Oil cooler
Cold air
Fig. 2
Year
Change in engine length, weight and COG
45
bore bridge dimension in this configuration was 11 mm.

In 2003, the thickness of the bore bridge water jacket core
was reduced to 3 mm and a closed deck configuration was
employed, reducing the bore bridge to 9 mm.
From 2004, the direction of cooling water flow was
modified from a cross flow for each cylinder to a
longitudinal flow. At the same time the bore bridge
water jacket core was abandoned and a Siamese cylinder
block configuration was employed. With respect to the
bore bridge cooling water path, 3 machined holes were
employed to guide cooling water from the heads to the
bore bridge, enabling to reduce the bore bridge
dimension to 6 mm.
Figure 5 shows the reduction of the cam gear train
width in 2000. Compared to the 2000 Mugen engine,
which employed standard reduction gears, the diameter
of the crank gears was reduced in the 2000 Honda
engine to use a gear train layout without reduction gears
(shown at the bottom of Fig. 5). In combination with
the use of a thin front cover, this contributed to reduce
total engine length.
The method of mounting the clutch would be also an
important factor, if the total length of the engine
including the clutch fitted at the posterior end of the
engine was considered. In Hondas 2004 engine, the
clutch basket was inserted in the rear end of the
crankshaft, and was fastened with a single bolt. This
helped to put the clutch position forward by 9.5 mm,
compared to the previous years engine.
reducing engine weight were closely connected to

manufacturing technologies. Some examples are
presented below.
Figure 6 shows the closed deck cylinder block
employed from the latter half of 2002. The key to the
development of this technology was a quality
enhancement technology for casting that employed highstrength shell cores, which was essential to reducing the
thickness of the bore bridge water jacket shell core to 3
mm, and a surface treatment technology by means of
which a Nikasil surface treatment was directly applied
to the inner surfaces of the cylinders. The use of these
technologies resulted in weight savings due to the
reduction of the total length of the engine, in addition
to reductions in friction and blow-by gas with the
achievement of increased stiffness around the bore.
Reduction gears (block idler)
A
2000
Mugen
A
2.1.2. Reduction of engine weight

Between 2000 and 2005, before the advent of the
minimum weight regulation, a diverse range and large
number of technologies, regardless of the magnitude of
their effects, were undertaken to achieve weight savings.
An engine weight of 88.6 kg in a 3.0 L V10
configuration had been achieved by the 2005 final engine
specifications. Many of the effective technologies in
Water jacket
between inserted
sleeve and
cylinder block
SECTION AA
2000
Honda
SECTION BB
Fig. 5
Gear train configuration
Drilled hole
(D = 3 mm)
Shell core
(t = 3 mm)
Water pump
Water pump
Radiator
Water pump
Radiator
Radiator
L-head
Block
R-head
2000 - 2001
L-head
Block
R-head
2003
(twin radiators for 2002)
Fig. 4
Schematics of water cooling systems
46
L-head
Block
R-head
2004 - 2008
Figure 7 shows a cross-section of the cylinder heads

with an integrated pneumatic valve return system
(PVRS) used from 2003, and a sand mold used in their
casting. Until 2002, the cylinder heads were divided into
cam cases and heads in order to create PVRS air
channels. In 2003, a casting technology using complex
head shell cores was developed, helping to enable the
PVRS to be integrated with the heads. This resulted in
the simple cylinder head structure and a weight saving
of 6.2 kg. (This figure should be considered as a
reference value because the other factors also contributed
to this weight saving.)
Laser-clad intake and exhaust valve seats were
developed to achieve longer mileage durability (as
demanded by the regulations) and higher intake potential,

as well as weight saving. Laser-clad valve seat has
higher heat resistance and wear resistance than the
previous press fit seats. The new valve seats were
adopted from 2004, and contributed to a weight saving
of 0.4 kg per engine, in addition to an increase in the
valve diameter (in the case of the intake valves, from
40.6 to 41.6, and in the case of the exhaust valves,
from 31.8 to 32.6), a greater degree of freedom in
valve layout, and increased cooling of the combustion
chamber from the valve seats. A manufacturing
technology enabling a stable, high-quality cladding is an
important element in valve seat cladding. In the present
case, detailed settings by the Material and Prototype
Departments helped to enable the creation of such a
technology.
Cylinder sleeve
Core shape design
Molding condition
Fig. 6
2.1.3. Reduction of COG

The V bank angle and the height of the crankshaft
have a significant effect on the COG. Aiming at equal
interval combustion, a V bank angle of 72 was used in
the 2000 Mugen Honda engine. In the case of the 2000
Honda engine, the emphasis was on reducing the COG
in order to contribute to total vehicle performance, and
a V bank angle of 80 was employed. In 2002, an angle
of 94 was used in order to increase power by reducing
inlet pressure wave interference, in addition to further
reducing COG height. From 2003, facing a greater
demand to increase engine speed in order to increase
power at the same time as reducing the COG, the
demands were balanced through the use of an angle of
90, which theoretically enables secondary vibration to
be cancelled.
In the case of the height of the crankshaft, Table 1
shows that it has steadily been lowered with each year.
The baseline for the COG height (0 mm) is the mating
faces of the skid plate fitted to the bottom of the vehicle
and the bottom of the engine, as specified in the Formula
One regulations.
The achievement of a compact conrod locus (the
envelope surface of conrod oscillation) is an important
technology in lowering the crankshaft while preventing
an increase in friction. The gap between the rotating
parts and the inner wall of the crankcase has a significant
effect on oil agitation resistance in the crankcase. 2005
test results showed that a rapid increase in friction
occurred at a gap of 3 mm or less (Fig. 8).
Drying condition
before casting
Water jacket core

(t = 3 mm)
Material specs of
sand and binder
Closed deck cylinder block and water jacket

core
Pneumatic air manifold
(green)
Head with integrated

pneumatic air passage
Cam case
(yellow)
2002
2003
Friction loss (kW)
5
Thrust gap from main journal to
counterweight
Radial gap from crank case to
conrod locus and counterweight
4
3
2
1
0
0
Gap (mm)
Fig. 7
Cylinder head with integrated PVRS
Fig. 8
47
Friction loss against closest gap
counterweight to be reduced to 58.55 mm. In 2005, the

radius of the counterweight was reduced to 50 mm, and
the crankshaft was lowered by 4.5 mm against the
previous years engine.
The point at which the gap between the rotating parts

and the inner wall of the sump case becomes the
narrowest is the lowest position of the conrod locus,
determined by the locus of the conrod bolt head. CAE
was therefore employed to study the conrod big end,
helping to make the big end compact gradually (Fig. 9,
Fig. 10).
With regard to the reduction of the radius of the
crankshaft counterweight, it was necessary to balance the
reduction of the radius and the achievement of reduced
weight while maintaining a crankshaft balance rate set
on the basis of friction and reliability considerations. Up
to the Italian Grand Prix (Round 15) in the latter half
of 2004, a counterweight manufactured from a tungsten
alloy with a specific gravity of 18 was press-fitted to the
crankshaft, following which it was covered with a plug
and welded in place (Fig. 11). However, in order to
further reduce the radius, from Round 16, the Chinese
Grand Prix, a configuration was employed in which a
separate tungsten alloy counterweight was directly bolted
to the crankshaft. This helped to enable the radius of the
2.2. Engine Stiffness

When the dynamic performance as a whole vehicle
is taken into consideration, increasing lateral bending
stiffness is known to have an effect on the stability of
the vehicle. Figure 12 shows an example of the
deformation of monocoque, engine, and gearbox under
a bending load calculated by the Chassis Development
Division. As indicated above, the engine was directly
connected to the monocoque and the gearbox, and was
thus an element in the stiffness of the vehicle. Increasing
the lateral stiffness of the engine would therefore
increase the stiffness of the vehicle as a whole, thereby
contributing to greater stability on the track.
Figure 13 shows the example of the methods of
increasing engine stiffness besides the use of the closed
deck cylinder block discussed in Section 2.1. Auxiliary
2004
4.2
2005
- Round 15, 2004
Round 16, 2004 -
M10X1.25
M9X1.0
Schematics of crankshaft counterweights
20
24.5
35
38.5
Fig. 11
48
50
2005
2004
Fig. 9
Fig. 12
Reduction of conrod locus
Conrod cap clamp thickness

Bolt head thickness
Bearing thickness
35
35
35
34
34
9.4
10
9.4
10
9.4
10
6.4
10
5.4
10
19.5
20.5
20.5
20.1
2000
2001
2002
2003
Bolt size
Pin journal diameter
Water pump and

water inlet/outlet
1.5
1.5
1.5
1.5
34
34
34
34
19.1
5.4
9
15
5.4
9
15
5.4
9
15
5.4
9
15
2004
2005
2006
2007
2008
Scavenging pump
and air separator
Lower mounts
Year
Fig. 10
Lateral bend calculation
Progress of conrod configuration
Fig. 13
48
Cylinder block with integrated oil/water pump

Table 2
equipments including scavenging pumps and water pump

were integrated in the 2002 cylinder block. The
scavenging pumps and air separators were positioned at
the side of the crankcase, which extended for almost the
total length of the engine, to enable oil to be collected
from each crankcase throw chamber and blow-by gas to
be centrifugally separated at the front of the engine. The
water pumps and water inlets and outlets were positioned
at the front of the engine due to the demands of the
radiator layout. In order to use these auxiliary devices
as elements in engine stiffness, from 2002, the
scavenging pumps and water pumps were integrated with
the cylinder block or sump case during casting. In
addition to increasing stiffness around the front lower
mounts, this also increased the cross-sectional secondary
moment throughout the engine, resulting in greater
stiffness in the engine as a whole.
As indicated above, in 2006, major changes were
made in the regulations that have effects on engine
frameworks, with stipulations demanding the use of V8
engines and setting figures for minimum weight and
minimum center of gravity. An example of a technology
to increase engine stiffness that made optimal use of
these regulation changes will be discussed below.
Figure 14 shows features of the 2006 engine related
to stiffness. There should be some weight margin for
changes to meet the 2006 minimum weight of 95 kg as
the weight of the 2005 engine had already been reduced
to 88.6 kg. In addition, the center of gravity of the 2005
engine was lower, at 163.5 mm, than the 2006 regulation
figure of 165 mm, meaning that the mass had to be
added to the vehicle at a higher position than the center
of gravity. Based on these conditions, 2006 engine layout
concept was to employ the extra weight margin partly
in increasing the joint stiffness of the upper mounts of
23.7
Vertical
Torsional
1.21
1.10
1.34
1.50
1.23
1.16
2.3. Development of Oil Pump

Because Formula One engines use dry-sump
lubrication systems, two types of oil pump are employed:
A feed pump to supply oil and a scavenging pump
(SCAV pump below) to collect oil. Torocoid pump
rotors with four inner teeth and five outer teeth were
employed for both the feed and the SCAV pumps for
their excellent volumetric efficiency. The sections below
will discuss the functions of the pumps and details of
their development.
19.2
73.5
2.3.1. Development of Feed Pump

The necessary oil volume and oil pressure for each
part of the engine varies with engine speed. The sliding
parts of the valvetrain need a high-volume oil supply
from low engine speeds, but the volume of oil demanded
does not increase beyond a specific level as engine speed
increases. By contrast, the crank pins, which use a center
oil supply that is affected by centrifugal force, need
higher volumes of oil at higher pressures as engine speed
Thick wall cam cover

Stiff design of top mounts
Fig. 14
Lateral
the engine and the monocoque. At the same time, the

span of the joints was increased in order to reduce the
degree of local deformation by reducing the input force
around the mount joints.
The front upper mounts were extended to the limit
of the dimensions specified in the regulations; the form
of the cam cover close to the front upper mounts was
modified to increase the section modulus, and the
thickness of the vertical wall of the cam cover was
increased to form stiffening ribs. The gear train layout
was modified to help enable the positioning of the lower
mounts as far as possible to the left and right, and the
shape of the lower mounts was designed to increase the
stiffness of the sump case near the mounts.
Table 2 shows the stiffness of the 2006 Formula One
engine and the 2006 vehicle as a whole based on
simulation results. The stiffness values are normalized
so that 2005 engine stiffness should be 1.00. The engine
stiffness is expressed per unit length because the change
from a V10 engine to a V8 engine stipulated in the
regulations reduced the total engine length from 581.5
mm to 490 mm, and direct comparisons therefore cannot
be made. As for the comparisons of total vehicle
stiffness calculation, the 2005 monocoque was modified
to fit the 2006 engine and gearbox. Figures for lateral,
vertical, and torsional stiffness are all higher for the 2006
model. The achievement of increased lateral stiffness was
one of the aims of the development, and this parameter
was increased by 50% in terms of total vehicle stiffness.
Difference between 2005 and 2006
29
Relative rigidity of 2006 engine (normalized

for 2005 rigidity = 1.00)
Relative engine rigidity

(per unit length)
Total vehicle rigidity
-Engine & gearbox = 06
-Monocoque = 05
2005 front mount position
19.2
2006 engine characteristics for increased

stiffness
49
increases. Figure 15 shows the estimated oil supply to

each part of the engine for the final engine
specifications.
The set oil supply pressure in Hondas initial Formula
One engine specifications was 700 kPa, but this figure
increased to 900 kPa in response to annual increases in
engine speed. There were two main reasons for this. The
first of these was to increase the volume of oil that might
contact to the piston ceiling close to top dead center by
increasing the injection speed of the piston cooling oil
jet. The second was that the supply of oil to the crank
pins using the center supply method against centrifugal
force necessitated higher oil pressures in order to achieve
a stable oil supply at high engine speeds. However, it
is known that if the oil supply pressure is too high, the
oil pressure pulse will increase, and there will be an
effect from negative pressure waves.
The following two points were focused on in order
to increase the efficiency of the feed pump:
(1) In order to prevent loss from pressure relief, the minimum
pump capacity should be set at a figure that enabled the
necessary oil pressure at low engine speeds to be
maintained.
(2) At high engine speeds, a reduction in oil volume and oil
pressure due to pump suction cavitation, and a consequent
loss of torque, should be prevented.
In the case of (1), the appropriate figure could be
estimated to a certain extent based on experience and
theory, and the aim could be relatively easily achieved
by modifying the specifications. In the case of (2),
however, not only was the parameter challenging to
predict, but the achievement of the aim would
necessitate major changes that were contrary to the
concept of a race engine, i.e., reducing the speed and
increasing the size of the rotor. The 2005 feed pump
rotor, of 54 mm in external diameter, was operated at
a high speed (12800 rpm), resulting in a severe decline
in oil pressure. The following measures were employed
against the two main factors in feed pump intake issues
in the high-speed range:
(1) An insufficient volume of oil was flying into the rotor
suction chamber during rotational transfer at high speeds:
A thick rotor presented a disadvantage in this case, and
it was therefore replaced with two thin rotors.
(2) The oil drawn into the rotor suction chamber leaked out
of the chamber due to the effect of centrifugal force: In

this case, a modified suction port shape was adopted in
which the opening to the exterior closed as the chamber
filled with oil (Fig. 16).
These measures increased the flow rate by 30% with
no changes to the basic specifications (Fig. 17).
The main issue of concern in the development of the
feed pump was potential fractures of the driveshaft due
to torsional resonance. The 12 feed pump driveshaft in
the original pump specifications was manufactured from
special carburized steel, and possessed a safety factor
approximately 100 greater than that of the average drive
torque. Nevertheless, fractures occurred frequently. A
simulation of shaft behavior showed that tertiary
torsional resonance at 17000 rpm and above was the
main factor in these fractures. The increase in resonance
was reduced by reducing the diameter of one section of
the driveshaft to 8 in order to reduce the shaft spring.
Fractures did occur following this when the rotor drive
method or the rotor itself was modified, but the
knowledge gained from the analysis discussed here
enabled the situations to be rapidly responded to by
altering the torsional stiffness of the driveshaft or the
rotor inertia.
2.3.2. Development of SCAV pump
The SCAV pump, by rapidly collecting oil that has
been used in lubricating or cooling the engine, enables
a sufficient quantity of oil to be maintained in the oil
tank and helps to ensure sufficient oil pressure when the
vehicle is cornering, in addition to helping to prevent an
increase in friction due to oil agitation by the moving
parts in the engine. This is particularly important inside
the crankcase, where oil agitation tends to generate
increased friction. Each of the crankcase throw chambers
Original port
Fig. 16
80
VLV train
60
Crank pin
Crank main
40
Oil jet
Others
20
Modified port
Comparison of suction port shape (shown as

pink area)
120
Flow rate (l/min)
Flow rate (l/min)
Leakage
protection
Centrifugal
force
leakage
t = 8 mm
x 2 rotors
Modified port
80
t = 8 mm
x 2 rotors
40
t = 16 mm
x 1 rotor
Total
0
0
0
5000 10000 15000 20000

Engine speed (rpm)
Fig. 15
4000
7000
10000
13000
Pump speed (rpm)
Oil distribution
Fig. 17
50
Effect of twin rotor and modified port shape
own inertia (Fig. 19). This design was not optimal for
stable collection of oil, because the oil did not collect
in the SCAV pump inlet when cornering G force was
produced in the opposite direction. The use of an oil trap
in the SCAV pump inlets from 2004 increased the
efficiency of oil collection (Fig. 20).
An internal compression SCAV pump was introduced
in 2003 in order to increase pump efficiency. Formula
One SCAV pumps collect a mixture of oil and blow-by
gas from inside the crankcase at an absolute pressure of
20 to 40 kPa, and send the mixture through exhaust
channels pressurized to levels of between 150 and 250
kPa. This represents a compression ratio of between 4
and 12. Used at such high compression ratios, a
compression pump is more efficient than a conventional
displacement pump (Fig. 21). Torocoid pumps are able
to be used for internal compression, and a torocoid pump
could be modified for the purpose simply by changing
the port shape for delayed opening. Pump damage when
oil alone was absorbed would become an issue, and this
was responded to by positioning a relief valve on the
rotor side. A compression ratio of 2 was employed to
help ensure performance and reliability. This measure
reduced SCAV pump drive resistance by 30% and
engine friction by 3 kW.
is a sealed structure closed off from the others. This

configuration helps to prevent pumping loss due to the
volumetric transfer of blow-by gas by the alternate
actions of the pistons; the gas exhausted by the falling
pistons in the V bank would be sucked up by the rising
pistons. For the same reason, it is desirable to prevent
any contact between the throw chambers in the SCAV
pump that collects the oil. The initial SCAV pump
design used one SCAV pump for multiple throw
chambers, but later, independent SCAV pumps were
used for each throw chamber in order to increase
energetic efficiency. In addition, to help ensure stable
collection of oil in the case in which the oil was
unevenly distributed to the front or rear, pumps were
placed at the front and rear of each throw chamber.
Figure 18 shows the oil system based on the final
specifications.
The main function of the SCAV pumps is the
collection of oil, but blow-by gas also plays an important
role in this. Just as a vacuum cleaner would not be able
to suck up dust in a vacuum, the SCAV pumps would
be unable to collect oil in the absence of blow-by gas.
For this reason, engine breathing supplies an optimum
volume of blow-by gas from the upper section of the oil
tank to the heads and the gear housings, where blow-by
gas is not normally present. However, oil mixed with
large quantities of blow-by gas can have a variety of
negative effects on the feed pump, including interfering
with filling and producing bearing damage. For this
reason, after the oil is collected by the SCAV pumps,
the oil and the blow-by gas are separated in an oil-air
separator that uses centrifugal separation, and are then
sent to the oil tank by means of separate channels.
When a person drinks water through a straw, the
water could not be drawn up if the end of the straw was
not completely immersed in the water. In the same way,
the SCAV pump inlets must be completely immersed in
oil in order for the pumps to collect oil. In the initial
SCAV pump design, oil struck by the crankshaft and the
conrod flew into the SCAV pump under the force of its
2.3.3. Reduction of weight of pump rotors

The rotors were initially formed from a sintered
aluminum powder material, but later a magnesium alloy
was employed for the inner rotor and a plastic material
for the outer rotor(1) in order to achieve weight savings.
SCAV
rotor
Fig. 19
Straight suction
To air box
Oil tank
3rd
SCAV
rotor
2nd
Main
Oil
cooler
Oil filter
Oil feed
pump
Fig. 20
Centrifuge
LHS
cylinder
head
Trap-type suction
t20
Crankshaft
Piston
oil jet
t11.5
t21.5
t16.5
t16.5
t16.5
P2
P2
Scavenging
pump
t16.5
t21.5
RHS
cylinder
head
Fig. 18
P1
P1
t11.5
t20
V1
Displacement pump
Oil system
Fig. 21
51
V2
V1
Compression pump
Comparison of pump work
This reduced the weight of the rotors by half. To help

prevent friction due to the use of soft materials, a unique,
low contact face pressure tooth shape was employed. In
another year, a sintered steel powder material was
employed for the feed pump rotor to help prevent
torsional resonance, an issue which was discussed above.
power loss in accordance with high water temperature

operation was restrained at the minimum.
Figure 23 shows the influence of water temperature
on engine power.
Until 2002, the radiator and the oil cooler were each
divided and positioned on the right and left of the
chassis. In order to simplify the system, including the
pipe layout, from 2003, the radiator was installed on the
left side and the oil cooler on the right side of the
chassis. In this configuration, the cooling water on the
right side of the engine bank was recirculated into the
water pump suction without passing the radiator, and
mixed with the water returned from the radiator on the
left side of the chassis. The configuration did not affect
performance and reliability.
The introduction of a Siamese block from 2004 in
order to reduce the total length of the engine had a
significant effect in changing the flow of cooling water.
The water sent from the water pump was allocated to
each bank of cylinder heads, and flowed from the front
to the rear, following which it flowed from the rear of
block to the front. This increased the flow rate in the
water jackets, thus increasing the rate of heat transfer.
This configuration also reduced the volume of cooling
water inside the engine from 3.2 L (2003) to 2.1 L
(2004), contributing to the achievement of weight
savings (Fig. 4). Figure 24 shows temperature
measurements taken inside the engine for each water
jacket configuration. The results show that increasing the
flow rate reduced the temperature of the combustion
chamber walls and the area close to the surface of the
deck. A water channel was also machined between the
2.4. Cooling System

Cooling system development in Hondas third
Formula One era commenced with the aim of reducing
the size of the radiator. The size of the radiator has a
significant effect on the vehicles aerodynamic
characteristics, i.e., is directly related to lap times.
Reducing heat rejection and enhancing heat transfer
coefficient are essential factors in reducing the size of
the radiator, and the development was focused on them.
The water jacket was made as small as possible, in
order to help reduce heat rejection by enabling only the
areas in which cooling was necessary to be cooled and
to help prevent unnecessary heat transfer from the
exhaust port. Figure 22 shows the design concept of the
water jacket. In addition, individual cylinder type water
jackets (Fig. 4) were applied to the cylinder heads in
order to reduce the difference in temperature between
cylinders. These measures reduced heat rejection by 25%
against a conventional engine, and reduced the size of
the radiator by 25%.
Increasing the temperature of the coolant, i.e.,
operating the engine at a high water temperature is an
effective method of enhancing heat transfer coefficient.
The entire water system was pressurized to 365 kPa in
order to help prevent boiling, and the flow rate was
increased to help ensure sufficient reduction of the
temperature of the walls of the combustion chamber
despite the increase in water temperature. In addition, the
ability to employ high-temperature water helped to
enable a greater degree of closure in the cooling option
(the degree of opening of the radiator exhaust port),
which is more advantageous from the perspective of
aerodynamic performance. While conventional engines
operated at a water temperature of 90C, during Hondas
third Formula One era, a temperature of 120C was
achieved in 2000, and the maximum temperature of
130C was achieved in 2008. At the same time, the
Output (kW)
Minimum water
jacket to prevent heat
transfer from exhaust
port
Heat
Water
10
Fig. 23
130
Effect of water temperature on engine

performance
Temperature comparison for different water flow systems
300
Water gallery
to prevent heat
transfer to
suction port
2003
2004
250
200
150
100
50
0
High-velocity water jacket

design to increase heat
transfer coefficient
Head
(between
exhaust valve
seats)
Optimization of flow mass

distribution between
cylinder head and sleeve
Head
(between inlet
valve seats)
Head
(bore-bridge)
Block
(bore-bridge
upper)
Block
(bore-bridge
middle)
Place
Fig. 24
Fig. 22
90
100
110
120
Operating water temperature (C)
80
Temperature (C)
Thermal insulation of
exhaust gas
2000
Honda
Minimum water jacket to

prevent heat transfer to
suction port
Long-reach and small-diameter
spark plug to make room for
water flow
Conventional
engine
Design concept of water jacket
52
Temperature comparison between 2003 and

2004
block bore bridges to help prevent an increase in

temperature between the bridges.
Cooling system development from 2008 aimed to
enhance the aerodynamic performance of the vehicle, and
therefore sought to reduce the size of the radiator. To
help enable this, it was necessary to increase the head
of the water pump and the water flow rate.
3. Conclusion
A compact, lightweight, low center of gravity, highstiffness and high-power Formula One engine framework
with an excellent vehicle fit was progressively developed
to reduce lap times on the circuit by enhancing total
vehicle performance. The following results were
obtained:
(1) The reduction of engine size and weight progressed
through a fusion of material technologies, manufacturing
technologies, and framework design, including the design
of the cooling and lubrication systems. As a result, in
2005 (the final year in which 3L V10 engines were used
in Formula One) Honda reduced the weight of its engine
to 88.6 kg, producing one of the lightest and the most
compact engine among the Formula One teams.
(2) Reciprocating system parts and a high-efficiency
scavenging pump were developed in order to lower the
center of gravity, helping to enable the height of the
crankshaft to be reduced to 58.5 mm with a minimum
increase in friction.
(3) Making maximal use of the stipulations of significantly
revised engine regulations in 2006, the stiffness of the
engine was increased for the sake of the dynamic
performance of the vehicle. 2006 engine contributed to
an increase of 50% in the vehicles lateral bending
stiffness against that of the previous years.
(4) Feed pump technologies were developed to accept high
speed operation and hence enable more precise response
to oil demands in each part of the engine, a lightweight
and compact design, and a reduction in drive loss.
(5) The use of independent SCAV pumps provided with oil
traps at their suction ports increased oil collection
efficiency. The use of an internal compression technology
reduced drive loss by 30%, and the use of substitute
materials in the rotors halved their weight.
(6) The optimization of the water jacket configuration
reduced heat rejection and helped to ensure reliability
with operation at high water temperatures while
controlling performance loss, thus contributing to
enhancing the aerodynamic performance of the vehicle.
Author
Kiyoshi KONDO
53
Hajime ENDO
Tetsuo GOTOU
Development of Reciprocating Parts and

Crankshaft in Hondas Third Formula One Era
Tetsuo GOTOU*
Nagao YANAGISAWA*
ABSTRACT
Reciprocating parts and crankshafts for race use must display reliability and achieve low friction levels. Increasing
engine speed is an effective method of increasing the power of natural aspiration engines, and Honda increased engine
speed to 19600 rpm in 2006, prior to the introduction of upper limits for engine speed. In 2005, Formula One regulations
were changed to increase the distance for which engines would be used from one race event (400 km) to continuous
use for two race events (1500 km). In order to balance performance under severe use conditions with the achievement
of reduced friction, the shapes, materials, and manufacturing methods used for the reciprocating parts and crankshaft
were modified to reduce weight and increase strength and stiffness. In FY2008, the weight of the reciprocating parts
was reduced to 358 g, representing a weight saving of 41 g against the FY2000 figure of 399 g. This paper will
discuss the development process, focusing on each component part.
1. Introduction
increasing the accuracy of CAE.

Performance demands were fulfilled and durability
and reliability and reduced friction were realized through
the development of new manufacturing methods, new
materials, and surface treatment technologies such as
diamond-like carbon (DLC) coatings.
This paper will discuss the history of the
development process with a focus on individual parts.
Like many other engine parts, the reciprocating parts

used in Formula One engines must reconcile multiple
conflicting performance demands. It is advantageous for
reciprocating parts to be heavy and robust, in terms of
resistance to high combustion pressures for long periods,
but the achievement of increased engine speed
necessitates lightweight parts.
In addition, reducing the weight and the contact
area of the reciprocating parts represents an advantage
in terms of the reduction of sliding friction, but going
too far in this direction can increase friction and result
in wear.
As in the case of previous engine development
programs, the development program for reciprocating
parts in Hondas third Formula One era developed a
large number of new technologies that were later
employed in races.
The introduction of CAE resulted in significant
changes in the approach to technological development.
In particular, from 2003 it became established procedure
for designers themselves to construct 3D models and
conduct CAE analyses, and this helped to enable the
optimal design of structural strength, resonance, flow
rates, temperature distribution, heat stress and other
parameters within short time periods.
The analysis of relevant phenomena using dyno tests
and measurement technologies was also effective in
2. Changes in Specifications of
Reciprocating Parts and Crankshaft
Table 1 shows annual changes in the use conditions,
specifications, and materials of the main reciprocating
parts, and Fig. 1 shows changes in the weight of the
parts.
The weight of the reciprocating parts increases in
some years. This is due, in 2002, to the change in the
cylinder bore diameter, in 2005 to responses to
regulations stipulating increased use distances, and in
2006 to the introduction of regulations prohibiting the
use of aluminum metal matrix composites (MMC) for
the pistons.
In 2004, despite the extension of use distances by the
regulations, the use of CAE analysis in MMC piston
design helped to enable the achievement of weight
savings in the pistons. The upshift engine speed was
basically increased each year, but regulations limited
engine speed to 19000 rpm in 2007.

54

Table 1
Year
Miledge (km)
Engine revolution (rpm)
BorexStroke-Type
Piston
Piston ring
Piston pin
Connecting rod
Crankshaft (Pin/Main)
Reciprocating parts and crankshaft specification history
2000
2001
2002
2003
2004
2005
2006
2007 & 08
400
1500
420
800
1350
17500
17800
18600
18800
19200
19200
19600
19000
95x42.24-V10
97x40.52-V10
97x40.52-V8
Bridge #rib
Radialy rib
# rib
Rectanglar
WR
Mono R
Titanium WR
18.7-MAS1C
17-MAS1C
17TiAl
17-HD200
I-section
Hollow
I-section
( 34/45)
( 36/48) ( 34/46) ( 33/45) ( 34/46)
( 35/52) NT100
GKHYW GKHYW GKHYW
GKHYW
NT100
of a weight saving of 15 g. However, both these piston

shapes resulted in the conrod small ends being covered
by the piston walls, increasing the challenge of providing
an oil supply around the conrod small ends and the
piston pins. To resolve this issue, and to achieve further
weight savings, the shape of the pistons was further
modified to use curb bridge ribs (# bridge ribs) with no
wall on the piston ceiling side. Changes continued to be
made, but this basic piston shape was adopted until 2008
(Fig. 3).
The weight of the pistons, which stood at 255 g in
2000, had been reduced to 210 g by 2004. The use of
MMC, which increased high temperature strength by
30%, contributed to this reduction in weight.
The reduction of the weight of the pistons did not
only increase their toughness against inertial forces, but
also reduced friction. However, it was not sufficient for
the pistons to merely be lightweight. It was also essential
for the sliding of the piston skirts to be smooth and
stable.
The sliding resistance of the piston skirts, which slide
at high speeds on an oil film, is determined to a greater
400
Weight (g)
350
Piston
Recipro
total
Conrod
big end
300
250
200
150
00 01 02 03 04 05 06
Model year (year)
Fig. 1
'07 and 08
Reciprocating parts weight history
3. Piston Development
The pistons are directly affected by the energy of
combustion gases, and must withstand maximum
combustion pressures of 7 t and inertial forces exceeding
2 t while being exposed to temperatures in excess of
2000C.
The general orientation of the development program
was to reconcile high strength with low weight. CAE
was introduced at an early stage to assist in the
realization of optimal designs. In the initial stages of
development, Honda CAE specialists were requested to
perform simulations, but by degrees designers themselves
became able to construct 3D models, and made
judgments regarding strength and stiffness from the
results of CAE analyses incorporating combustion
pressure and inertial forces. CAE also helped to enable
designers to study temperature distribution, thermal
deformation, and thermal stress, and the ability to predict
lifespans and allowable operating conditions for parts
from the results of such simulations increased the
accuracy with which the shapes of parts were evaluated
(Fig. 2).
The use of these analytic techniques led to a variety
of changes being made in the shapes of pistons. The
initial external shape of the pistons featured a skirt
extending around the circumference, and ribs radiating
for equal lengths in eight directions. To reduce the
weight of the pistons, their shape was modified to use
curb ribs (# ribs). This helped to enable the realization
Temperature distribution
Thermal deformation
Thermal stress distribution
Fig. 2
Examples of piston CAE results
Radial rib
Fig. 3
55
# rib
# bridge rib
Piston structure history
extent by the shape of the skirt surface than the specific

friction coefficient of the material forming the skirt
surface. Rough and deep streaks on the piston skirt
surface increase oil retention, but sliding resistance also
increases. Sliding resistance is low when shallow
striations are employed, but they encourage scuffing of
the skirts due to the breakage of oil film. Scuffing was
a concern when designers halved the depth of the streaks
in order to reduce friction, and the issue was responded
to through the use of steel plating and a plastic coating
containing metallic pigments, and by modifying the skirt
forms.
The initial method of designing the piston skirt form
involved a burner method, in which the piston ceiling
is exposed to the flame of a burner to measure thermal
deformation. A basic form would be established on the
basis of the measurements and damage to the skirts
following durability tests would then be studied to make
further minute adjustments. The increased accuracy of
CAE design technology, however, helped to enable the
use of a surface pressure method, by means of which the
form of the skirts was designed on the basis of CAE
analyses of thermal distortion and skirt stiffness values.
This method helped to enable the realization of skirt
marks largely corresponding to development targets
through the use of simulations alone.
The detachment of the pin clips was a piston issue
that remained unresolved until the end of Hondas third
Formula One era. As engine speed increased annually,
engine blow occurred frequently as a result of the
detachment of the pin clips. The clips would fracture and
detach from their grooves, and the issue was therefore
responded to by increasing the strength of the clips
through shot peening and nitriding, but the issue
reoccurred as engine speed increased.
It was hypothesized that the detachment of the clips
resulted from the collapse of the clip grooves. As
countermeasures, the diameter of the clips was increased
in order to reduce surface pressure and increase strength,
and an oil jet was introduced to cool the clips and
provide lubrication. It was thought that the issue of
detachment of the pin clips had been resolved with the
introduction of these measures, but the issue reoccurred
with an increase in combustion pressure.
Due to the regulations stipulating periods in which
design changes were prohibited, further countermeasures
could not be effected. In endurance tests on a test bed
conducted in 2008, one-third of the tested engines
stopped as a result of detachment of the pin clips.
However, this only occurred after a distance of 1000 km
had been exceeded, and therefore did not represent an
issue during actual driving on the track, including race
operation.
As test verification proceeded, it was theorized that
the collapse of the clip grooves that was resulting in
detachment of the clips was not the result of buckling
due to the load from the pins, but of wear resulting from
repeated minute deformations of the clip groove due to
combustion pressure and inertial forces. Because it was
not possible to prevent the deformation of the clip
grooves, countermeasures focused on increasing

slidability between the clips and the clip grooves, and
increasing the wear resistance of the clip grooves.
However, the effectiveness of these measures was not
verified.
4. Development of Oil Jet

For aluminum pistons, cooling is an important factor
in realizing strength and dimensional stability in the
piston material and reducing thermal stress.
In the initial stages of development, four or six holes
were used per cylinder in the oil jets employed to cool
the pistons (Fig. 6). The flow rate oil injected was
slightly higher than 10 l/min.
As development proceeded, increasing engine speeds
and power output increased the thermal load on the
pistons. The flow rate of oil injected was steadily
increased by means of increasing the diameter of the oil
jet holes, but the use of MMC, a material with good
high-temperature strength, a low linear expansion
coefficient, and a high level of resistance to thermal
stress, to manufacture the pistons during this period
meant that fundamental modifications were not sought.
In 2006, the use of MMC was prohibited. Honda
returned to the conventional 2000 aluminum material for
the manufacture of Formula One pistons, but cracks
occurred frequently in the aluminum pistons when
attempts to increase engine speed commenced. The
mechanism of the cracks occurring in the thin areas of
the ribs (Fig. 4) and the centers of the bridge ribs (Fig.
5) could not be explained by CAE analyses of
combustion pressure and inertial forces, but the results
of thermal analysis showed that the heating of the piston
ceilings and increased temperatures added a tensile
thermal stress exceeding 50 MPa to the stress that had
Thermal stress:
50 MPa
Fig. 4
Crack due to thermal stress in thin area of rib
Thermal stress:
75 MPa
Fig. 5
56
Crack due to thermal stress in bridge rib
already been studied, and this resulted in the cracking.

Figures 4 and 5 show the results of thermal analyses
of the thin area of a rib and the center of a bridge rib
respectively.
Because the centers of the piston ceilings bore
combustion pressure, strong tensile stresses were
produced on them. In addition, their temperatures were
increased by longer exposure to flame. This resulted in
cracking due to a decline in the strength of the materials.
The ignition timing for each cylinder was retarded by
an average of 6.5 as a countermeasure. However, the
results of endurance tests on a test bed showed that
retarding the timing resulted in a power loss of more
than 7 kW against optimum ignition timing, and also
resulted in a decline in fuel efficiency.
This made the cooling of the piston ceilings an urgent
issue. However, simply increasing the volume of oil used
in the oil jets would also increase friction, and was
therefore not an effective countermeasure.
The thin areas of Formula One pistons are only 2 mm
or less thick, and therefore even if the volume of oil was
increased in one area, the cooling effect would tend not
to extend to surrounding areas. An even spread of
cooling oil over a wide area was therefore desirable. The
six holes used per cylinder for the oil jets were increased
first to 12 holes and finally to 24 holes (Fig. 6). Because
the centers of the piston ceilings, where the temperature
increase was greatest, were shielded by the conrod small
ends, it was challenging to directly introduce oil to this
section, but this issue was resolved through the use of
Inlet
an oil jet in which the point of oil discharge had been

raised to a level at which it almost made contact with
the piston ceilings. In addition, the up-down motion of
the pistons extended the distance of travel of the points
at which the oil made direct contact, and this expanded
the range of direct contact of the oil jets (Fig. 7).
However, it had been necessary to increase the size
of the oil jet in order to raise the point of oil discharge,
and the incidence of cracks due to engine vibration then
became a concern. While the oil jet could be used on
the inlet (In) side, the issue could not be resolved on the
exhaust (Ex) side by the deadline for engine submission.
The 2007 homologation engine finally included an oil
jet using 19 holes per cylinder.
The use of an equivalent volume of oil per cylinder
to the previous oil jet with six holes per cylinder reduced
the maximum temperature of the pistons by 25 C. The
achievement of strength increases in the pistons helped
to enable the previously necessary 6.5 retardation of
ignition timing to be reduced to 1.5, and power loss
from the retardation no longer represented an issue.
Direct measurements were initially relied upon for
adjusting the flow rate of the multi-hole oil jet, but
increases in the accuracy of CFD meant that the
adjustments could eventually be made using CAE alone.
Because it was also possible to predict the piston
temperature distribution resulting from the use of the oil
jet, it became possible to design the strength of the
pistons in tandem with the oil jet.
Exhaust
Inlet
Exhaust
Final generation 24-hole oil jet
Early generation 4-hole oil jet
Fig. 6
Evolution of oil jets
Long cooling distance

Short cooling distance
High
jet
Low
jet
High
jet
Low
jet
At TDC
At BDC
Fig. 7
Efficiency of high oil jet
57
5. Development of Piston Rings
increased oil consumption to the 30 km/l range. It

therefore could not be used in races, and was only used
in the Grand Prix qualifying round.
In 2003, regulations prohibited any change of engine
parts between the qualifying round and the race, meaning
that the ring could no longer be used in the qualifying
round. From 2003, a twin-ring configuration was
employed in which an identical expanded ring was
incorporated in the oil ring groove. The use of the new
rings increased the engine speed at which fluttering
occurred by more than 1000 rpm against the previous
rings, and returned oil efficiency to the 100 km/l range.
In 2006, the steel expanded rings were replaced by
titanium in order to increase fluttering toughness by
achieving further weight savings. Titanium displays a
high level of aggressivity, making damage to the piston
groove side or the expander a concern. To respond to
this concern, a tungsten carbide coating (WCC) was
applied to the rings and a DLC coating was applied to
the expanders (Fig. 9). The effect of the reduction in
weight achieved through the use of titanium was
tremendous, preventing the increase in pressure in the
crankcase due to fluttering not only at wide-open throttle
but also when the throttle was off, and helping to enable
oil consumption to reach 150 km/l. Figure 10 compares
the status of pressure inside the crankcase for the steel
and titanium rings.
In mass production engines, one important function of
the piston rings is to enable heat to escape from the
pistons to the cylinders. In Formula One engines, by
contrast, the pistons are cooled by powerful oil jets. In
addition, the heat transferring surface area of the expanded
rings was also extremely small. The contribution of the
piston rings to cooling was therefore minimal, and no
issues arose as a result of using titanium, a material with
a low rate of thermal conductivity.
Maintaining the benefits provided by the expanded
rings necessitated strict control of the taper angle, leading
to the development of a dedicated measurement jig. This
jig incorporated a ring groove of the same dimensions as
the pistons and an external cylinder of the same
dimensions as the sleeves, and was able to accurately
reproduce the state of the rings when fitted. When a ring
was fitted on the jig, it was scanned by the probe of a
form measurement device through two measurement
windows, enabling measurements to be taken for the entire
ring (Fig. 11). Only rings within a tolerance range of a
mere 6 were employed in race engines.
In order to achieve weight savings and reduce sliding

friction, Formula One engine piston rings are formed
from top rings and oil rings.
The initial piston ring configuration was identical to that
used in mass production vehicles. The rings were formed
from a top ring with a rectangular cross-section and an oil
ring with a spacer expander inserted between rails.
In 2002, expanded piston rings were developed in order
to reduce friction and the decline in oil efficiency and blowby gas leaks produced by fluttering. These were ultra-fine
rings of 0.9 mm in width and 1.4 mm in thickness, with a
rear expander to provide tensile force (Fig. 8).
In the initial year of development, a single ring
configuration without an oil ring was tested. This
configuration boosted power by more than 10 kW, but
Expander
Outer ring
Steel
Piston
Fig. 8
Expanded piston ring
CrN
DLC
Ti
Piston
Fig. 9
WCC
Titanium expanded piston ring

Red: Former
steel ring
Pressure
(kPa)
Blue: Titanium
expanded ring
Time (sec)
Fig. 10
Rise of crank case pressure due to piston

ring fluttering
Probe
Expander
Outer ring
Scanning
Outer jig
Inner jig
Fig. 11
Expanded ring taper angle measurement jig
58
6. Conrod Development
horizontal dimensions of these sections, ultimately they

are assemblages of three thin sheets. The torsional
stiffness of these sections is determined by the cube of
the thickness of the sheets, and does not reach a high
figure. By contrast, the torsional stiffness of structures
with hollow sections such as circular pipes and square
pipes is proportional to the cube of the horizontal and
vertical dimensions of the structure, and a high figure
can be achieved even if the structure is formed from
thin sheets.
Based on the theory that the ideal cross-section for
the conrod shafts was not an I- or H-section but a hollow
section, development efforts commenced to produce a
hollow conrod (this was termed a box conrod due to
the fact that the hollow section was rectangular). The
structure was realized through the use of diffusion
bonding (Fig. 13).
Evaluation of the form of the conrods using CAE
analyses and the use of technologies to assist in
hollowing the rods helped to reduce the weight of the
conrods by 30 g, enhance the touch of the valve recesses
in the direction of the circumference, and increase the
degree of freedom in setting the compression ratio.
However, regulations established to reduce costs
prohibited the use of this distinguished technology from
2005, and it was therefore only used for a two-year
period, from 2003 to 2004.
The conrods transmit the up-down motion of the

pistons to the crankshaft and convert it into torque. They
are therefore subjected to powerful compressive and
tensile forces. Theoretically, thrust should not be
generated in the conrods, but thrust is actually generated
due to the deflection of the crankshaft and the inclination
of the pistons. This thrust must therefore be limited by
position adjustments.
Normally, the side faces of the conrod big ends
contact the crankshaft to limit thrust, but if the motion
is stable and the thrust load is sufficiently small, thrust
can be limited by the side faces of the conrod small ends
making contact with the pistons. In this latter
configuration, the relative sliding speed of the thrust
faces is low, and friction is therefore reduced. In the
initial year of development the conrod big ends were
used to limit thrust, but later the small ends were used
in order to achieve the effect described above. Because
this would increase wear on the sliding faces, a DLC
coating was applied to the sides of the conrod small end
and oil grooves were engineered in the sides of the
piston bosses and a fluorine plastic coating applied.
These measures helped to enable the achievement of a
3 kW increase in power.
In addition, measurements taken using the link
method showed that the motion of the crankshaft
generates a torsional resonance in the conrod shafts, the
amplitude of which reaches 5. The fact that torsional
resonance was generated was known before these
measurements from circumstantial evidence, such as the
facts that if the piston recesses were not elliptical, they
would strike the valves, and that the actual strength of
the conrods was much lower than their unit tensile and
compressive strength. For this reason, limits were placed
on the reduction of the section area of the conrod shafts
in the quest to achieve weight savings by the need to
maintain torsional stiffness, rather than by simple tensile
and compressive strength.
A normal conrod shaft has an I-section or an Hsection (Fig. 12). No matter how large the vertical and
Fig. 13
Hollow conrod
7. Bearings Development
Fig. 12
Issues involving main bearings were rare during

Hondas third Formula One era, and the bearings
development program therefore chiefly focused on the
development of conrod bearings. This was the case
because of the severity of the use environment for the
conrod bearings in Formula One engines. The diameter
and width of the conrod bearings used in Formula One
engines are similar to those used in light mass production
engines with up to 660 cc displacement, and their
dimensions are almost identical.
Former I-section conrod
59
Table 2
However, the loads on the conrod bearings during

four cycles at wide-open throttle are constantly several
times higher than those on the production car conrod
bearings (Fig. 14). Their sliding speed is 2.6 times
higher, and sliding resistance generates up to
approximately 15 times as much heat.
In addition, cavitation attacks are induced by the
inertial force of the reciprocating parts, which
exceeds 5 t.
For these reasons, the attempt to ensure durability
was a prime concern from the initial stage of
development. A change in any condition, such as
increased engine speed, increased frequency of use of
wide-open throttle, or increased oil aeration, would
frequently result in damage to bearings, leading to the
seizure of the conrods. The development program was
particularly challenging in years in which the regulations
were altered to double the use distance demanded from
Formula One engines.
The 2004 increase in the use distance of the engines
from 400 to 800 km was successfully responded to by
reducing the weight of the reciprocating parts, and by a
variety of other initiatives including modifying the shape
of the relief and adjusting the bearing clearances.
With the advent of the 2005 regulation stipulating
that an engine must have a life of 1500 km, the materials
of the linings and overlays were modified and their
thickness adjusted, in addition to which the shape of the
relief was modified, the clearances were altered, and
high-viscosity oil was employed. Despite these efforts,
however, conrod bearing seizure occurred frequently in
endurance tests on a test bed.
Finally, the tests showed that a copper alloy with
added silicon (silicon bronze) displayed a life of 1500
km when the contact pressure was increased two-fold or
more. In addition, no serious damage occurred and no
issues were caused by bearing seizure even when lowviscosity oil was used.
The design concept of silicon bronze differs from that
of the conventional conrod bearings. The conventional
bearings were bi-metal, bringing together the strength of
steel and the heat dissipation and slidability of copper,
and were produced by sinter bonding. However, because
Si-Cu
Thermal conductivity
0.2
Youngs modulus
159
50
(MPa)
570
420
(GPa)
148
210
15 m
Overlay
(PbSnInCuSiN)
0.2 mm
1.8 mm
Lining
(CuPbln)
Fig. 15
Backmetal:
Steel
Former bearing layer
20 m
20
Overlay
(PbIn)
0
20
20
20
40
2 mm
Vertical bearing load (kN)
(W/mK)
40
40
Former
metal
the strength of the surface was determined by the copper,

and heat transfer to the conrod was determined by the
steel, the bearings also combined the undesirable
properties of both materials (Fig. 15).
While costly, silicon bronze is stronger than steel,
and displays a level of heat dissipation and slidability
close to those of copper. The back metal was produced
from this material with the aim of achieving a dramatic
increase in toughness (Fig. 16). The Youngs modulus
of silicon bronze is lower than that of the conventional
conrod bearings, and it displays a greater degree of
crushing at identical contact pressures, giving it a good
ability to follow the deformation of the conrod big ends.
The ability to maintain the contact pressure also
represented an advantage from the perspective of heat
dissipation (Table 2).
The contact pressure was initially set low in order to
protect the conrod bolts, but insufficient contact with the
conrods prevented the material from displaying its full
heat dissipation performance. Increasing the contact
pressure helped to enable the expected heat dissipation
performance.
Following this, a copper alloy with added silicon and
nickel (Corson alloy) with good strength, thermal
60
40
F1
Production car
60
Horizontal bearing load (kN)
Fig. 14
Characteristics of bearing backmetal

materials
Conrod bearing loads at WOT, Ps peak

revolution
Fig. 16
60
Lining =
Backmetal:
Si-Cu
Si-Cu bearing layer
input shaft was adopted (Fig. 18).

The use of the high-strength material GKHYW
(corresponding to AFNOR 32CDV13) and application of
a technology helping to enable the removal of the brittle
layer of the oil holes also proved effective, and from
2003 there were no instances of fractures of the
crankshaft due to torsion.
In order to reduce the counterweight radius, from
2004 a separate weight was produced from a tungsten
alloy (heavy metal below) with a specific gravity of
17.5 g/cm3. This counterweight was then bolted to the
crankshaft. The production time for this part, with its
different rate of balance, was reduced from three
months to three weeks, and it could also be rapidly
adjusted (Fig. 19).
Crankshaft fractures due to torsion had been
prevented, but cracks in the pin fillet R due to bending,
while rare, continued to occur. To respond to this issue,
a stress relief groove with a shape designed using CAE
conductivity, and sliding performance was employed in

order to respond to further increases in engine speed.
No active development programs were conducted in
the area of main bearings, but in 2005 roller bearings
were introduced as a measure to reduce sliding friction.
To achieve resistance to high loads, the outer wheels and
the rollers were manufactured from tool steel, and the
optimum roller diameter to disperse surface pressure was
selected by studying diameters in 2 m increments.
However, despite these elaborated measures, pitting
of the sliding surfaces and fractures of the retainers
remained issues, and completing the minimum distance
of 1500 km set that year represented a challenge. In
addition, issues in the seals of the throw chambers for
the journal bearings resulted in communication between
the chambers, and the resulting pumping loss was judged
as having largely canceled out a 4 kW increase in power
achieved through the reduction of friction. The use of
the roller bearings was therefore ceased.
8. Crankshaft Development
Crankshaft
The crankshaft converts the combustion pressure

acting on the pistons into torque and outputs this torque.
Because combustion pressure and inertial forces mutually
act on the crankshaft, the achievement of bending and
torsional strength and the adjustment of the specific
frequency of vibration of the crankshaft are important
factors in the strength of the crankshaft. Oil supply lines
also fulfill an important function, helping to ensure a
stable supply of oil to the bearings and the counterweight
employed to reduce the vibration and friction imparted
to the engine by the reciprocating parts.
Preventing fracture of the crankshaft due to torsional
resonance and reducing friction and vibration are not
easy matters in high-speed and high-power engines such
as Formula One engines. For this reason, from the engine
concept stage, studies were conducted using CAE and
past data with a focus on the vee angle (limited to 90
from 2006) and order of ignition in order to determine
the diameter, length, and weight of the pins and journals
as well as the mounting method of the clutch. When the
actual engine was produced, measurements of the speed
at which the crankshaft resonated, the torsional angle of
the crankshaft when resonating, and crankshaft friction
were taken, and counterweight specifications providing
a good overall balance were determined.
The important points were to attempt to ensure that
no resonance peak existed in the standard speed range
of 15000 to 19000 rpm, and that the torsional angle of
the crankshaft when resonating was sufficiently small.
In the initial configuration, the clutch was connected
to the crankshaft by a fine input shaft, but from 2003,
the clutch was directly connected to the crankshaft in
order to prevent crankshaft resonance from occurring in
the standard engine speed range, by reducing the speed
at which the crankshaft resonated (Fig. 17). The
transition to V8 engines in 2006 necessitated an increase
in the speed at which the crankshaft resonated, and in
this year a method of attaching the clutch via a short
Clutch
Fig. 17
Crankshaft
Fig. 18
Fastened clutch
Input shaft
Clutch
Transmission with input shaft clutch
Heavy
metal
Dowel pin
M9 Bolt
Former weight
Fig. 19
61
Tightened weight
Crankshaft counterweight history
was machined in the lower section of the fillet R. This

prevented the occurrence of cracks in the fillet R,
resulting in no further incidence of fracture in the
crankshaft unit (Fig. 20).
The stability of the volume and pressure of the oil
supply from the crankshaft pin journals has an effect on
the lifespan of the conrod bearings.
In Formula One engines, oil is not supplied to the
crankshaft from the main journals, as in mass production
vehicles. A center oil supply method is employed, in
which oil is supplied from the front end of the crankshaft
(Fig. 21).
The supply of oil from the main journals as in mass
production vehicles would supply oil against centrifugal
force. In the case of a high-speed crankshaft, this results
in significant loss due to pressure loss and leaks. By
contrast, center oil supply, in which oil is supplied from
the center of the front end of the crankshaft, is a highly
efficient method, using centrifugal force to pressurize the
oil. However, even when using a center oil supply
method, it is only the first throw to which oil can be
supplied without resistance. The oil sent to the later
throws will to a certain extent be supplied against
centrifugal force, necessitating a layout in which the oil
channels deviate as little as possible from the center of
the crankshaft. In addition, the use of a center oil supply
method does not necessarily resolve all oil supply issues.

Pin metal seizures also continued to occur as a result
of factors including momentary oil supply abnormalities,
due to instability of supply from the oil tanks, and oil
pressure pulsations, due to the pin oil supply holes being
struck by the bearings.
Development of a hollow crankshaft was commenced
in order to achieve a weight reduction of 1 kg or more
and comprehensively resolve these issues. Friction
welding was employed to help enable the main journal
of the crankshaft to be made hollow, after which it was
able to function as a combined oil channel, oil tank, and
damping chamber (Fig. 22).
The hollow crankshaft exceeded initial predictions in
helping to increase oil supply. The instantaneous
minimum oil pressure that had previously been negative
due to oil pressure pulsation returned to 500 kPa, and
prospects for durability were also good. However, in
2006 the regulations prohibited the use of the welded
crankshaft, and the new crankshaft was never employed
in races.
9. Conclusion
(1) In the engine development process conducted during
Hondas third Formula One era, development efforts
related to the reciprocating parts and crankshafts, which
can be considered the core of the engine, produced
technologies that helped to enable the engine to satisfy
the stipulation for a lifespan of 1500 km at an upshift
engine speed of 19600 rpm, without having to protect
the parts through retardation of the ignition timing.
(2) The achievement of weight savings, reduction of sliding
resistance, and modification of the lubrication method
helped to reduce friction and realize a high level of
durability and reliability.
(3) During the third Formula One era development process,
designers became able to formulate 3D models and
conduct CAE analyses, helping to enable the realization
of greater accuracy in initial designs and more rapid
response to issues. The pistons and bearings are crucial
to the achievement of durability and reliability, and the
ability to reflect the results of analyses of cooling and
heat dissipation performance in designs in particular
helped to enable responses to be effected to demands for
increased lifespans.
(4) While the individual technologies developed for Formula
One engines and the development methods employed
Original point of crack
Stress relief groove
Fig. 20
Crankshaft stress relief groove
Oil
inlet
Fig. 21
Crankshaft center oil supply passage
Friction welded
Fig. 22
Hollow crankshaft
62
will not necessarily be used in mass production vehicles,

there are numerous points of similarity in terms of the
realization of high levels of efficiency and reliability.
The authors hope that this paper will function as a
reference that provides support for future development
projects.
Author
Tetsuo GOTOU
63
Nagao YANAGISAWA
Shuichi HAYAKAWA*
Kazushi OGIYAMA*
Masanori TATE*
ABSTRACT
From 2000, the development teams involved in Hondas Formula One engine development program worked as
one in development efforts focused on the achievement of increased power through increased engine speed. As part
of these efforts, the configuration of the valvetrain was reexamined from the bottom up, and new mechanisms were
developed.
To enable the realization of high speed and high lift in the valvetrain, a finger follower system was employed as
the valve drive method in place of a bucket tappet system. In addition, the application of a measurement method
applying Bezier curves to the acceleration of the valve lift curves and the optimization of the gear train reduced
fluctuations in the angular velocity of the camshafts and enhanced valve motion.
Furthermore, it was necessary to reduce the excessive friction due to increased engine speed. In response to this,
the pneumatic valve return system was advanced by reexamining the configuration of the cylinder heads. The
development of electronically-controlled regulator system also contributed to reducing friction.
1. Introduction
In response to this trend, a jet-type air control system

(J-valve mechanism) was newly developed to modify the
pneumatic valve return system (PVRS). This system
enabled oil agitation resistance in the PVRS cylinder to
be reduced, resulting in reduced motored friction of the
valvetrain.
Looking back on development efforts from 2000
onwards, this paper will provide an overview of
technologies developed to increase engine speed and
reduce friction in Formula One valvetrains with the aim
of achieving increased power, and will discuss the
performance of these technologies.
Competition in technological development is

particularly intense in Formula One, the pinnacle of the
worlds auto races. Merely maintaining competitiveness
represents technological stagnation or regression
constant advances are demanded.
In the development of valvetrains for Formula
One engines to respond to these demands, increased
speed and higher lift are sought in order to increase
power, but this generates the issue of controlling
increases in friction.
With increased power and higher lift as the targets,
the valve drive mechanism was redesigned, new
materials were employed, and surface treatments such as
diamond-like carbon (DLC) were applied, in addition to
the achievement of maximal weight savings in the
moving parts. Moreover, efforts including reexamination
of the valve lift characteristic and the gear train layout
reduced lift load at the same time as enhancing the
motion of the valves in the high-speed range, thus
increasing the allowable speed for the valvetrain.
Changes in Formula One regulations placed
restrictions on the materials able to be used, set limits
on engine speed, and stipulated that components must be
capable of long mileage. This made the reduction of
frictional losses in the engines an important issue, and
resulted in a significant change in the direction of
development efforts.
2. Overview of Valvetrain
Table 1 shows the main specifications of the
valvetrain used in Hondas 2008 Formula One engine,
and Fig. 1 shows an external view of the components
of the valvetrain.
A variety of studies were conducted of the valve,
including those on the angle of the intake and exhaust
valves from the perspectives of increased volumetric
efficiency and enhanced combustion. For the intake side,
a compound valve layout that also displays an angle of
inclination in the longitudinal direction of the engine was
adopted. The intake and exhaust valves were the
components in which the maximum weight savings were
sought, and development proceeded on materials for
valves so that weight could be minimized within the ratio

64

Table 1
Valvetrain main specifications
Year
Engine code
Head
Control
Air bottle
PVRS
Gear train
Lift , valve timing
Material
IN
Surface treatment
Cam
Lift , valve timing
Material
EX
Surface treatment
Stem diameter
Valve diameter
IN
Material
Surface treatment
Valve
Stem diameter
Valve diameter
EX
Material
Surface treatment
Material
Finger follower IN/EX
Surface treatment
Air spring seal
IN/EX
IN
Reciprocating mass (g)
EX
Maximum engine speed for valvetrain (rpm)
Upshift engine speed (rpm)
for supply pressure and discharge pressure.

Torque was transmitted by the gears from the crank
to the camshaft. This helped prevent delay in the valve
timing due to high drive torque and resonance generated
in the chain or the belt. The gear train was designed to
reduce fluctuations in the angular velocity of the
camshaft caused by vibration, thus enhancing valve
motion and reducing friction.
2008
Honda RA808E
J-Valve
P2/P3EAR control
Carbon bottle 570 cm3
AVRS-G
13.5 mm 19/64
SCM435 nitriding
DLC
13.0 mm 19/64
SCM435 nitriding
DLC
5.8 (hollow stem 3.5)
41.6
Ti6246
DLC
5.8
32.4
KS64411Ta
DLC
SNCM815 carburizing
DLC
PTFE
50.0
45.6
20300
19000
3. Technologies for Increased Engine Speed

3.1. Changes in Speed Targets
Figure 3 shows changes in the maximum allowable
engine speed for valvetrain performance in the Honda
Formula One engine development process from 1990 to
2008. In the valvetrain, the seating load increases rapidly
due to valve bounce at a higher engine speed. This might
reduce durability and reliability of each component of
the valvetrain, resulting in a failure. In response to this,
valve motion should be enhanced so that valve bounce
would not occur even at the maximum engine speed. The
maximum allowable engine speed for valvetrain
performance can be defined as the limit speed up to
which valve bounce would not occur, taking the overrevving projected to occur in the circuit driving
environment into consideration. At the commencement
of development efforts, increases of approximately 400
rpm per year were achieved, and this contributed to
increased power. From 2004, regulations specified long
mileage, and the mileage for each engine consequently
increased. From 2007, maximum engine speed was
limited to 19000 rpm, increasing the frequency of wide
open operation of the throttle valves. These changes
made increased valvetrain durability and reliability
necessary, and development therefore proceeded in order
to achieve a maximum engine speed for valvetrain
performance of 20000 rpm or more.
range of specific weight to stiffness as stipulated in the

Formula One regulations.
A camshaft and finger followers was used for the
opening and closure of the valves, and a DLC coating
was applied to the sliding components. In addition, the
application of a PVRS using compressed air in place of
normal metal springs reduced reciprocating mass, and
prevented the surging at high speeds that originated in
the metal springs. Figure 2 shows a view of the PVRS
as fitted in the vehicle. The system incorporated an air
bottle filled with high-pressure air and two regulators,
3.2. Reduction of Reciprocating Mass

Figure 4 shows a cutaway view of the valvetrain and
the reciprocating mass of each component of the
valvetrain. Weight savings were achieved by the
application of methods including the use of a finger
3.5 L
2.4 L
V8
View of valvetrain
Air bottle
V10
2000
Engine speed (rpm)
Fig. 1
3.0 L
Engine speed
limitation
V12
V10
Material
limitation
Mileage guarantee
420 km
400 km
2008
2007
2006
2005
2004
2003
2002
2001
2000
1992
1991
1990
Year
Air regulator
Fig. 3
Fig. 2
800 km
1350 km
1500 km
View of air spring system
65
Changes in maximum allowable engine speed

for valvetrain
follower mechanism to drive the valves, instead of a

bucker tappet system, and the reduction of the diameter
of the PVRS cylinder. In addition, efforts in materials
development and the use of CAE enabled the durability
and reliability of the valves to be ensured with the
minimum weight increase, despite restrictions on the
materials that could be employed and the requirement for
long mileage. These efforts reduced the reciprocating
mass of the intake-side valvetrain by 21.7 g (30%)
between 2000 and 2008.
From the perspective of sliding performance, a DLC

coating was applied to the cam lobe and the finger
follower top pads in order to help ensure durability and
reliability at high speeds and high hertz stresse levels, and
to reduce friction. In the development of the DLC coating,
factors, including the balance of film hardness between
the parts and the formation of the films, were optimized.
By comparison with the initial stage of DLC development,
an increase in surface pressure limit of 17% in high-load
operating environments was achieved, and friction was
reduced by approximately 2.0 kW per year.
From the perspective of finger follower geometry,
positioning the camshafts almost exactly on the stem of
the intake and exhaust valves brought the lever ratio
close to 1 and reduced surface pressure, in addition to
enhancing valve motion by increasing stiffness.
Furthermore, positioning the pivot of the finger follower
on the same axis as that of the intake and exhaust valves
enabled the finger followers to be lengthened and helped
to ensure the formation of an oil film on the sliding
surfaces.
The oscillation of the finger followers and the inertial
load on the arms generated a large moment on the valve
stems, raising concerns over bending of the valve stems
or fracture of the valve heads. Minimization of the
bending moment on the valve stems was an effective
means of responding to these concerns. The curved shape
for the bottom pads was therefore designed to enable the
amount of movement of the contact points with the valve
stems to be reduced and their positions to be located in
the center of the valve stems.
3.2.1. Development of finger followers

The drive system for the valves in a Formula One
engine must use lightweight parts in order to achieve
increased power, and at the same time must enable
increased valve lift in order to increase the volume of
intake air. Using the previous bucket tappet system, it
was necessary to increase the diameter of the bucket
tappet to achieve higher lift, and it was challenging to
balance this necessity with the achievement of weight
savings. Because of this, finger followers were employed
from 2002 onwards, enabling weight to be reduced by
13.5 g and valve lift to be increased by 1.0 mm or more.
Figure 5 shows a comparison of the valve drive systems
used in Honda Formula One cars.
The shape of the finger followers was determined
through studies using 3D models and CAE based on a
variety of considerations, including frictional losses at
high speeds and high-hertz stress levels as well as the
local stiffness and strength depending on inertial loads
and the direction of rotation of the cam.
Camshaft
Shim
PVRS
piston
Cotter
Seal
Valve
80
Reciprocating mass (g)
Finger
follower
3.2.2. Development of intake and exhaust valves

In order to increase volumetric efficiency in the
Formula One engine, it was necessary to increase the
diameter of the head sections of the intake and exhaust
valves while reducing the diameter of the valve stems.
It was also necessary for the valves to be capable of
withstanding the inertial forces and seating loads
generated by acceleration of more than 6000 G under
combustion gas temperatures of 700 C or more.
Considering these demands, materials possessing high
heat resistance strength were developed, and CAE was
employed to enable selection of appropriate valve shapes
(Fig. 6). Combustion gases and intake air generated a
Others
PVRS piston
Valve
Finger follower
70
60
50
40
30
20
10
0
2000
Fig. 4
2002
Year
2008
Reciprocating mass of parts
2001 model
2008 model
Tensile stress
by inertia
Tensile stress
by seating
Thermal stress
Fig. 5
Comparison of valvetrain layouts
Fig. 6
66
CAE for inlet valve
Inlet valve
shape
50 mm/rad2 or more.
In the design of the valvetrain, considerations of
reliability and durability and the realization of increased
speed made it essential to set a level of valve lift
acceleration that considered the achievement of a balance
between the lift load for the PVRS and inertial forces.
A Bezier curve was therefore applied to the lift curve.
This type of curve enables designers to freely set valve
lift acceleration, and at the same time, vary the lift curve
in real time, simplifying the realization of the desired lift
curve. By this means, the negative acceleration of valve
lift was optimized with consideration of the PVRS
pressure characteristic, enabling the valve lift load and
inertial forces to be balanced to the limit point and the
maximum allowable speed for the valvetrain to be
increased by 1000 rpm. In addition, in order to balance
increased lift with a high compression ratio, the positive
acceleration of the valve lift was designed to display two
stages, with the level of lift being reduced only when
the clearance between the pistons and the valves was
low. This enabled the compression ratio to be increased
by 0.3. Figure 8 shows the two-stage acceleration
designed using the Bezier curve.
significant difference in ambient temperature between the

fireface and the back of the intake valves, and there were
concerns that high levels of heat stress might be
generated on the exteriors of the heads, resulting in seat
chipping. Analyses were therefore conducted of stresses
generated by heat and by inertial forces, and an umbrella
shape was developed that would prevent the level of
stress from exceeding the fatigue strength of the material
during operation. The swaying of the valves at high
speeds also generated partial contacts when the valves
were seated on the valve seats. This caused stress to be
concentrated at the points of connection between the
valve heads and valve stems. As a countermeasure, the
valve stems were tapered in order to reduce stress.
Honda began using titanium aluminum (TiAl)
materials in 2002, and employed them in race vehicles
until 2005. The use of these materials enabled the
diameter of the valve stems to be reduced from 6.6 to
4.5 and the reciprocating mass of the valves to be
reduced by 5.0 g against conventional titanium (Ti)
materials. TiAl alloys are intermetallic compounds, and
the particular machining employed for these materials
produces cracks around which residual compressive
stress exists. Because of this, the valves frequently broke
during their first period of use. In order to resolve this
quality issue, quality control was implemented by
subjecting all valves to eddy current testing rather than
fluorescent penetrant inspection in order to detect cracks,
enabling the realization of a level of durability and
reliability high enough to meet the standards of the long
mileage regulations.
From 2006, restrictions were placed on materials in
Formula One, and Honda had no option but to change
its materials. It was necessary to hollow the valve
umbrellas and valve stems in order to achieve an
identical or lower weight using Ti as had been possible
using TiAl. However, in the initial stages, insufficient
strength and stiffness resulted in frequent breakages of
the valve heads. Measurements of the stress on the valve
heads and the temperature of the heads were therefore
taken, and analyses were conducted of factors including
the motion of the valves during firing. The optimal
umbrella shape was selected by feeding the results back
into CAE, enabling the realization of a low weight and
a high level of stiffness, equivalent to those available
from TiAl.
3.4. Reduction of Fluctuation in Cam Angular Velocity

Torsional vibration was one of the factors that had a
negative impact on valve motion. This generated a
nonlinear resonance phenomenon in the camshafts.
Fluctuation in the angular velocity of the camshafts is
thought to occur due to fluctuations in the motion of the
camshafts as a whole resulting from resonance with the
20
60
45
10
30
15
0
0
45
90
135
180
225
-5
0
270
-15
-10
-30
Speed
Acceleration
-15
-45
-20
14
Inlet lift curve of 2008 model
Bezier acceleration
Previous acceleration
40
30
10
20
Lift (mm)
12
10
0
Difference in lift
between previous
acceleration and Bezier
4
2
0
100
125
150
175
Cam angle (deg)
Fig. 8
67
-60
Cam angle [deg]
Fig. 7
3.3. Optimization of Cam Profile

The cam profile has a wide range of effects on
performance. The achievement of increased volumetric
efficiency by increasing valve lift and the optimization
of the duration angle and the opening and closing timing
of the valves enables low- and medium-speed torque to
be increased. The stabilization of combustion increases
drivability, and the realization of a higher compression
ratio increases fuel efficiency. Figure 7 shows the lift
curve for the intake valves employed in Hondas 2008
Formula One engine. Study of the valve lift curves of
Formula One engines using 3D v simulations showed
that the positive acceleration of the lift curves reached
Acceleration [mm/rad2]
15
200
Bezier lift curve
-10
-20
-30
225
Acceleration (mm/rad2)
Lift [mm], Speed [mm/rad]
Lift
gear train, and the twisting of the camshafts between

cam lobes due to the reaction force of the valve lift. It
instantaneously increases or decreases the speed of
rotation of the cams, reducing the motion of the valves.
This makes it necessary to reduce the fluctuation of the
angular velocity of the camshafts at high speeds. The
following factors can be indicated as reducing
fluctuations in the angular velocity of the cam:
(1) The drivetrain moment of inertia
(2) The drivetrain spring constant
(3) Gear backlash
(4) Cam drive torque
(5) The damping coefficient of the dampers, etc.
Given that the cam drive torque is determined
prioritizing vehicle dynamics performance in a Formula
One engine, the remaining four factors were focused on.
In addition, it was essential to reduce variation in the
angular velocity of the cams without increasing the
weight of the engine or raising its center of gravity, two
factors that have a significant effect on vehicle body
performance.
in order to control fluctuation of the angular velocity

of the camshafts in the high-speed range increased the
valve motion speed by 1500 rpm, and reduced friction
by 6 kW.
3.4.2. Angular Velocity Reduction System with Gears
(AVRS-G)
Figure 11 shows the configuration of the AVRS-G.
Figure 12 shows fluctuations in the angular velocity of
the camshafts. When the AVRS was employed, the
fluctuation in the angular velocity of the rear ends of the
camshafts became higher than that of the front ends, and
there was therefore a margin for further reduction of
fluctuations. The issue was considered to originate in the
fact that the rear ends of the camshafts to which the
weights were affixed were the free ends of the shafts.
The AVRS-G was therefore developed. This mechanism
employed gears to connect both the front and rear ends
of the intake and exhaust camshafts, and increased the
spring constant of the camshafts in their entirety. The
use of the AVRS-G reduced the difference in twisting
between the front and rear ends of the camshafts at the
same time as reducing fluctuations in angular velocity,
thus assisting in reducing the stiffness of the camshafts.
This enabled the high specific gravity tungsten alloy
used for the rear weights in the AVRS mechanism to be
replaced by steel rear gears and the hollow diameter of
the camshaft to be increased, resulting in a weight saving
of 2 kg around the camshaft while maintaining identical
valve motion.
3.4.1. Angular Velocity Reduction System (AVRS)

The AVRS was a mechanism that reduced
fluctuation in the angular velocity of the cams by
attaching weights to the rear ends of the camshafts (on
the side opposite the gear train) in order to increase the
moment of inertia and shift the resonant frequency of
rotational vibration to the low-speed range. Figure 9
shows the configuration of the AVRS and Fig. 10
shows its effect in reducing fluctuation in the angular
velocity of the cams.
The use of weights to increase the moment of inertia
of the camshafts and the adjustment of gear backlash
3.4.3. Cam damper

While never employed in racing due to Hondas
withdrawal from Formula One, the achievement of even
Rear mass
Camshaft
Rear gear
Cam gear
Head idler gear
Fig. 9
View of AVRS
Fig. 11
2000
Conventional valve train
2000
Angular velocity
[deg/sec]
2000
2000
Angular velocity
[deg/sec]
View of AVRS-G
AVRS
AVRS-G
AVRS
Engine speed [rpm]
Fig. 10
Engine speed [rpm]
Fig. 12
Angular velocity of AVRS (2002 engine)
68
Angular velocity of AVRS-G (2003 engine)
higher lift was studied in order to further increase power,

making it necessary to correct a resulting decline in
valve motion. As a means of achieving this goal, cam
quills with a damping effect were developed to be fitted
between the cam gears and the camshafts, and viscous
dampers were attached to the rear ends of the camshafts.
Figure 13 shows the configuration of the cam damper
system, and Fig. 14 shows its effect in reducing
fluctuations in angular velocity.
The stiffness of the quill shafts connecting the cam
gears and camshafts and the viscosity of the silicon
encased in the viscous damper were adjusted to achieve
the desired damping coefficient. Cam drive torque and
resonant frequency were calculated for each
specification, and specifications that did not produce
extreme peaks of fluctuation in angular velocity in
simulations were employed.
The combined use of the quills and the viscous
dampers reduced fluctuations in angular velocity, and
increased the valve motion speed by 1100 rpm against
the AVRS-G.
valve motion and reduced lift load. In addition, the

reduction of the level of oil in the PVRS cylinders was
studied, and in 2005 the employment of a J-valve
mechanism in an evolution of the PVRS enabled the oil
level to be reduced to zero. These advances reduced
friction by 20 kW in the 2005 V10 engine.
4.2. J-valve Mechanism
4.2.1. Aim of development of J-valve mechanism
PVRS is generally employed in the valvetrains of
Formula One engines, which are high-speed engines, and
Honda also employed the system from 1992.
Conventional PVRS used a check/relief mechanism,
and supplied and discharged air with one-way valves
fitted at the inlet and outlet of the PVRS cylinder to
control pressure. However, the volume of air that could
be used on a race was determined by the capacity of the
bottle fitted in the vehicle, and it was necessary to
control the volume of air consumption from the outlet.
To enable this, a system was designed in which oil
accumulated in the PVRS cylinder, and the discharge of
the oil was prioritized over the discharge of the air.
However, the oil generated agitation resistance during
valve lift, and became a factor in increased friction.
Figure 16 shows the differences between the check/
relief and J-valve mechanisms. Figure 17 shows the
correlation between the level of oil in the cylinder,
friction, and air consumption. The newly developed Jvalve mechanism controlled air consumption by using a
single orifice for supply and discharge of air. By
separating air consumption from the production of
Cam quill
Viscous damper
Cam damper system
Angular velocity
[deg/sec]
2000
2000
2.4 L V 8
Engine speed
limitation
V10
Material
limitation
5.0
Fig. 13
Valvetrain friction (kW)
3.0L V 10
V10 engine
V8 engine
V8
Mileage guarantee
1350 km
2008
2007
1500 km
2006
2003
2002
800 km
Year
Fig. 15
Engine speed [rpm]
Fig. 14
2001
2000
Cam damper
420 km
2005
400 km
2004
AVRS-G
Valvetrain friction performance
Check/relief
Angular velocity of cam damper (2008 engine)

Check valve
J-Valve
Relief valve
4. Friction Reduction Technologies
Orifice jet
4.1. Progress in Friction Reduction

Figure 15 shows the progression of reduction of the
motored friction of valvetrains in Honda Formula One
engine developments from 2000 to 2008.
In the initial stage of development for Formula One,
the reduction of reciprocating mass and the control of
fluctuations in the angular velocity of the cam enhanced
Oil
:Air flow
Fig. 16
69
Comparison of PVRS layout
of the stem seals was increased by reexamining the

configuration of the cylinder heads and employing a
layout that enabled forced oil supply only to the seals.
friction, this mechanism obviated the necessity for oil in

the PVRS cylinder, and reduced friction by 3 kW. In
addition, doing away with the one-way valves and the
air channels on the cylinder heads resulted in a weight
saving of 1 kg.
4.2.3. Control
The J-valve mechanism was fundamentally designed
as a configuration in which there was no oil inflow to
the PVRS cylinder. However, the seals did allow tiny
amounts of oil into the cylinder, which would increase
the internal pressure of the cylinder if not dealt with. An
increase in the internal pressure of the cylinder would
lead to increased friction, and in addition seal ruptures
and sliding abnormalities could occur and lead to engine
troubles. In addition, the volume of oil inflow to each
PVRS cylinder was unknown. To respond to this issue,
a system of air line purge control was developed
enabling forced discharge of the oil that had accumulated
in the PVRS cylinder. Air line purge control enabled
variable control of the pressure in the PVRS cylinder by
means of electronic control of each EAR unit, and forced
air flow through the air channels. Figure 20 shows a
diagram of the P2 EAR unit.
The forced discharge of oil in the air line purge was
conducted by increasing the pressure on the inlet side
4.2.2. Configuration of J-valve mechanism

Figure 18 shows the system configuration of the Jvalve mechanism. One section of the system incorporates
an air bottle, a primary decompression mechanism, and
an electric air regulator designated P2 (P2 EAR), which
regulates the supply pressure via an air injector. Air is
supplied to the engine from this section. The outlet side
is provided with an electric air regulator designated P3
(P3 EAR), which similarly regulates discharge pressure
using an air injector. The development of these two EAR
units enabled variable control of pressure.
Figure 19 shows a section view of the PVRS
cylinder. Two concerns in the development of the J-valve
mechanism were: 1) the feasibility of a configuration of
the seal rings that helps prevent oil from flowing into
the cylinder; 2) the durability and reliability of the stem
seals in the absence of oil accumulation. To respond to
these concerns, the seal rings were formed from PTFE
jackets with a ripple shape and rubber rings with an X
shape. This produced a stable tensile force and enhanced
oil discharge performance. The durability and reliability
200
3.0 kW
-2.0
Limit of air consumption
-1.0
0.0
1.0
2.0
3.0
4.0
10
8
6
4
2
0
-2
-4
-6
-8
-10
Friction (kW)
Air consumption (cm3/km)
Setup of Check/relief
PVRS piston
PVRS liner
Jacket
(seal ring)
Quad ring
(seal ring)
5.0
Oil line
Oil in PVRS chamber (cm3)
Fig. 19
Friction of Check/relief system
Friction of J-Valve system
Air consumption of Check/relief system
Air consumption of J-Valve system
Fig. 17
Section view of PVRS cylinder
Relation between air consumption and friction

Air bottle
Air return OAS
P3 EAR assembly
Front
IN
IN
EX
Air injector
Air injector
Oil tank
EX
Air bottle
P3 sensor
P1R sensor
P2 sensor
Air charge
Primary pressure-reduction parts

(diaphragm-type)
Rear
From engine (P3)
Air injector
To engine (P2)
P2 EAR assembly
Fig. 18
P1 sensor
Mechanical
regulator
Fig. 20
J-Valve system configuration
70
View of P2 EAR
and creating a pressure differential with the outlet side.

In order to achieve complete oil discharge, it was
necessary to create a pressure differential that was able
to generate an air flow speed of 255 L/min, and to use
a volume of 8 L or more of air. The air line purge was
automatically conducted when a preset number of laps
(distance) had been covered, and the oil was periodically
discharged from the PVRS cylinder throughout the race.
The air line purge could also be conducted manually in
the event of abnormal functioning or during pit work.
When long mileage regulations for Formula One
engines came into effect, the amount of use of the idling
range (5500 rpm and below) increased. Because the
inertial load in the valvetrain is low in the idling range
and the lift load becomes dominant, cam surface pressure
is high. This raised durability and reliability concerns
over the sliding of the cam. To address this issue, a
control method that helped prevent excessive lift loads
in the idling range was developed which reduced the
surface pressure of the cams and helped to ensure
reliability and durability. Lift loads during idling were
reduced by reducing the air channel pressure to a level
lower than that of during race driving, using the P3 EAR
unit. Figure 21 shows an overview of lift load switching
control during idling.
(3) The reduction of valve lift load through the enhancement

of valve motion, the reduction of frictional losses through
the use of a DLC surface treatment, and the reduction of
oil agitation resistance through the use of a J-valve
mechanism helped to reduce motored friction of
valvetrain by 23 kW at an engine speed of 18000 rpm.
High engine speed

Engine speed
Maintenance of speed
Low engine speed
Delay
time
High
Low
Enhancement of valve motion

Target air-line
pressure
Decline in cam surface pressure
Time
Fig. 21
Pressure control during idling
5. Conclusion
During Hondas Formula One engine development
program between 2000 and 2008, the company
developed a valvetrain reconciling the achievement of
high speed with low friction. The following results were
achieved:
(1) Finger followers were employed as the valve drive
method, and materials development was conducted and
measurement technologies and CAE introduced to
respond to materials and long mileage regulations,
resulting in a 21.7 g (30%) reduction of the reciprocating
mass of the valvetrain.
(2) To respond to the necessity for increased power in order
to maintain competitiveness, valve lift acceleration and
the gear train layout were reexamined, enabling the
allowable speed for the valvetrain to be increased by 2700
rpm.
Author
Shuichi HAYAKAWA
71
Kazushi OGIYAMA
Masanori TATE
Development of Induction and Exhaust Systems

for Third-Era Honda Formula One Engines
Ken NISHIMORI*
Masayoshi TAKAHASHI*
Yasuhiro MOTOHASHI*
Ryuichi FURUKAWA*
ABSTRACT
Induction and exhaust systems determine the amount of air intake supplied to the engine, and as such are critical
elements affecting engine output.
In addition, the layout of the induction and exhaust systems affects the vehicles aerodynamic performance, and
so it must be considered together with vehicle development.
At first, there were few CAE software and computer resources available, and induction and exhaust system
components were produced by measurement and guesswork so that development was largely performed on a trial and
error basis, but in recent years, the 3D-CAD and CAE software has advanced so quickly, and computer resources
have expanded so much, that development is done by simulation.
The enhanced phenomenon elucidation and forecast precision have made it possible to shorten the time it takes to
determine specifications and reduce development costs.
1. Introduction
the driver, the engine weight accounts for a large

percentage of the total, so when designing engine
components, one has to make them light while
maintaining their necessary functions.
The development of induction and exhaust systems
was mainly about lowering flow resistance in 2000,
when the initial development was being done, but since
2006, it has been possible to predict dynamic effects
during the process of design examination, meaning that
one can carry out optimal design, with consideration of
vehicle package, in a short amount of time. This paper
discusses the content of this development.
The first thing required of a racing engine is power

performance, but transient characteristics and vehicle
package also impact performance. As shown in Fig. 1,
the induction and exhaust systems are placed outside the
engine, and have a large impact on the package.
Vehicle weight and inertia influence dynamic
performance, but at a vehicle weight of 600 kg including
2. Induction
Fig. 1
2.1. Induction System

2.1.1. Reducing flow resistance
One technique for increasing volumetric charging
efficiency is to reduce induction system pressure loss
(below, pressure loss). The main factors relating to
pressure loss are air filter performance and airbox
(below, ABX) form.
Sometimes impurities drift in when the ABX is
opened after running, so an air filter is required to avoid
having to retire from the race. Because it is located
within the air intake passage, however, the filters flow
resistance directly affects intake resistance of the engine.
Air filters, therefore, must be designed to reduce flow
resistance.
Engine with complete induction and exhaust

72
Hondas third-era Formula One engines originally

used a sponge-type air filter from an overseas
manufacturer, but in 2004 Honda began examining a
nonwoven fabric-type air filter. In 2004 a Grand Prix
event was held in Bahrain, in a dusty area, and the
sponge-type filter was not able to trap the fine desert
sand and presented the crew with a big challenge.
At first, they used a dry filter that was installed on
mass-market vehicles in dusty areas. However, although
this increased the scavenging ratio, pressure loss also
increased and power dropped by 4 kW, so the use of this
filter was limited to certain events.
Subsequently, an investigation was begun on how to
achieve enhanced engine durability, reliability and a
higher scavenging ratio without diminishing
performance.
A wet, nonwoven fabric filter was developed with
high scavenging ratio and low pressure loss, which was
used starting with the Italian Grand Prix, one of the last
events of the season.
Starting in 2006, engines were limited to eight
cylinders, which led to frequent induction fires caused
by backfiring. Similar phenomena were also seen in
CART series V8 engines, along with the phenomenon of
induction systems exploding and components scattering
in CART races where premixed methanol fuel were used.
In Formula One engines, the air filters suffered melting
damage (Fig. 2), and simultaneously with this seeping
into engines, CFRP components were burnt and vehicles
left unable to run.
Engine controls were changed as a countermeasure,
while the filter was changed to a highly fire-resistant
material.
After that, there were no more troubles in actual
driving from filters burning as a result of backfiring.
Because the flow resistance of an air filter is
proportional to the square of flow velocity, it is
important to evenly distribute and reduce the velocity of
the air-flow through the filter in order to limit pressure
loss.
Within the scope allowed by the layout and
aerodynamic performance, one needs an efficient
velocity transformation to the maximum filter area that
fits inside the engine cover.
Although it looks like nothing more than a container

that simply supplies air to the engine, the ABX is an
important component that takes air flowing at a relative
speed of 300 km/h and turns it into a homogenous, lowspeed flow of air to be sucked into the engine.
Figure 3 illustrates results of ABX CFD examination
in 2008.
To reduce flow velocity in a limited package, one has
to efficiently expand the air passage in cross-section. The
velocity of air taken into the ABX contains a large
horizontal constituent.
The layout of the ABX in the vehicle is such that the
aperture is high to protect the driver, the ABX is frontmounted on the engine to concentrate the mass, and there
are surfaces with much curvature in the front. These
front surfaces are subject to separation, which causes
variance in the amount of specific intake air volume to
each cylinder and a drop in performance. Therefore, the
form was optimized using CFD to enhance filter
efficiency.
At first, the ABX was developed on the assumption
that air is brought in to the ABX inlet homogenously.
But a wind-speed simulator (below, WSS)(1) implemented
in 2007 was used to simulate actual driving conditions
and measured pressure distribution on the inlet duct. This
made it clear that the distribution was not homogenous,
as illustrated in Fig. 4. The results confirmed that engine
power under these circumstances showed a 5 kW power
loss as compared to a homogenous condition.
Figure 5 shows CFD results at low vehicle speed.
The flow that stagnated at the lower end of the roll hoop
rises while containing a Z-direction constituent up to a
point close to the air intake and comes flowing into the
Fig. 3
Fig. 2
Damaged air filter
Fig. 4
73
ABX CFD result
Total pressure distribution at inlet to ABX
In a V10 engine, intake interference between the

banks causes the quadratic standing waves in the ABX
to amplify, with the result that the secondary constituent
of the pulsation in the port attenuates, interfering with
dynamic air intake.
A splitter is an effective way to block interference
between the banks. Figure 7(i) shows a splitter on the
2004 (V10) ABX. Adding on the splitter helped enhance
power by 7 kW.
The item indicated by the letter a in the figure is
a component installed to plug the small gap, which
boosted power by 2 kW. However, because of the
tolerance of this component and the variance that
occurred when it was installed, the gap could not be
completely plugged, and eliminating the impact the gap
had on power was a continuing issue.
Fig. 7(ii) shows the splitter used since 2006.
The changes made to the regulations in 2006 required
the use of V8 engines. In addition, the previously
described VIS could no longer be used, so the team was
required to expand the torque band with a fixed intake
pipe length.
The effect of inter-cylinder interference within the
induction system in a V8 engine is greater than that in
a V10 engine (Fig. 8), and torque characteristics change
depending on the form of the induction system. Because
of this, inter-bank interference of the four cylinders in
the center was controlled, while promoting active
interference in the four cylinders in front and back,
ABX. As a result, separation occurs on the lower surface

of the part where the ABX comes in (below, the
snorkel).
During development of the 2009 model, the air-flow
status of the ABX aperture was considered during
examination of vehicle aerodynamics, and development
went forward using CFD to make inlet pressure
distribution almost uniform.
2.1.2. Using intake pulsation
In an induction system, the intake pulsation caused
by the opening and closing of the intake valve interacts
dynamically within individual cylinders, among cylinders
and between the cylinder banks. By controlling these
interactions, one can try to enhance engine performance.
Figure 6 illustrates a variable-length intake system
(below, VIS).
A VIS uses the fact that the intake pipe length has a
dominant effect on engine speeds at which resonant
supercharging effect can be achieved. By using this
system, intake pipe lengths are continuously achieved
that are appropriate for each engine speed. In addition,
corrections are made that correspond to changes in intake
temperature.
CFD result of ABX inlet on chassis (150 kph)
(i) 2004 ABX splitter
Fig. 7
(ii) Local splitter for V8
Schematic view of ABX splitter
300
Torque [Nm]
Fig. 5
250
200
Without ABX and without splitter
With ABX and without splitter
With ABX and with splitter
150
6000
Fig. 6
8000
10000
Fig. 8
Schematic view of VIS system
74
12000 14000 16000

Engine speed [rpm]
Effect of splitters
18000
20000
which increases volumetric charging efficiency (below,

v) over a wide range of engine speeds and flattens
torque characteristics.
Figure 9 shows a test ABX for adjusting intake
pulsation. Inserts made by a rapid prototyping machine
are attached to and used on the interior of an ABX built
to a large size. The ABX is made with lightweight CFRP
to ensure rigidity in those areas where the inserts are
attached.
The form-selection process switched to testing with
the above test ABX, which made it possible to find the
correlation between ABX wall form and power
characteristics in a short amount of time. They were also
able to be validated on 3Dv simulation(2).
Examination of the making of inserts by rapid
prototyping machine can be done until just before the
test, and furthermore, the replacement with an optional
part taking account of test results could be performed
quickly.
Also, correlation with 3Dv simulation was
conducted simultaneously, with enough precision to
forecast the power-performance benefit at the design
stage.
was revised with the aim of enhancing reliability, while

using the sensor in the nose as a reference point.
In addition, an air inlet was added to the engine cover
to allow fresh-air induction (Fig. 10), and circuit tests
were thus conducted.
Because the engine cover has a high flow rate on the
surface and a high internal pressure, a reverse flow of
hot air from within the engine cover occurs if
unmodified. It was not possible to take countermeasures
in 2008. In the design of the 2009 vehicle, however, this
item was incorporated from the start, and a design was
chosen that suppressed this temperature increase inside
the engine cover.
2.2. Inlet Port
Points to watch in the design of inlet ports include:
Charging stroke
Fuel distribution
Suctorial dynamic effect
Charging stroke and suctorial dynamic effect impact
the absolute air-flow rate, while fuel distribution and
flow-velocity distribution impact combustion speed.
2.2.1. Reducing charging stroke
In cylinder head design, the following are done from
the time of layout to reduce charging stroke.
Reducing valve stem diameter
Adjusting valve form
Adjusting valve layout
Adjusting form close to valve seat
Smoothing-out component surface on inlet port
Adjusting port form
The valve head shape contributes to reducing flow
resistance.
The valve margin thickness and angle R close to the
valve seat face are adjusted, and a thin form of valve
margin thickness is used as the final form.
As the valve lifts, the effective area expands. If at
this time the distance between the cylinder wall and
valve face is short, the effective aperture between the
valve and seat cannot be used efficiently. Therefore, the
valve pitch was decided by taking into account the valve
faces position relative to the cylinder wall.
In addition, the exhaust valve face was set in a
2.1.3. Reduction of intake temperature

Reducing intake temperature is one technique for
increasing specific intake air volume.
In a Formula One engine, the injector is located in
the upper part of the intake pipe inside the ABX, and
the fuels latent heat of vaporization is used to reduce
intake temperature.
The injector height and angle are adjusted to
maximize this effect.
In addition, a great distance from the inlet valve seat
face enhances the temperature reduction benefit, but also
leads to decreased engine response and drivability, so
these considerations must be balanced.
It has been confirmed that intake temperature is
higher than ambient temperature, and that the
temperature difference between an air intake sensor
located in the induction system and one in the nose is
approximately 4C. However, it was found that the
actual intake temperature difference can be more than
4C, depending on the peak brake power revolution as
measured from drive-shaft torque in driving status. The
reason for this can be speculated to be due to heating
of the induction system by the increased temperature
inside the engine cover.
As a countermeasure to this, engine control mapping
Fig. 9
Picture of test ABX and inserts
Fig. 10
75
Additional air inlet on engine cover
position so that it would not interfere with air inducted

through the inlet.
The valve angle is one factor determining cylinder
head size. For the 2005 Formula One engine, the valve
angle was made small, and it was the most lightweight
of all Honda Formula One V10 engines.
However, engines with small valve angles cannot get
enough effective area relative to valve lift, which makes
it difficult to increase power even if one does achieve a
compact combustion chamber.
The engine in the first half of 2006 took the design
of the 2005 engine into account, and the valve placement
emulated that engines as well, but power could not be
increased for the above reasons, so the valve layout was
changed mid-season.
Experiments were done with changing to a valve
angle equivalent to that of 2004, but the ability to install
to the chassis for racing was a necessary condition for
making a layout change mid-season, so in fact, the
finalized specification was intermediate between 2004
and 2005.
Because of homologation, which froze development
of engines themselves after 2007, the valve layout was
maintained as is.
A compound valve layout is a technique to achieve
the compact combustion chambers needed to produce
good combustion.
In a parallel valve layout, because port form and flow
direction are arranged in relative conformity to each
other, it requires no great effort to set a port form
immediately over the seat. During the initial examination
of the compound valve layout, the seats upstream form
was mapped together with the valves in parallel, and the
effective area relative to the parallel valve port used as
a base was reduced. However, because the actual flow
showed no direct correlation to the valve stem (Fig. 11),
the form was designed with no consideration for the
valve stem, thereby minimizing the loss in effective area.
As a result, power was increased by 4 - 6 kW, including
the valvetrain design.
In designing the port form, one chooses a form that
tries to control the separation that occurs on the side

towards the lower face inside the port (a). As for the
form of the upper face close to the seat, designers are
conscious of overall air flow. Since the master stream
of air flows in a different direction from that of the valve
stem, the part directly over the seat (b) is not expanded
(Fig. 12).
2.2.2. Fuel distribution
In the early stages of engine development in 2000,
the main focus was on reducing intake resistance. During
development in 2004, a port to control separation from
the inner face was examined as a technique to reduce
intake resistance. Expansion of port volume was
approved and v was expected to increase, but power
actually went down. The discrepancy expected in the
measured v could not be identified, and it was
considered that perhaps some factor other than mass
flow-rate was impacting power performance. It is
considered that, although this specification controls
separation, there is significant fuel adhering to the
internal wall as a consequence. Figure 13 shows results
of a check of fuel adhering to the port internal wall.
Where the color is darker in the figure, fuel is adhering
to the internal wall of the port. It can be confirmed that
ports with reduced intake resistance have more fuel
adherence.
(b)
(a)
Fig. 12
CFD result of 2004 inlet port

Lower
Upper
Base
Minimal flow resistance
Less fuel sticking
Fig. 11 Influence of shape around valve seat

(Left : Parallel Valve Right : Compound valve)
Fig. 13
76
Fuel sticking area with inlet port when

conscious of flow (middle)
By controlling fuel adherence, a power increase of 2

kW was achieved. It was confirmed that besides reducing
intake resistance, the volume of the inlet port and
volume allocation in sectional area also affect power. It
was simultaneously confirmed that the form of the port
is also related to the form of gas mixture in the cylinder.
possible to design an optimized induction system at the

earliest stages of development, and overall performance
of the induction system as a whole as well as
development efficiency have been enhanced.
To increase efficiency of inlet port development, the
method of filling ports with adhesives and again
applying CNC processing techniques was adopted in
2004. This offered the advantages of being able to
produce optional parts in a short time so they could be
put into use more quickly, and of being able to minimize
the effect of individual differences on engine
performance.
2.2.3. Dynamic effect

Following on the fact that v does not increase in
enlarged ports with reduced flow resistance, in the
middle of the 2004 season an investigation was initiated
into forms that would narrow-down the inlet port within
the range where flow resistance would not be decreased.
This concept accounted for 3 - 6 kW of power gain and
was used in the 2004 Japanese Grand Prix. Since then,
Honda models up to 2008 were given similar distribution
in sectional area.
Analysis testing with a single-cylinder engine was
done to accelerate port development. However, in
comparisons of Formula One engine port performance,
multi-cylinder engines (V8 with intake interference) and
single-cylinder engines (with no intake interference) have
sometimes shown completely opposite trends (Fig. 14).
The cause of this is intake interference between
cylinders in a V8 engine. Two approaches can be taken
as countermeasures: controlling intake interference with
the ABX; and optimizing ports in places where there is
intake interference.
During development of the 2009 engine, the inlet port
form was examined with the objective of promoting flow
within cylinders (tumbling). Although there was a trend
for the effective area to grow smaller, no difference was
found in performance, and no great discrepancy was
found in v at this time. These results could not be
explained just from the perspective of reducing charging
stroke.
3. Exhaust System
3.1. Exhaust Port
During the study of single-cylinder engines, it was
learned that flow rates from exhaust ports rose to nearly
the speed of sound and caused choking. In steady flow
tests, the flow velocity did not come close to the speed
of sound due to equipment capacity, and thus choking
was not recognized.
While in choke, flow depends on cross-sectional area
and the state of the fluid, so performance tests were
conducted after expanding the part with the smallest
cross-sectional area. Some enhancement of performance
was seen with the single-cylinder engine, but none was
observed in the V8 engine. In a single-cylinder engine,
choking occurs during the period of valve overlap from
blowdown, but in the V8 engine, there is a range where
pressure within the exhaust pipe is high compared to that
of a single-cylinder engine because of the effect of intercylinder interference, so it is believed that a high
pressure differential such as would cause choking does
not exist, so the same effect could not be achieved.
The cause of this is that, at exhaust-pipe diameters
that would be realistic on board a vehicle, it is not
possible to produce a diffuser effect that would cause
choking of the throat of the exhaust port of a V8 engine,
as shown in Fig. 15. However, by using expanding pipe
that brings flow velocity close to the speed of sound
within the exhaust port, staying within the range of
equipment that can be mounted on board and without
expanding port diameter, performance could be enhanced
at high engine speeds.
2.2.4. Inlet port development techniques

When designing an engine, the platform is first
decided, then the form of the ports included in this
platform is designed, and finally the induction system is
designed.
However, advances in 3D CAD and CFD have now
made it possible to predict induction system performance
with high precision at the design stage. It is therefore
10
8
1.4
1.2
12000
13000
14000
15000
16000
17000
18000
19000
-4
-6
-8
V8 on dyno
Single cylinder engine on dyno
(correlated to V8)
Exhaust valve lift
Mach number
Ps [kW]
2
0
Mach number at throat of exhaust port (V8)

Exhaust valve lift
-2
Mach number at throat of exhaust port (single cylinder)
1.0
0.8
0.6
0.4
0.2
0
-10
Engine speed [rpm]
90
180
270
360
450
540
630
Crank angle [deg]
Fig. 14
Ps of inlet port on V8 and single cylinder

engines
Fig. 15
77
Calculated mass flow of exhaust port
720
3.2. Exhaust System

3.2.1. System overview
The exhaust system of a Formula One engine consists
of three parts: primaries, a collector and a tail. Figure
16 shows these in assembled form.
In exhaust systems used for testing on a dyno, the
three components are each configured separately, which
takes account of the need to change components because
of damage, and the need for tuning between the collector
and tail. In exhausts for actual vehicle testing and racing
use, however, the collector and tail are welded together
and are used as a single piece. Although making the
three components into a single piece would make the
exhaust lighter in weight, the primaries are kept separate
to enable spark plug maintenance, during which the
cylinder head covers are removed. The material used was
Inconel 625, with pipe thickness of 0.7 mm.
In the world of Formula One, which is under no
emissions regulations, the exhaust system is designed for
high power. Broadly speaking, there are two points to
focus on: reducing exhaust loss and using the dynamic
effect of exhaust pulsation.
To reduce exhaust loss, pressure loss has been
consistently reduced by enlarging the pipe bending-angle
and bending-radius within the limits allowed for the
layout under the vehicle cover (cowl).
On the other hand, the primaries are gathered together
to utilize exhaust interference. As a result, within the
collector, exhaust pressure causes reflected waves to
form from the open edge. These reflected waves have
an effect on internal pressure in the cylinders during
valve overlap, and the amount of residual gas can be

decreased. This leads to an increase in specific intake air
volume as a result, making it possible to alter power
characteristics in relation to engine speed. This uses the
dynamic effect of exhaust pulsation. With the reflected
waves in the collector alone, however, the range of
engine speeds in which pulsation can be used is limited,
so stepped pipes (steps) were implemented in the
primaries (Fig. 17).
At 17500 rpm, a gain of 4 - 8 kW was realized as
compared to an exhaust without steps.
While Honda was competing in Formula One during
the third era, a forward exhaust system was used in 2007
only (Fig. 18), while a backward exhaust system was
used in all other years (Fig. 19).
In order to enhance car aerodynamic performance to
compensate for the engine power that decreased with the
change in regulations that stipulated V8 engine use, it
was necessary to increase the degree of freedom of
aerodynamics design. A forward exhaust system has less
negative impact on engine power than a backward
exhaust system, and allows a greater degree of freedom
in the aerodynamics design of the rear of the engine.
However, all the high-temperature parts were
contained within the engine cowl, and the exhaust outlet
was placed in a position that did not promote ventilation
from the outlet, so there was frequently heat damage,
including to vehicle parts. For that reason, Honda
returned to the backward exhaust system in 2008.
Tail
Collector
For dyno
Fig. 16
Primary
F1 exhaust system (right-hand side for car)

Fig. 18
Forward exhaust system
Section AA
A
Fig. 17
Stepped primaries
Fig. 19
78
Backward exhaust system
3.2.2. Compact and lightweight technology

The exhaust system has a major impact on a racing
cars vehicle dynamics in terms of component size and
weight. Therefore, the technology of packaging the
components themselves was always being advanced.
Because an exhaust system is made by bending
cylindrical pipes, the primaries and collectors require
space within the cowl and have an impact on
aerodynamics.
With the exhaust system with non-circular sections
(compact exhaust) developed in 2008, the aim was to
achieve a space-saving, lightweight exhaust system that
retained engine power, durability and reliability, rather
than deciding on an exhaust form in a way that depends
on the aerodynamics concept based on the cowl.
For the primaries, a non-circular section structure,
three-dimensional bend layout and ellipsoidal stepped
pipes were used and design was conducted that took
account of heat damage to the engine itself. The
production method proposal ensured molding accuracy
by pressworking that took advantage of the know-how
of an exhaust system manufacturer.
To make the collector lighter, it was considered to
use a shared-pipe outer wall where the collectors
gathered together, and to use a long collector in which
a part of the primary would be taken into the collector
and sharing of the wall further increased, and
components were thus produced.
When sharing the pipe outer wall, if using
pressworking, which is the conventional way, more dies
are needed and material yield declines, which raises
production costs, so a new precision casting technology
was used. The result was a form with wall thickness of
0.7 mm (Fig. 20).
To deal with the increased thermal load that comes
from sharing the pipe outer wall, Ren 41 was used,
which is suitable for casting and has greater hightemperature fatigue strength than conventional materials.
Results showed that in dyno durability tests, component
life was extended about 60% over that of ordinary
Inconel 625 collectors.
An image of the entire compact exhaust system is
shown in Fig. 21.
These specifications allowed the cowl line to draw
50 mm closer to the engine than before. However, in
order to make effective use of limited space while
minimizing the impact of using a non-circular section,
a layout was chosen that extended primaries to a length
greater than standard specifications and wrapped them
around to the engine front.
Figure 22 shows results of checking engine power
under these specifications. This shows the compact
exhaust power gain under the condition that the
primaries have been extended. This is for a single
bank, but there was an increase in power of 4 kW on
average from 7500 rpm to 11500 rpm. The factor that
increased power is believed to be that, since the
volume of the collector gather was approximately
34% less than the conventional specifications, there
were changes in exhaust pulsation in a certain engine
speed range, which caused engine power

characteristics to change.
3.2.3. Torque boosting technique
During races, engine speeds of 16000 rpm and higher
are used about 90% of the time. However, torque
characteristics at lower engine speeds are critical at the
start of the race and when the car is accelerating from
the apex of the curve, and such characteristics affect lap
times and race results.
For that reason, the exhaust system is developed not
only for power at high engine speeds but also for torque
characteristics at low engine speeds. Specifically, the
form of the internal wall of the collectors has undergone
optimization since 2005.
Figure 23 shows a cross-sectional view of collectors
with internal walls.
The form of the internal wall in a standard collector
is decided by the collection angle of the pipes as well
as the pipe diameter, and the wall edge forms a curve.
If left unmodified, the cross-sectional area of the pipe
would remain unchanged up to the gather as it went to
the open edge.
If the wall height were intentionally raised and the
cross-sectional area restricted, there would be no
Fig. 20
Difference from standard exhaust [kW]
Fig. 21
Compact exhaust system
10
-5
7000 8000 9000 10000 11000 12000 13000 14000 15000 16000 17000 18000 19000
Engine speed [rpm]
Fig. 22
79
Casting collector
Performance of compact exhaust system
decrease in power at high engine speeds as a result of

increased pressure loss, and the phase of reflected waves
would be delayed. This is thought to be caused by the
fact that the distance to the open edge was increased and
exhaust gases were kept below supersonic speed.
As a result, exhaust pulsation characteristics change
at low engine speeds and combustion gas flow inside the
cylinders is impacted, causing changes in torque
characteristics. Starting in 2005, the optimization of the
form of the internal wall of the collectors was constantly
pursued as one means of enhancing torque characteristics
at low engine speeds.
Since 2008, the regulations have prohibited the use
of traction control systems, making it even more
necessary to enhance torque characteristics from the
point of view of drivability.
However, there are areas where optimization of wall
form alone is not sufficient compensation, and there is
also the effect of inconsistent combustion, so the issue
of drivability was not completely resolved.
An effective way to enhance drivability is to flatten
engine torque at partial throttle (i.e., engine torque when
the throttle is neither fully closed nor fully open) at low
engine speeds.
To increase development efficiency, a test exhaust
system with variable pipe diameters and lengths was
used (Fig. 24) as well as simulation to study form,
thereby verifying the drivability enhancement effect. This
part discusses two results that demonstrate the
effectiveness of development.
The first was an exhaust system with connecting
pipes (i.e., balance pipes) (Fig. 25). The goal was to
cause changes in exhaust pulsation characteristics with
these connecting pipes. Simulation was used to select
connecting pipe width, length and connecting points, and
specifications were selected that enabled residual gas to
be reduced.
The second was the 4-2-1 exhaust system. The basic

stance was to alter exhaust pulsation characteristics just
as for the balance pipes mentioned above. Specifications
were selected while confirming the power with the test
exhaust system. Ultimately, the primary length was made
50 mm longer than in a 4-1 exhaust system, with 360o
assemblies of #1-#4, #2-#3, #5-#8 and #6-#7 (i.e.,
connecting cylinders with the ignition phase offset 360o).
Power check results are shown in Fig. 26.
The balance pipes yielded a power gain of 12 kW at
engine speeds of 8500 rpm and 10500 rpm, but a drop
of 2.5 kW at 17000 rpm.
The 4-2-1 exhaust system increased power by an
average of 8 kW at engine speeds from 8500 rpm to
10500 rpm and yielded about the same results as the
base exhaust (4-1 exhaust system) at 17000 rpm and up.
Figure 27 compares results of engine torque at
partial throttle.
Both the balance pipes and 4-2-1 exhaust system
showed torque flattening and superior performance as
compared to the base exhaust.
Difference from standard exhaust [kW]
Fig. 25
With internal wall

Without internal wall
20
15
10
5
0
-5
-10
-15
-20
6000
Balance pipes
4-2-1
8000
10000
Fig. 26
200
14000
16000
18000
Performance of balance pipes and 4-2-1

Standard exhaust
Balance pipes
4-2-1
Throttle opening rate %

35%
150
Cross-sectional view of collector

Torque [Nm]
12000
Engine speed [rpm]
Section AA
Fig. 23
Balance pipes
25%
100
50
14%
10%
4%
-50
-100
7000
Fig. 24
Fig. 27
Test exhaust system
80
9000
11000
13000
Engine speed [rpm]
15000
Torque characteristics at partial throttle
However, drivers were not able to experience the

superiority of the balanced pipes in circuit tests. Also,
the connecting pipes caused the weight to increase, and
there were still issues with durability and reliability
because of scattered cracks in the weld to the base pipe.
It was estimated that the weight increase would be
approximately 220 g over that of the base exhaust.
4. Conclusion
Through engine development during Hondas thirdera Formula One activities, we have learned the
following about induction and exhaust systems.
(1) The importance of induction and exhaust system design
that is mindful of the vehicle package became apparent
once again, and development techniques were created
that extract maximum performance from Formula One
cars.
(2) Even with racing engines, it is necessary to be aware of
torque characteristics at low and medium speeds, and is
important to design exhaust systems as a technique for
their enhancement.
(3) Induction system development techniques were created
that use CFD and can predict dynamic characteristics
from the design stage. If it is possible to predict the
dynamic characteristics of exhaust systems in the future,
this will enhance development efficiency even more.
(4) Test part-production techniques were implemented to suit
a short development cycle, and the time required to
optimize power characteristics and determine
specifications was shortened.
References
(1) Nakamura, S., Motohashi, Y., Hayakawa, S.:
Development of Wind Simulator Equipment for Analysis
of Intake Phenomena in Formula One Engines, Honda
Era Activities), P. 89-94
(2) Hanada, N., Hiraide, A., Takahashi, M.: CFD Technology
for Formula One Engine, Honda R&D Technical Review
2009, F1 Special (The Third Era Activities), P. 82-88
Author
Ken NISHIMORI
Ryuichi FURUKAWA
81
Yasuhiro MOTOHASHI
Masayoshi TAKAHASHI
Naoki HANADA*
Atsushi HIRAIDE*
Masayoshi TAKAHASHI*
ABSTRACT
Simulation technology has advanced markedly in recent years, and various types of CFD models have come into
use for Formula One engine development.
However, to use such simulation modeling it is necessary to establish simulation technology for the unique conditions
for Formula One engines, such as that for high engine speed.
The pressure of each engine part was measured and in-cylinder gas motion and fuel spraying behavior were also
measured using a single-cylinder optical engine, enabling CFD technology that can be applied to Formula One engines
to be created.
1. Introduction
are high, as high as 20000 rpm. At high engine speeds,

several factors affect output: increased piston speed and
gas flow rate, fuel spray behavior and cylinder
interference in the intake and exhaust systems resulting
from pressure fluctuations within the intake and exhaust
pipes, the motion of gas coming into the cylinders, and
so on. We began work on developing a simulation model
that could express these impact parameters.
Using data from pressure measurements for each part
and in-cylinder measurements taken with an optical
engine, a simulation model usable for the development
of Formula One engines was validated.
This paper explains CFD technology for the flow of
air during intake and exhaust, in-cylinder gas motion, incylinder fuel behavior, and combustion.
Simulation technology has advanced markedly in

recent years. Even in many fields for Formula One,
simulation technology has been established and has come
into use for development. CFD requires extensive
computer resources and has become more practical as
hardware and computing technology have advanced.
CFD modeling uses a variety of sub-models such as
turbulence models, and it is crucial to validating whether
the phenomenon being calculated can be expressed. For
example, in the case of automobile reciprocating engines,
which use a wide range of operating speeds, the pressure
fluctuation within the intake and exhaust pipes will differ
at low engine speeds and high engine speeds. A
calculation technique validated for just low engine
speeds or just high engine speeds would not necessarily
be compatible with the other. Dependable validation of
simulations helps make it possible to create CFD
technology that is effective for development.
CFD technology for commercial engines(1)-(5) has been
developed by members of a special team. At first, since
the Formula One engine is also a reciprocating engine,
it was thought that the CFD technology should also be
compatible with the Formula One engine as is, and the
attempt was made to apply the technology to Formula
One engine development. When actual simulation was
performed, however, it became clear that the technology
could not express Formula One engine phenomena. Since
commercial engines are normally used at low engine
speeds, low engine speeds of about 2000 rpm are an
important analysis condition of simulations too. Since the
highest priority of a Formula One engine is output, it is
necessary to conduct evaluation when the engine speeds
2. Intake and Exhaust Systems

The intake and exhaust systems of a reciprocating
engine impact volumetric efficiency and as such are
important components determining engine output.
Being able to analyze pressure and gas flow of intake
and exhaust systems and predict performance in advance
provide effective ways of determining specifications in
a short period of time when developing Formula One
engines, where enhancements are expected on a daily
basis. To achieve such predictions of performance, it is
necessary to perform simulation studies in advance, and
the results should be cross-checked with the
measurement data.
Intake and exhaust system simulations were originally
performed using one-dimensional gas dynamics and
engine performance simulation software WAVE made by
Ricardo Software. Subsequently, for the intake system,

82
an in-house software capable of three-dimensional

compressible fluid analysis was used, with the aim of
enhancing calculation accuracy. For the exhaust system,
usable three-dimensional analysis software could not be
found and so we continued to use WAVE.
three-dimensional compressible fluid calculation

accuracy.
Based on the measured intake pipe pressure,
boundary conditions, mesh density, flow coefficient,
filter model and so on were investigated and a series of
enhancements were made until a simulation that could
be used for the analysis of intake interference in the
airbox of a Formula One engine was created.
Figure 2 shows measurement and calculation results
of pressure within intake pipes in each of the cylinders
in one bank of a V10 engine. Engine speed was 17000
rpm, and the crank angle of 360 deg is top dead center
of intake stroke. Intake interference within the airbox
causes the different intake pressure in all the cylinders.
While the amplitude and phase of pressure varies with
2.1. Intake System

The current Formula One engine intake system is
equipped with an airbox, and the pressure fluctuation
(pulsation) that occurs within the intake pipe of each
cylinder propagates and reflects inside the airbox and is
transmitted to other cylinders. As a result, airbox shape
is one factor affecting engine output.
On the other hand, the bodys aerodynamic
performance greatly affects lap time. Therefore, the
shape of the body cowl is determined by aerodynamic
characteristics. As for airbox shape, the degree of
freedom is restricted by the shape of the cowl.
Using WAVE, at first, performance was evaluated
with the air intake form only, but in order to achieve
both body aerodynamic performance and engine output,
a large amount of time was spent during development
on selecting the airbox shape. We tackled this issue so
that simulations could be utilized for performance
evaluation and phenomenon analysis including also the
influence of airbox shape.
The simulation software used was Hondas threedimensional compressible fluid analysis software(1) in
which a one-dimensional model is coupled.
The simulation model is shown in Fig. 1. The airbox
portion was expressed in three-dimensional form, and the
portion from the downstream part of the air intake
throttle to the cylinder and exhaust pipe was expressed
by one-dimensional and scalar models.
The greatest feature of Hondas software is that it
solves compressible fluids by not only one-dimensional
but also three-dimensional calculation.
Although it is possible to do similar calculations by
combining Ricardo Softwares three-dimensional fluid
analysis software (VECTIS) with their one-dimensional
gas dynamics simulation software (WAVE), these
software programs did not satisfy our target standards of
Test
CFD
Pressure [Pa]
200
Cylinder 6
150
100
50
0
0
180
360
Crank angle [deg]
540
720
180
360
Crank angle [deg]
540
720
180
360
Crank angle [deg]
540
720
180
360
Crank angle [deg]
540
720
360
Crank angle [deg]
540
720
Pressure [Pa]
200
Cylinder 7
150
100
50
0
0
Pressure [Pa]
200
Cylinder 8
150
100
50
0
0
Pressure [Pa]
200
Cylinder 9
150
100
50
0
0
3-dimensional
model
Pressure [Pa]
200
1-dimensional
model
Cylinder 10
150
100
50
0
0
Fig. 1
Simulation model
Fig. 2
83
180
Intake pressure of V10 engine at 17000 rpm

Table 1
Parameters for WAVE simulation
Parameter
Old value
New value
Flow coefficient at pipe end (inflow)
Auto
0.8
Flow coefficient at pipe end (outflow)
Auto
0.5
Ambient temperature of exhaust side
Atmospheric temperature
400 K
1.0
2.0
Multiplier of heat transfer of cylinder wall

when intake valves are closed
WAVE, a model used for the intake side was designed

such that intake pipes are open to the air, with each
cylinder being independent. Because the left and right
exhaust systems are independent, the WAVE model used
one bank with five cylinders.
Figure 5 shows V10 engine validation results.
Exhaust pressure was measured at a downstream
position approximately 40 mm from the exhaust valve
seat of cylinder 10. Because the calculation model has
five cylinders, the data used was of cylinder 5. The
initial calculation results (b) differed from the measured
results (a).
A parameter study to revise the variables used for
simulation shown in Table 1 yielded the results shown
each cylinder, there is a particular difference between

cylinder 10 and the other cylinders. It can be seen that
intake interference causes the volumetric efficiency of
each cylinder to change and has a direct impact on
engine output.
The calculation results reproduce these characteristics
with good accuracy, demonstrating that Hondas software
is capable of analyzing intake interference within the
Formula One engines airbox.
The slim and curving airbox shape required for good
body aerodynamic performance is due to reduced volume
and is characterized by a weak attenuation of pressure
fluctuation and lowered volumetric efficiency. However,
because practical simulation has been achieved, it is now
possible to satisfy both a slim airbox shape which can
contribute to body aerodynamics and engine output
enhancement by optimizing air flow in the airbox.
(a) Test
(b) WAVE / not validated
(c) WAVE / validated
300
Pressure [kPa]
2.2. Exhaust System

WAVE was used to analyze the exhaust system and
examine its shape.
This explanation uses a V10 engine as an example.
A 5-into-1 collector exhaust system, as shown in Fig.
3, was used. The WAVE model is shown in Fig. 4.
Because the effect of intake interference in the airbox
cannot be expressed in one-dimensional simulation
12000 rpm
200
100
0
Pressure [kPa]
300
14000 rpm
200
100
0
Pressure [kPa]
300
Fig. 3
Exhaust pipe of V10 engine
16000 rpm
200
100
0
Pressure [kPa]
300
17800 rpm
200
100
0
0
180
360
540
Crank angle [deg]
Fig. 5
Fig. 4
WAVE model of 5 cylinders for V10 engine
84
Exhaust pressure of V10 engine

(#10 cylinder)
720

Table 2
in (c) in Fig. 5, with enhanced calculation accuracy. The

important revised parameter is the flow coefficient for
the end of the exhaust pipe. In the WAVE calculation,
exhaust pressure changes greatly depending on the value
of the flow coefficient, and ultimately volumetric
efficiency and output values also change. Estimating the
flow coefficient correctly is clearly the most important
aspect of one-dimensional simulation.
In addition, this result shows that there are limits to
analyzing exhaust systems with one-dimensional
simulation. One-dimensional analysis has many subjects
to be addressed such as followings. Naturally, it is not
possible to estimate pressure loss in each part with onedimensional analysis. General usefulness of flow
coefficient is low for changes in the form of the exhaust
pipe or changes in flow velocity. To some extent it is
possible to analyze and examine flow within exhaust
pipes with one-dimensional simulation, so this can be
used in the development of exhaust systems. However,
the calculation accuracy is not necessarily sufficient; for
example, revision of the flow coefficient is inevitable
when engine specifications change. Development of
three-dimensional simulation, as used with the intake
system, is needed in the future as a way to resolve this
issue.
Calculation conditions of in-cylinder gas

motion
Calculation region
Whole engine
Number of calculation cycle
4 cycles
Mesh density
(for combustion chamber)
1st to 3rd cycle: 2.0 mm

4th cycle:
1.5 mm
(Half the above size is used

near the wall.)
simulation model. To correctly express intake and

exhaust pulsation, the calculation covered the entire
single cylinder engine from air inlet chamber to exhaust
pipe.
The calculation conditions are shown in Table 2.
These conditions were selected to enable calculation
results to reproduce PIV measurement results and at the
same time cause intake pulsation, volumetric efficiency
and in-cylinder turbulence energy to converge. The
reason that mesh density is changed with the number of
cycles is to ensure both calculation accuracy and reduced
calculating time. Except for turbulence intensity, the
calculation results showed no differences between 2 mm
and 1.5 mm of mesh densities, but for turbulence energy,
changes in cycles could not be expressed at mesh density
of 2 mm. For that reason, calculating time was saved by
using a coarser mesh to calculate the first three cycles,
and then a finer mesh was used on the fourth cycle to
ensure calculating accuracy.
Figure 8 shows visualization and simulation results
in a motored condition at 10000 rpm. The crank angle
shown in the figure is the angle after top dead center.
Measurements were performed through a cross-section
3. In-cylinder Behavior
Enhancing the combustion status is indispensable for
achieving engine performance enhancements. Besides
that, the current Formula One engine has longer
combustion duration than commercial engines because of
its high engine speed and big bore. Therefore, shortening
the combustion duration is necessary.
To shorten combustion, it was considered necessary
to adjust the in-cylinder gas motion and fuel distribution.
To do this, it is indispensable to be able to predict incylinder gas motion and simulate fuel spray and mixture
distribution formation.
Seeding
equipment
Airbox
Head
Laser
3.1. In-cylinder Gas Motion

Validation on in-cylinder gas motion has been
performed on commercial engines and an attempt has
been made to apply the results to Formula One engines,
but no correlation was found to actual engines. Thus we
started over again and redid the process from validation
of the Formula One engines in-cylinder motion.
Figure 6 gives an illustration of the system for
measuring in-cylinder motion. An optical engine(6) was
used. Two high-repetition-rate YAG lasers were used
with a YAG 532 nm second harmonic wavelength. The
camera used was Vision Researchs Phantom V7, and
resolution was 656 328 pixels. Hollow resin tracers
were used, with particle size of 40 m and density of
36 kg/m3. Based on tracer size and density, it is thought
that air flow traceability will have a 95% attenuation rate
for fluctuations of approximately 300 Hz. The
measurement data took the mean for 100 cycles.
The simulation software was three-dimensional fluid
analysis software (VECTIS). Figure 7 shows the
Camera
Engine
PC
Controller
Fig. 6
Fig. 7
85
Schematic of PIV measurement system with

optical single cylinder engine
Simulation model of in-cylinder gas motion
under the intake valve, as shown in the upper right of

the figure. Two intake pipe lengths of 200 mm and 150
mm were used to compare differences in gas motions.
The left column is from the 200 mm intake pipe, and
the right column is from the 150 mm intake pipe.
In-cylinder motion in a Formula One engine consists
of two vortexes, as demonstrated by the example of
crank angle 216 deg ATDC: a large vortex centered on
the bore, and a vortex that forms on the bottom of the
intake port. The images show that the twin vortex
configuration does not change with the difference in
intake pipe length, but the gas inlet velocity, vortex
center and vortex speed are affected by intake pipe
length. This means, in other words, that intake pulsation
has an impact on not only volumetric efficiency but also
combustion.
Calculation results showed a flow pattern that closely
matches visualization results. This demonstrates that if

a model shape and calculation conditions are chosen
such that intake and exhaust pulsation can be expressed,
it is possible to accurately simulate in-cylinder gas
motion at the high speed of 10000 rpm.
These initiatives have made it possible to predict gas
motion in a Formula One super-high-speed engine. In
addition, this research has renewed the researchers
awareness of the importance of measurement that helps
one to grasp phenomena when creating simulation
technology. Without PIV measurement results, what to
set as the number of cycles and mesh density cannot be
determined.
The fact that turbulence intensity is not validated is
an issue, and analysis of this, including its impact on
combustion, should be performed.
3.2. In-cylinder Fuel Behavior
The distribution of fuel within the cylinder is one
factor affecting the quality of combustion, and it is
important to predict this in advance with simulation.
However, there are issues with measuring fuel
distribution within cylinders, and measurements are not
easy. For that reason, analysis by simulation is valuable
in terms of advancing phenomenon analysis.
Injectors of Formula One engine are typically
installed close to the end face of the trumpet. The reason
is because volumetric efficiency is increased as a result
of charge cooling by evaporation latent heat of the fuel.
Because of this installation, fuel injected toward the
interior of the intake port is affected by gas motion
which fluctuates because of intake pulsation, and then
it enters the cylinder. Therefore, to calculate in-cylinder
fuel distribution, the necessary calculations must include
fuel behavior inside the intake port.
Simulation validation was performed with the
behavior of the fuel droplets flowing into the cylinder
which was photographed with an optical engine. Using
two kinds of injectors, it was evaluated whether the
simulation can express the difference produced from the
difference in the fuel spray characteristic. One was a
pintle form with fuel pressure of 1.2 MPa, and the
second was a six-hole plate form with fuel pressure of
10 MPa. These were characterized by respective Sauters
mean diameters of approximately 40 m and 16 m,
respectively.
Using an optical engine(6), spray droplets were filmed
directly with a strobe as a light source. Operating
conditions were: motored at 10000 rpm, and wide open
throttle.
For simulation, VECTIS was used, the same as for
in-cylinder motion calculation. An entire single cylinder
engine was modeled and a fuel spray calculation was
added. To calculate fuel spray, the following method was
used. The state of spraying with the injector alone was
matched with measurement results in advance. Then, the
fuel spray calculation conditions were input into the
single cylinder engine calculation.
The fuel spray form of the injector alone is shown
in Fig. 9.
Laser
sheet
View
Simulation
PIV
126 deg ATDC
Simulation
PIV
180 deg ATDC
PIV
216 deg ATDC
Simulation
Vortex center
2 vortexes
(a) Port length 200 mm
Fig. 8
(b) Port length 150 mm
Comparison of flow vectors between PIV

measurement and simulation results at
10000 rpm
86
This shows that with both the 1.2 MPa and 10 MPa
injectors, the calculations were able to sufficiently
express the actual spray form.
It was learned that for in-cylinder mixture forms to
converge in calculations of fuel spray in the engine,
about five cycles are necessary. Figure 10 shows the
cycle history of in-cylinder average A/F and A/F
variance when calculating sprays in an engine using a
10 MPa injector. Fuel injected from the trumpet end face
begins to enter the cylinder on the second cycle after
injection, but the amount is small. This figure can be
read as showing in-cylinder fuel behavior converging
from the fourth to fifth cycle. To give some margin, it
was decided to conduct the analysis with five cycles.
Figure 11 shows the results of photographing incylinder spraying and the calculation results.
The measurement results are colored to represent the
degree of photographed brightness. In the calculation
results, spray droplets are represented by dots, and the
broken line shows the approximate area in which the
droplets occur.
With the pintle, the fuel spray has large particles and
the spray has great inertia. That is why, as the
photograph taken at crank angle 120 deg shows, the
spray entering the cylinder from the intake valve passes
below the exhaust valve and collides with the cylinder
wall. In contrast, the high pressure injector spray has
little inertia, and after it enters the cylinder, it rides the
flow of air down the cylinder toward the piston. In
addition, the spray entering from the intake valve enters
not only from the exhaust side but also from the port
bottom side that is opposite it.
To make the fuel pass through the entire combustion
chamber, it would be effective to reduce particle size and
inertia.
The simulation largely expressed the measurement

results, and by matching the spray with the injector alone
beforehand, it was possible to predict in-cylinder fuel
behavior. In addition, the fact that gas motion inside both
the intake port and the combustion chamber was
correctly solved is one factor for accurate calculation.
3.3. Combustion
To directly evaluate the influence of gas motion and
fuel behavior on combustion, and to predict amount of
residual gas, efforts were focused on combustion
simulation as the next step.
The software used was VECTIS, but the combustion
simulation module(4) made by Honda was incorporated
into VECTIS.
120 deg Test

ATDC
CFD
150 deg Test

ATDC
CFD
180 deg Test

ATDC
CFD
CFD
Test
CFD
Test
210 deg Test

ATDC
(a) Pintle Pf = 1.2 MPa
Fig. 9
Fuel spray form

240 deg Test
ATDC
14
A/F and variance
CFD
(b) Multi hole Pf = 10 MPa
Averaged A/F
12
10
CFD
8
6
4
Variance of A/F
(a) Pintle Pf = 1.2 MPa
(b) Multi hole Pf = 10 MPa
0
1
inj = 100 deg
inj = 144 deg
Cycle
Fig. 11
Fig. 10
Simulation convergence of averaged A/F and

variance of A/F in cylinder
87
Comparison of fuel droplets in cylinder

between measurement with optical engine
and simulation with VECTIS at 10000 rpm
References
Pressure [MPa]
The subject of the simulation evaluation was not the

Formula One engine itself, but rather a single cylinder
engine with the same specifications as the Formula One
engine. The calculation model was the same as the incylinder motion and fuel spray model shown in Fig. 7.
As a calculation procedure, at first, in-cylinder
motion and fuel behavior in a motored condition were
calculated up to the point just before ignition on the fifth
cycle for convergences of gas motion and mixture
concentration. After that, combustion calculation was
performed only by a model of the combustion chamber
except intake and exhaust elements.
In-cylinder pressure measurements and calculation
results for a single cylinder engine at 15000 rpm and
17750 rpm are shown in Fig. 12. In the combustion
simulation used, it is necessary to tune a constant that
adjusts combustion speed, but the simulation was able
to reproduce in-cylinder pressures at different engine
speeds with the same constant. However, this is a
different value from the constant used for commercial
engines and it was confirmed that it is necessary to tune
a unique value for Formula One engines. The outlook
is that Hondas own combustion simulation can be
applied to Formula One engines as well, and the research
is at the stage of confirming whether it has general
usefulness for an even greater number of operating
conditions.
10
8
6
4
2
0
300
(1) Takabayashi, T., Hotozuka, Y., Tsushima, H.: The Threedimensional Pulsation Flow Simulation and Its
Application for Engine Intake and Exhaust Systems, 2000
FISITA World Automotive Congress, F2000A091
(2000)
(2) Ishikawa, N., Hiraide, A., Takabayashi, T.: Air/Fuel
Distribution Simulation in a Port-Injected Gasoline LeanBurn Engine, SAE Paper, 2001-01-1230 (2001)
(3) Zhu, G., Reitz, R. D., Xin, J., Takabayashi, T.:
Characteristics of Vaporizing Continuous MultiComponent Fuel Sprays in a Port Fuel Injection Gasoline
Engine, SAE Paper, 2001-01-1231 (2001)
(4) Yang, X., Ohashi, T., Takabayashi, T., Kubota, S.,
Katsuyama, H., Urata, Y.: Ignition and Combustion
Modeling with G-Equation in Spark Ignition Engines,
The 13th International Pacific Conference on Automotive
Engineering (2005)
(5) Ohashi, T., Yang, X., Takabayashi, T., Urata, Y., Kubota,
S., Katsuyama, H.: Ignition and Combustion Simulation
in HCCI Engines, SAE Paper, 2006-01-1522 (2006)
(6) Yanagisawa, N., Hanada, N., Hamakawa, T., Ogiyama,
k., Goto, T.: Measurement Technologies for Formula One
Engines, Honda R&D Technical Review 2009, F1
Special (The Third Era Activities), p. 101-108
Test
CALC
330
360
390
420
450
Pressure [MPa]
Crank angle [deg]

at 15000 rpm
10
8
6
4
2
0
300
Test
CALC
330
360
390
420
450
Crank angle [deg]

at 17750 rpm
Fig. 12
Comparison of cylinder pressure between

measurement and simulation
4. Conclusion
Creating simulation models has been attempted to
allow development of Formula One engines with more
performance more quickly.
Simulations which use various models cannot
demonstrate their ability unless reliable validation is
conducted for each subject of calculation.
In the future, we hope to use the model validated and
created with Formula One to develop CFD technology
for engines that will be friendlier to the global
environment.
Author
Naoki HANADA
88
Atsushi HIRAIDE
Masayoshi TAKAHASHI
Combustion Diagnosis of Formula One Engine

Using Micro-Cassegrain Sensor
Naoki HANADA*
Yuji IKEDA**
Atsushi HIRAIDE*
Atsushi NISHIYAMA**
Kohei YAMADA*
ABSTRACT
Combustion diagnosis of a Formula One engine during wide open throttle (WOT) acceleration and deceleration
operations was performed using a micro-Cassegrain system.
The air/fuel mixture (A/F) in each cylinder was measured, which helped in the development of controls to minimize
the torque loss due to unstable combustion.
In transient conditions where acceleration and deceleration are performed repeatedly, the air/fuel mixture around
the spark plugs becomes too rich or too lean, and results in unstable combustion. In addition, the air/fuel mixture
formation is different inside each cylinder.
The fuel distribution to each cylinder needed to be controlled to a high degree of accuracy, and to do this, a highly
responsive air/fuel ratio sensor (LAF sensor) is required.
1. Introduction
the cylinder so that the appropriate A/F is provided to

each cycle.
As shown in Fig. 1, as the indicated mean effective
pressure (IMEP) rises, the misfiring occurs in some
cycles. This is because the appropriate A/F is not formed
around the spark plugs, resulting in unstable combustion.
It is important to analyze why the misfiring occurs and
to take preventive measures.
Existing combustion diagnosis methods include
measuring the induction pressure, in-cylinder pressure,
exhaust pressure, and exhaust gas characteristics; and
Flame
speed
A/F
IMEP
To increase the engine output, it is important to find

a way to efficiently combust the air/fuel mixture.
Combustion involves a series of processes, including
forming the air/fuel mixture inside the cylinder,
performing an effective ignition, promoting combustion
with turbulence, and scavenging. Making the initial
flame more stable by performing an effective ignition
can minimize unstable combustion and lead to increased
thermal efficiency.
Although analysis of the air/fuel mixture formation
and flame propagation inside all the cylinders was also
required, priority was given to analyzing the engine issue
as described by the driver. The A/F around the spark
plugs, which directly affects the ignition performance,
was measured using a micro-Cassegrain sensor.
In Formula One engines, nearly all throttle operations
are WOT or in fully closed positions. For this reason,
fuel control is required that matches sudden changes in
the amount of induction air, and if the fuel correction is
inappropriate, unstable combustion will occur and the
drivability will be reduced.
The flame will not be able to propagate quickly if
the air/fuel mixture is too lean or too rich.
To enhance the drivability, fuel need be supplied to
Cycles
Fig. 1

** Imagineering, Inc.
89
Combustion characteristics under acceleration
measuring the A/F ratio between the cylinders with the

LAF sensor. Regarding the measurement of the
combustion variation between cylinders, published
reports include those that describe pressure (1), (2), ion
probes (3) , the IR absorption method (4), (5) , and the
chemiluminescence measurement method(6), (7).
As a new combustion diagnosis method, a microCassegrain sensor was used to directly measure the A/F
around the spark plugs, and measurement was performed
of the corresponding time series variations and
combustion variations between cylinders. In this method,
time series analysis can be performed up to 20000 rpm.
This document is a report about the results of the
unstable combustion analysis performed under the
transient operation of a Formula One engine.
The combustion pressure for 500 cycles was measured

with an engine speed of 16500 rpm and WOT.
The air excess ratio (set Lambda) was set at 0.95.
The deviation of combustion pressure between
cylinders was 16.7%, and so it was shown that the builtin micro-Cassegrain sensors did not greatly affect the
combustion variation.
It was also confirmed that negative effects from the
rpm, loads and Lambda value and the plasma from early
discharge were not detected by the micro-Cassegrain
sensors.
Sapphire glass was installed on the end of the microCassegrain sensors as a countermeasure for heat and
pressure resistance.
Figure 4 shows an overall view of cylinders 5 to 8
to which the micro-Cassegrain sensors were attached.
Figure 5 shows a view of the optical fiber attached to
a spark plug in a plug hole.
Figure 6 shows the device configuration. The
Cassegrain sensor output value, in-cylinder pressure,
induction pipe pressure, exhaust pipe pressure, LAF
sensor output value, and crank angle signal were
synchronized and imported.
SPB2000 (made by Imagineering, Inc.) and DS2000
(made by Ono Sokki Co., Ltd.) were used as the data
processing devices. SPB2000 is equipped with a bandpass filter and interference filter for each radical, and the
light signal from each cylinder is converted into an
electrical signal by the photomultiplier.
2. Measurement Device
A V8 race engine in 2008 that had a displacement
of 2400 cm3 was used. Figure 2 shows the exterior of
the engine.
A micro-Cassegrain system (made by Imagineering,
Inc.) was used to measure the A/F around the spark
plugs.
When the flame passed a certain point (about 0.1 x
0.1 mm), the chemiluminescence intensities of four
r a d i c a l s ( O H , C H , C N , C 2) f r o m t h e f l a m e
chemiluminescence was measured using a band-pass
filter and a photomultiplier.
The intensity ratios of these radicals have an
unambiguous correlation with the equivalence ratio in the
flame(8)(16). This means that an approximate measurement
can be made of the flame propagation speed and the A/
F in the flame.
The measurement system was comprised of an optical
fiber sensor that measures the local chemiluminescence,
hardware that processes the chemiluminescence intensity
ratio in high-time resolution, control software, and a
monitor.
The sensor type used was an M10 spark plug with a
built-in Cassegrain optical element. The distance of the
focal point from the ignition position was 3 mm.
The effect on combustion by inserting the sensor into
the spark plug was checked. Figure 3 shows the results
of the evaluation of the combustion variation using the
pressure sensor. Normal spark plugs were installed for
cylinders 1 to 4, and spark plugs with the built-in microCassegrain sensors were installed for cylinders 5 to 8.
Target lambda: 0.95
Pmax, a.u. [%]
120
100
107.74
100.66 98.59 104.35 97.27
90.99 93.15
80
60
Normal plug
Sensor plug
40
20
0
#1
#2
#3
#4
#5
#6
#7
#8
Cylinder
Fig. 3 Comparison of maximum cylinder pressure

(Pmax) using normal plug and sensor plug
Fig. 4
Fig. 2
107.25
Honda Formula one engine
90
Photograph of overview of cylinders to which

micro-Cassegrain sensors were attached
Fig. 5
3.1. Combustion Characteristics at 18000 RPM Engine

Speed, WOT Load, and Steady Condition
Figure 7 shows a comparison of the results with
different vehicle velocities and air box specifications.
The horizontal axis is Lambda, and the vertical axis is
the average values and the variation range. The
combustion difference and the time series combustion
variation were compared between cylinders 5 and 6.
The faster the vehicle velocity, the higher the IMEP
becomes because of the lam pressure. The highest pressure
was reached when the set Lambda was around 0.9.
However, P max and IMEP values varied between
cylinders. The combustion pressure variation range also
varied according to the Lambda value.
At 120 km/h and 300 km/h in the standard air box,
the average A/F and the A/F around the spark plugs were
approximately the same, which is a desirable result.
The air box inlet shape was changed and
measurement was performed at a vehicle velocity of 300
km/h. The A/F around the spark plug in cylinder 5
became richer by about 0.5 from the set value. The inlet
shape affects the air flow pattern, and so it also likely
affects the A/F distribution inside the combustion
chamber.
No significant change was seen in the FPS within a
Lambda range of 0.8 to 0.95. The same trend has been
reported in another combustion analysis for Formula One
engines(7).
However, it should be noted that the FPS varies
greatly regardless of the Lambda.
Photograph of plug hole in which plug with

sensor was attached
V8 engine Cylinder 1 Cylinder 2 Cylinder 3 Cylinder 4

Ex. Ex.
Ex. Ex.
Ex. Ex.
Ex. Ex.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
In.
(Rear)
(Front)
Ex. Ex.
Ex. Ex.
Cylinder 5 Cylinder 6
TDC
Cylinder pressure
Intake pressure
LAF output, etc
Fig. 6
Ex. Ex.
Ex. Ex.
Cylinder 7 Cylinder 8
Optical input
SPB2000
PC
DS2000
PC
DS2000
PC
3.2. Combustion Characteristics at 14000 RPM Engine

Speed, Partial Load, and Steady Condition
Figure 8 shows the combustion characteristics of
cylinders 5 and 6 with a throttle position of 30%, an
Schematic diagram of measurement system
120 km/h
300 km/h
300 km/h (Other Intake Box)
Cylinder 5
IMEP [MPa]
Pmax [MPa]
The hardware is comprised of an amp, filter, and A/

D converter (14-bit). It can import data simultaneously
and perform measurement with a resolution of a crank
angle of 1 at an engine speed of 18000 rpm. The light
emission data and pressure data for each cylinder can be
checked on the monitor.
3. Results and Discussion
Cylinder 6
10
9
8
7
6
2.0
1.8
1.6
1.4
1.2
1.1
Lambda
The ram pressure changes in accordance with the

vehicle velocity, then the air flow pattern inside the air
box changes. Therefore, the air inflow amount of each
cylinder also changes. The air inflow amount also
changes in accordance with the shape of the air box. In
particular, unstable combustion may be triggered under
transient conditions. The combustion analysis of the V8
Formula One engine was performed under such
conditions. In the analysis, the Lambda around the spark
plug, the flame propagation speed (FPS), reaction zone
thickness, IMEP, P max, induction pressure and exhaust
pressure were calculated, and the combustion difference
and time series combustion variation were compared.
1.0
0.9
0.8
FPS [m/s]
20
15
10
5
0.8
Fig. 7
91
0.9
LAF-L
1.0
0.8
0.9
LAF-L
1.0
Combustion characteristics at 18500 rpm

and WOT
engine speed of 14000 rpm, and steady condition. It

shows that the Lambda value shifts from the set value,
and that the flame propagation value and the value of
its variation changes significantly.
For cylinder 5, it is shown that the Lambda around
the spark plug is lean under all conditions, and that a
bias occurs in the A/F distribution around the spark plug
under a partial load. For cylinder 6, the Lambda value
is closer to the set value than it is for cylinder 5.
IMEP [MPa]
Pmax [MPa]
Cylinder 5
However, its FPS variation is larger than for cylinder 5,

and it can be found that significant combustion
instability occurs even though the formation of the air/
fuel mixture is even.
In this way, the relationship between the set Lambda
and the Lambda around the spark plug is greatly
different even in cylinders that adjoin each other, and
the corresponding time variation may also vary greatly.
This is proof that although the macro-level state of all
the cylinders can be ascertained from the pressure and
LAF sensors, the detailed state around the spark plugs
cannot be ascertained, and so there is a limit to the
optimization that can be performed based on these
measured values.
Cylinder 6
10
6
2
2.0
3.3. Combustion Characteristics in Transient

Conditions
Combustion analysis in transient conditions was
performed to confirm for misfiring and torque loss in
transient conditions and identify why they occur.
Measurement was performed at the set Lambda of 0.8
while accelerating at a ratio of 3000 rpm/sec from the
engine speed of 9000 rpm. Figure 9 shows the P max,
IMEP, Lambda and FPS for cylinders 5 and 8. The
Lambda graph shows three types of data. (a) is the
Lambda around the spark plug derived from the
Cassegrain sensor, (b) is the value from the LAF sensors
installed on the exhaust pipes of each cylinder, and (c)
is the set Lambda. The horizontal axis is the number of
cycles, and the 4500th cycle is the start of acceleration.
Combustion variations that involve misfiring were
observed from cycles 4530 to 4540. The IMEP also
dropped dramatically.
1.6
Lambda
1.2
1.1
0.9
FPS [m/s]
0.7
20
15
10
5
0.8
1.0
0.8
0.9
LAF-L
1.0
Combustion characteristics at 14000 rpm and

30 % throttle opening
Cylinder 5
Pmax [MPa]
10
0
2
1.0
0.8
1.2
Lambda
(a) MCS
(b) LAF
(c) Target
0.8
0.6
30
30
20
10
0
4500
4550
Cycle
4600
(a) MCS
(b) LAF
(c) Target
1.0
0.6
FPS [m/s]
Lambda
1.2
Fig. 9
Cylinder 8
2
IMEP [MPa]
IMEP [MPa]
Pmax [MPa]
10
FPS [m/s]
Fig. 8
0.9
LAF-L
20
10
0
4500
4550
Cycle
4600
Cyclic variations of combustion characteristics under transient condition
92
In cylinder 5, the Lambda value dropped but did not

reach misfiring. However, the A/F around the spark plug
was too rich, leading to a drop in output.
On the other hand, the misfiring phenomenon can be
clearly seen in cylinder 8. In contrast to cylinder 5, it
misfires because the A/F is too lean. After the misfiring,
combustion does not start again immediately. It starts
after four cycles.
This is likely the effect of the rigid vortex that occurs
in the air box. A recirculating vortex is formed in the
air box by the inertia force of the induction air. This is
likely to result in different air fuel ratios between
cylinders. This fuel bias state occurs even in steady
conditions. A corresponding amount of fuel injection is
set. However, it was found that a setting based on a
steady condition is insufficient under a transient
condition.
Because the vehicle is always driven under transient
conditions during a race, engine calibration is required
for these conditions.
includes the value around the spark plug, the LAF

measurement value, and the set value.
Looking at the FPS, IMEP, and Lambda around the
spark plug, it can be seen that there is no ignition in the
first cycle after the fuel recovery. However, in the next
cycle the mixture is rich and there is an ignition, and in
the following cycle, the mixture changes to a lean one.
It can be clearly observed that the Lambda around the
spark plug changes alternately between rich and lean,
and there is significant variation between cycles.
It was demonstrated that the micro-Cassegrain sensor
can measure the Lambda for each cycle and has a high
responsiveness.
On the other hand, because the responsiveness of the
LAF sensor is low, it cannot measure the behavior of
each cycle. It detects the Lambda value after fuel cut
recovery with a delay of eight cycles.
4. Conclusion
The results obtained when combustion diagnosis of
a Formula One engine has been performed using a
micro-Cassegrain sensor are described below.
(1) Except under certain conditions, in WOT operations with
an engine speed of 18500 rpm, average cylinder Lambda
and Lambda around the spark plugs are approximately
equal and so are in a desirable state.
(2) Under the operating conditions described above, the FPS
is not influenced by the Lambda.
(3) The tendencies of the A/F distributions between cylinders
differ under WOT and partial throttle conditions.Even if
the average A/F in the cylinder is set to an appropriate
value using the LAF sensor, the A/F around the spark
plug may not be appropriate.
(4) The A/F behavior during transient operation was shown
to be different from that during steady operation. When
torque loss or unstable combustion occurred during
transient operation, the A/F around the spark plug was
too rich or too lean. Further, it was shown that the A/F
behaviors were different between cylinders even during
the same cycle.
(5) In the future, highly accurate fuel control will be required
under transient conditions, and to achieve this, highly
responsive measurement instruments will be required.
The micro-Cassegrain sensor provides this function, and
by using this instrument, it may be possible to accelerate
future engine development.
Pmax [MPa]
3.4. Combustion Characteristics in Fuel Cut and

Recovery Cycles
In order to perform torque control, the fuel injection
control was performed in which the fuel was cut during
deceleration and the fuel injection was recovered during
acceleration. At this time, sometimes unstable
combustion occurred during fuel recovery. In response,
the fuel recovery state was simulated on a test bed, and
the combustion state was measured. Figure 10 shows
the results. It shows the P max, IMEP, MFB 0 to 10%,
Lambda and FPS cycle data from the fuel cut to recovery
for cylinder 6. The same as in Fig. 9, the Lambda graph
10
5
MFB 0 - 10% [deg] IMEP [MPa]
0
2
1
0
100
50
0
Lambda
1.5
LAF sensor delay
(a) MCS
(b) LAF
(c) Target
References
1.0
(1) Neo, G. H., Collings, N.: Pressure Data Analysis of

Formula One Racing Engines, SAE Paper, 970061
(1997)
(2) Alten, H., Illien, M.: Demands on Formula One Engines
and Subsequent Development Strategies, SAE Paper,
2002-01-3359 (2002)
(3) Shimasaki, Y., Sakaguchi, J., Maki, H., Suzuki, H.,
Kondo, N., Yamada, T.: Study on Combustion
Monitoring System for Formula One Engines Using
Ionic Current Measurement, JSAE Technical Paper, No.
FPS [m/s]
0.5
30
20
10
0
6900
6950
7000
Cycle
Fig. 10
Cyclic variations of combustion characteristics

in fuel cut recovery test
93
20055017 (2005)
(4) Ikeda, Y., Nishiyama, A., Kawahara, N., Tomita, E.,
Baritaud, T.: Application of IR Absorption Method to
a Racing Engine, 2007 JSAE Annual Congress
proceedings, No.13-07 (2007)
(5) Ikeda, Y., Nishiyama, A., Bariatud, T.: Flame Speed
Measurement of a Racing Engine by IR Method and
Chemiluminescence Method, 14th Lisbon International
Symposium on Applications of Laser Techniques to
Fluid Mechanics (2008)
(6) Ikeda, Y., Nishiyama, A., Kim, S. M., Takeuchi, A.,
Winklhofer, E., Baritaud, T.: Cyclic Variation of Local
A/F and Mixture Quality in SI Engine Using Local
Chemiluminescence, Proc. of 7th international
symposium on internal combustion diagnostics, p. 277284 (2006)
(7) Ikeda, Y., Jeong, H., Nishiyama, A., Baritaud, T.:
Cylinder to Cylinder A/F Fluctuations Measurement
in a Racing Engine by Chemiluminesence, Porc. of 7th
International Conference on Modeling and Diagnostics
for Advanced Engine System, p. 535-540 (2008)
(8) Lee, J. G., Kim, K., Santavicca, D. A.: Measurement
of Equivalence Ratio Fluctuation and Its Effect in Heat
Release During Unstable Combustion, Proc. of the
Combustion Institute, 28, p. 415-421 (2000)
(9) Ohyama, Y., Ohsuga, M., Kuroiwa, H.: Study on
Mixture Formation and Ignition Process in Spark
Ignition Engine Using Optical Combustion Sensor,
SAE Paper, 901712 (1990)
(10) Ikeda, Y., Kaneko, M., Nakajima, T.: Local A/F
Measurement by Chemiluminescence OH*, CH* and
C2* in SI Engine, SAE 2001 Transactions, Journal of
Engines, Section 3, 110 (2001-01-0919) (2001)
(11) Ikeda, Y., Nishihara, H., Nakajima, T.: Measurement
of Flame Front Structure and Its Thickness by Planar
and Local Chemiluminescence of OH*, CH* and C2*,
SAE 2001 Transactions, Journal of Engines, Section
3, 110 (2001-01-0920) (2001)
(12) Ikeda ,Y., Kojima, J., Hashimoto, H.: Local
Chemiluminescence Spectra Measurements in a HighPressure Laminar Methane/Air Premixed Flame, Proc.
of the Combustion Institute, 29, p. 1495-1501(2002)
(13) Kojima, J., Ikeda, Y., Nakajima, T.: Multi-point timeseries observation of optical emissions for flame-front
motion analysis, measurement Science and Technology,
Vol. 14, P. 1714-1724 (2003)
(14) Ikeda, Y., Kawahara, N., Tomita, E.: Time-series A/F
Analysis in a SI Engine by Micro-Local
Chemiluminescence Technique, Proc. of 6th
COMODIA 2004, Paper No.C3-3 (2004)
(15) Kawahara, N., Tomita, E., Takeuchi, A., Arimoto, S.,
Ikeda, Y., Nishiyama, A.: Measurement of flame
propagation characteristics in an SI engine using microlocal chemiluminescence technique, SAE Paper, 200501-0645 (2005)
(16) Ikeda, Y., Nishiyama, A., Kawahara, N., Tomita, E.,
Arimoto, S., Takeuchi, A.: Spark-plug-in sensor for
initial flame analysis and its flame structure in an SI
engine, SAE Paper, 2005-01-0644 (2005)
Author
94
Naoki HANADA
Atsushi HIRAIDE
Yuji IKEDA
Atsushi NISHIYAMA
Kohei YAMADA
Development of Wind Simulator Equipment for

Analysis of Intake Phenomena
in Formula One Engines
Satoshi NAKAMURA*
Yasuhiro MOTOHASHI*
Shuichi HAYAKAWA*
ABSTRACT
Formula One vehicles repeatedly accelerate and decelerate, reaching maximum speeds of 300-370 km/h and facing
maximum G forces of 4.0 G. Under driving conditions such as these, the vehicles wind speed (ram pressure) has a
significant effect on engine power. However, due to issues involving the fitting of measurement devices onto vehicles,
the prioritization of vehicle set-up work, and other factors, time has not been available for measurements to be taken,
and analyses of air intake phenomena due to ram pressure during circuit driving have not been conducted.
A wind simulator able to reproduce ram pressure during circuit driving was developed, enabling efficient
development of intake systems using bench tests.
In addition, dyno equipment with identical inertia characteristics to those of a vehicle drivetrain and a vehicle
simulator system able to simulate body characteristics, throttle and accelerator work, and the conditions of different
courses were added to the system. This enabled circuit driving modes to be used in bench tests, making it possible to
establish appropriate intake settings for each circuit.
1. Introduction
circuit driving modes enabled engine phenomena during

circuit driving to be reproduced.
The developed equipment, Real Vehicle Dyno bench
(RV bench), consisted of a wind simulator, dyno
equipment displaying identical inertia characteristics to
the drivetrain of an actual vehicle, and a vehicle
simulation system.
The vehicle simulation system was based on the
Vehicle Simulation Module (VSM) produced by AVL
List GmbH.
In addition to controlling engine torque based on
engine speed and throttle opening via the dyno
equipment, the system reproduced driving conditions
through substitution of different vehicles, drivers, and
courses, enabling lap times to be evaluated. The system
featured a variety of modes, including cornering
(deceleration and acceleration), full lap driving, and antistall (idling and take-off).
Because the system enabled evaluation of the effect
of the parts employed on lap time in addition to power
output, it provided race teams with a more informed
perspective, and increased the efficiency of development.
This paper will provide an overview of the system,
its correlation with circuit driving, and the results of its
use in the analysis of engine air intake phenomena.
Recent Formula One regulations have reduced the

number of days for running tests and placed limitations
on the annual driving distance in order to reduce costs.
This has necessitated a thoroughgoing reduction of
engine-related running tests via higher accuracy of
performance predictions at the diagram stage and the use
of bench tests in place of running tests.
In order to realize this goal, items from running tests
were not simply substituted in bench tests; a device was
developed that would enable phenomena in the engine
during circuit driving to be clarified.
The factors that affect dynamic performance in
engine-related running tests are changes in the vehicles
wind speed (ram pressure) and atmospheric conditions
(temperature and humidity). Conventional bench tests for
intake system parts were unable to reproduce the
vehicles wind speed (ram pressure), temperature, and
humidity conditions in the high-speed range for all
circuits, and there were therefore cases in which
specifications determined on the basis of bench test
results were ineffective and were not applied.
A wind simulator (ram pressure generator) was
developed as a means of resolving this issue. In addition,
the use of a vehicle simulation system able to employ
95
2. Overview of Equipment
it was sent to the first-floor test room.

Three air ducts were set up in the first-floor test
room, supplying air at a uniform wind speed
(reproducing the vehicle wind speed during circuit
driving) to the air box inlet, radiator, and oil cooler.
Wind could be blown by the equipment in response
to simulated vehicle speed (0-360 km/h) while
meteorological conditions (temperature and humidity)
2.1. Overview of Wind Simulator

Figure 1 shows a section view of the RV bench, and
Fig. 2 shows an overhead view of the test room. Figure
3 shows the bench test equipment with an engine fitted.
The air blowers were designed for transient
performance enabling wind speed to be varied to
reproduce typical acceleration (1.7 G) and deceleration
(4.0 G) conditions for Formula One vehicles. In order
to minimize temperature increases through adiabatic
compression, the wind speed distribution was rendered
uniform by positioning variable-flow cooling fins, the
shapes of which had been designed based on CFD, in
the air ducts.
To enable the temperature of the air used in the
system to be adjusted appropriately in response to
temperature-setting commands, a boiler and a
refrigeration machine were positioned in the second-floor
test room, and the air produced here was sent to the firstfloor blower room.
Three blower motors were used to bring the air in
the blower room to the target wind speed, after which
First floor test room
Air duct
Radiator
Oil cooler
Air box
Fig. 3
Overview of test room
Inlet duct
Exhaust duct
Engine exhaust fan
Exhaust duct
Inlet duct
Inlet duct
Second floor
Air conditioning unit
Air flow
Air ducts
F1 car
Blower room
First floor
Exhaust ducts
Blower motors
Fig. 1
Section view of RV bench
Blower fan room
First floor
L-side fan
L- dyno
Exhaust
duct
F1 chassis
Center fan
Current plate
R- dyno
R-side fan
Fig. 2
Top view of 1F test room
96

Table 1
Control range
Windspeed
0-360 km/h (Max 380 km/h)
Acceleration and deceleration
Max 4.0 G
Temperature
10-40 C
Humidity
30-90% (Rh)
Table 2
In order to achieve a correlation between the wind

speed distribution during circuit driving and in the RV
bench, airspeed was measured by taking measurements
of the pressure distribution at the air box inlet using a
pitot tube sensor (Fig. 4).
It was determined that the wind speed distribution in
the RV bench was changed by 1) the vehicles
windscreen, and 2) changes in the air flow produced by
the drivers helmet.
The effect of these factors on pressure distribution
and engine output was analyzed in order to enable the
pressure distribution in the RV bench to be correlated
with the pressure distribution during actual driving.
Figure 5 shows the complete RV bench test room,
and Fig. 6 shows the total pressure distribution at the
air box inlet.
When the total pressure distribution during actual
driving at an engine speed of 18500 rpm and vehicle
speed of 300 km/h [total pressure distribution (a)] was
compared with the total pressure distribution in the initial
RV bench under the same conditions [total pressure
distribution (b)], it was discovered that while the total
pressure distribution at the air box inlet in the RV bench
was uniform, part of the lower section of the air box
displayed a lower pressure during actual driving.
System specifications
<Dyno motor>
AC-motor mass Inertia
0.767 kgm2
2,000 rpm
M dnom
P mon
400 kW
N max
3700 rpm
<Gear box>
Max input speed
Transfer power
Gear ratio
22000 rpm
400 kW 2 (Max 800 kW)
9
0.015 kgm2
Inertia
Acceleration
Control speed
130000 rpm/sec
9 kHz cycle
were controlled. The equipment was centrally controlled

from an operating panel and a control panel, enabling
wind speed, temperature, humidity, transient speed, and
exhaust duct air flow to be controlled collectively.
The equipment was linked to an engine-side dyno
controller and a vehicle simulation, creating a system in
which a vehicles wind speed could be varied in response
to vehicle speed during a circuit driving simulation. The
system was able to reproduce the atmospheric conditions
of all the circuits used in Formula One racing. Table 1
shows the control range for the system.
2.2. Ultra-low Inertia Dyno Equipment
If the system was able to operate in circuit driving
modes, it would become possible to conduct control
system development, previously only possible using
running tests, in a more concentrated manner. This
would enable a consistent development process, in which
control system development would be conducted
following the evaluation of hardware and devices.
Ultra-low inertia dyno equipment displaying identical
inertia characteristics to the drivetrain of an actual
vehicle was introduced to the system to enable circuit
driving modes to be realized. Table 2 shows the
specifications of the AVL-manufactured equipment.
Pitot sensor
Fig. 4
Pitot sensor position
Wind direction
3. Correlation with Circuit

Air duct
3.1. Pressure Distribution at Air Box Inlet

The wind simulator produced a uniform distribution
of wind speed in the air ducts. However, under actual
driving conditions, the pressure distribution at the air box
inlet is not uniform, and it was necessary to reproduce
this condition in the bench tests. Naturally, the ideal
situation is to ultimately render wind speed uniform.
Helmet angle=16.5
Fig. 5
97
View of RV bench test room
One difference between actual driving conditions and

the RV bench was the windscreen used on the vehicle.
The windscreen, a plate that sits at the front part of the
cockpit and blocks wind, is shown in Fig. 7.
A dummy windscreen that produced the same effect
as the actual windscreen and a vertical fin were
positioned at the outlet of the RV bench (Fig. 8), and
measurements were taken. The total pressure distribution
under these conditions [total pressure distribution(c)]
reproduced the lower pressure in the lower section of the
air box, indicating that the system was now able to
simulate actual driving conditions.
Next, in order to determine the effect of the total
pressure distribution at the air box inlet on engine output,
comparative tests were conducted on the initial RV

bench (b) and the RV bench following modification (c).
At vehicle speeds of 240 km/h or more, there was a
difference of 6 kW in output, equating to a difference
of 0.12 sec in lap times. In order to correct the uneven
pressure distribution, consideration of the vehicle shape
and positioning of the roll hoop were important factors
in addition to the shape of the air box.
3.2. Change in Amount of Fuel Compensation between
Cylinders in relation to Vehicle Speed
Given the fact that optimization of the air/fuel ratio
in each of the cylinders will increase power, fuel
compensation values have been studied using
conventional bench testing equipment, and refined on the
circuits used for Formula One races. The system
discussed here, however, enabled simulation of
atmospheric conditions and maximum vehicle speeds,
and thus did away with the necessity for refinement of
fuel compensation values on actual circuits. This enabled
the person-hours necessary for formulating settings to be
Z (mm)
Area measured
140
135
130
125
120
115
110
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
140
135
130
125
120
115
110
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
0 5
10 15 20 25 30 35 40 45 50
140
135
130
125
120
115
110
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
0.70
0.71
0.72
0.73
0.74
0.75
0.76
0.77
0.78
0.79
0.80
0.81
0.82
0.83
0.84
0.85
0.86
0.87
0.88
0.89
0.90
0.91
0.92
0.93
0.94
0.95
0.96
0.97
0.98
0.99
1.00
0 5 10 15 20 25 30 35 40 45 50
Y (mm)
0 5 10 15 20 25 30 35 40 45 50
Y (mm)
Y (mm)
Total Cp distribution
(18500 rpm, WOT, 300 km/h)
Total Cp distribution
(18500 rpm, WOT, 300 km/h)
Total Cp distribution when 225 mm

windscreen is added
(18500 rpm, WOT, 300 km/h )
(a) Track
(b) RV bench
(c) RV bench after modification
Fig. 6
Total Cp at inlet air box
Dummy windscreen
Windscreen
Vertical fin
Fig. 7
Windscreen
Fig. 8
98
Vertical fin and dummy windscreen
reduced and data for pit strategy, such as the amount of

fuel to be carried by the vehicle, to be formulated in a
more timely fashion. Figure 9 shows the fuel
compensation values for each cylinder resulting from
different ram pressures. The cylinders are compared at
vehicle speeds of 120 km/h and 340 km/h, with engine
speed set at 18500 rpm. On the low vehicle speed and
high vehicle speed sides, increases in ram pressure result
in changes in the flow of air in the air box, and the fuel
allocation to each cylinder (fuel compensation) therefore
also changes. As a result, the fuel compensation values
for the forward cylinders (#1, 2, and 5), which receive
minimal air intake at 120 km/h, increase by 2-4% at 340
km/h. Fuel compensation increases output by 3 kW.
was insufficient to reproduce the temperature

environment of an actual vehicle.
The results of analysis of the temperature increase in
the front of the air box showed that the increase was
synchronized with changes in the oil temperature in the
oil tank positioned between the engine and the bulkhead
rather than radiant heat from the engine and the exhaust
pipe, and the front of the air box was thus being affected
by heat from the oil tank.
The fact that the bulkhead was an enclosed structure
with no channels for cooling air resulted in an increase
in the temperature of the front wall of the air box, and
a consequent increase in the temperature of the intake
air. Results showed that the temperature of the wall
reached 90 C.
To help resolve this issue, a duct for cooling air was
added between the bulkhead and the engine.
As Fig. 12 shows, this enabled cooling air to flow
around the oil tank and between the V-banks, reducing
the temperature of the air box wall to 45 C and the
temperature of the intake air at the front of the air box
by 6 C.
3.3. Effect of Intake Air Temperature in Engine Cowl

A monocoque used in an actual vehicle was placed
in the RV bench, and temperature measurements were
conducted in the air box and the engine cowl.
Measurements were taken of the:
(1) Temperature of the intake air at the air box inlet
(2) Temperature distribution inside the air box
(3) Temperature distribution around the bulkhead
(4) Temperature distribution around the rear of the engine
(5) Temperature distribution of the pneumatic air cylinder
(6) Temperature distribution around the oil cooler
(7) Temperature distribution around the radiator
Figure 10 shows the interior of the air box.
The temperature distribution in the air box (Fig. 11)
was the point of focus among the measurement results.
The results showed that the temperature at the front of
the air box was 8 C higher than the temperature at the
rear. The cylinders at the front were taking in higher
temperature air than the cylinders at the rear.
This was a phenomenon that could not be reproduced
without using a monocoque and engine cowl in the
system; the use of an engine in isolation in bench tests
Front
Rear
Inlet air box
Engine
Front side
NE: 18500 rpm
120
115
At 120 km/h
At 340 km/h
110
105
Rear side
L_Bank
R_Bank
100
95
90
Fig. 10
View of engine cowl and air box
85
80
35
75
5
4
3
2
1
0
-1
-2
-3
-4
-5
30
Temp. (C)
Value
Compensation value (%)
Reduction
6C
Difference
(front-rear)
8C
20
15
10
5
0
Entry
Cylinder No.
Fig. 9
25
Track
RV dyno
With cooling duct
Speed-fuel distribution for each cylinder
Fig. 11
99
Front
Rear
Temperature distribution in air box
Because, based on the regulations, Formula One

engines have not been able to employ variable induction
systems from 2006, when these types of changes occur
in the temperature of the intake air, the engine speed for
peak output changes. An increase in the temperature of
the intake air may reduce intake efficiency and become
a factor in reducing output.
Heat insulating sheets were tested as a measure
against radiant heat, but it was found that temperature
increased over time, and that the sheets had no effect in
10-lap mode.
Scavenging by the cooling duct would be an effective
measure, but it would be necessary to study the shape
of the cooling duct with consideration of its effect on
the aerodynamic performance of the vehicle.
5. Conclusion
(1) A wind simulator (ram pressure generator) and a vehicle
simulation system with circuit driving modes were
developed, and used to reproduce engine intake
phenomena during circuit driving on the bench.
(2) The establishment of a correlation between circuit driving
and bench test conditions enabled the length of time
necessary for air box development, the formulation of
settings for fuel compensation between cylinders, the
selection of intake pipe length, and other tasks to be
reduced.
(3) The use of a monocoque vehicle in the system enabled
the effects of heat damage to be verified, contributing to
the achievement of increased power in actual race
vehicles.
Ideal flow for

cooling wind
Airbox
Oil tank Airbox
V bank
Fig. 12
Cooling duct (ideal flow)
4. Substitution of Data in the System

The RV bench enabled data for settings measured in
running tests to be verified in bench tests. The following
items were substituted in bench tests:
(1) Pressure and temperature measurements from new
vehicle shakedowns
(2) Radiator and oil cooler function checks
(3) Verification of vibration and heat damage (verification
of reliability of parts)
(4) Verification of performance in relation to atmospheric
conditions and ram pressure
(5) Verification of performance and reliability of electrical
system
(6) Evaluation of drivability using vehicle simulation
(7) Evaluation of lap time, evaluation of fuel efficiency
(8) Settings for control data
(race start, traction control, torque mapping for each
circuit)
Author
Satoshi NAKAMURA
100
Yasuhiro MOTOHASHI
Shuichi HAYAKAWA
Measurement Technologies for Formula One

Engines
Nagao YANAGISAWA*
Kazushi OGIYAMA*
Naoki HANADA*
Tetsuo GOTO*
Takeru HAMAKAWA*
ABSTRACT
In order to attain an accurate grasp of physical phenomena in the high-speed measurement environment of a Formula
One engine, it is necessary to be equipped with measurement systems capable of high-volume, high-speed sampling.
In addition, the accuracy of sensors and the vibration resistance of measurement equipment are important factors. It
is also necessary to reduce the size and weight of measurement systems in order to conduct measurements during
circuit driving. To respond to these performance demands, a variety of high-performance engine measurement systems
were used in Formula One development projects.
This paper will discuss combustion diagnosis using a misfire detection system, measurement of combustion pressure
and crankshaft behavior during circuit driving, oil pressure measurement, friction measurement, and the visualization
and measurement of the fuel spray and in-cylinder flows.
1. Introduction
Table 1
For all automotive engines, including the engines

used in Formula One vehicles, engine development is
inextricably bound to technological progresses and
evolutions in the field of measurement. It has become
standard procedure throughout the industry to determine
specifications through the analysis of phenomena such
as the vehicles dynamic behavior, as typified by
vibration and other parameters, combustion pressure,
which is intimately related to the level of power of the
vehicle, and the internal pressure of the inlet and exhaust
pipes, which determines filling efficiency.
As Table 1 shows, the evolution of measurement
technologies between Hondas second Formula One era
(1983-1992) and third Formula One era (2000-2008)
made it possible to conduct onboard measurements of
combustion pressure, vibration, and dynamic behavior in
an actual vehicle up to an engine speed of 20000 rpm.
This technological evolution resulted from increased
efficiency in packaging, for example the reduction of
sensor size, increases in CPU speeds, and increases in
measurement capacity, which helped to enable highaccuracy measurements in shorter periods.
With these advancements in measurement
technologies, the range of areas in which measurements
could be taken increased, and it became possible to
obtain a more accurate understanding of engine
phenomena.
As a result, it became possible to rapidly determine
Comparison of contents of measurements
Item
Second era
Third era
Combustion
pressure
Only one cylinder is measured

(Only dyno)
All cylinders are measured

(Dyno and circuit)
Crankshaft
twist vibration
Only dyno
Dyno and circuit
Engine
vibration
Only dyno
Dyno and circuit
Gear train
vibration
Measurement is impossible
Measurement of all gears

is possible (Only dyno)
engine specifications and resolve issues, and thus to

supply high-quality engines within a shorter time frame.
At the end of 2006, new regulations placed
restrictions on changes to almost all engine
specifications. From 2007, engine development focused
on the achievement of increased power and reliability by
means of small changes within the scope allowed by the
regulations, and it was necessary to change development
methods accordingly.
It was important to make accurate decisions on
whether or not to employ specific proposals within a
restricted scope of potential changes, necessitating a
transition to methods incorporating computer aided
engineering (CAE) in order to study results in advance.
The establishment of an accurate correlation between
actual phenomena and CAE results is an important factor
in the introduction of CAE, and thorough studies were
therefore conducted using single-cylinder engines.
This paper will provide an introduction to the latest
measurement technologies for Formula One engines.

101
2. Measurement Technologies
2.1. Combustion Measurements
2.1.1. Misfire detection system (MDS)
Because of the high speed of Formula One engines,
the combustion process is completed in a short period.
In addition, the on-off load when switching from wideopen throttle to fully-closed throttle is constantly
repeated. Because of this, the frequency of use of the
engines in an unstable state of combustion is high. A
misfire detection system (MDS) that is able to diagnose
the state of combustion under these conditions using
ionic current was developed, and employed in all races
from 2002.
Figure 1 shows the mechanism of generation of ionic
current. It has long been known that radical ions are
generated during the combustion process. Ionic current
is detected using the following two processes, and the
waveforms have two peak values(1):
(1) Detection of relatively long-lived C3H3+ radical ions
(chemical ions) generated when the flame surface passes
through the electrode during the initial stage of
combustion.
(2) Detection of NO2+ radical ions (thermal ions) generated
by thermal dissociation of N2 in combustion gases at
temperatures of 2000 K or higher.
Figure 2 shows the principle of ionic current
measurement. With the central electrodes of the spark
plugs used as ion probes, a positive electrical potential
of approximately 300 V is impressed, and the
C 3H 3 + ,CHO + , and NO 2 + radical ions generated by
combustion are captured.
The ignition coil energy is stored in the condenser
and used as a power source.
Hondas Formula One engines utilized a condenser
discharge ignition (CDI) system with a discharge time
of 70 s. This minimized the effect of ignition noise,
helping to enable the greater part of the combustion
period to be monitored.
Lean air-fuel mixtures are used during low fuel
consumption operation, but in this state engine
hesitation can occur and acceleration performance out
of corners can decline. Engine hesitation results from
misfires, or combustion states close to misfire, due to
the leanness of the air-fuel ratio falling below the limit

for combustion with a rapid increase in the volume of
intake air.
The MDS was developed in order to analyze this
phenomenon.
Figure 3 shows a comparison of detected MDS
values for the A/F setting for maximum power and a
lean A/F ratio during acceleration. For the maximum
power A/F ratio setting, the transition from a lean to a
rich ratio occurs in a short period. The period during
which MDS values are generated is also minimized, and
combustion stabilizes rapidly. However, for the lean A/
F ratio setting, the period during which MDS values are
generated continues, and an unstable combustion state
continues. This generates conspicuous hesitation.
The development of a technology that helped to
enable monitoring of the state of combustion during
vehicle operation made it possible to set appropriate A/
F ratios for the atmospheric conditions of each circuit.
2.1.2. Combustion pressure sensor integrated with
spark plug (PS plug)
Combustion pressure measurements were conducted
constantly during the Formula One engine development
program, but the following technological issues arose
with regard to measurements using combustion pressure
sensors:
(1) Low degree of freedom of layout
Limitation of areas for mounting of sensors due to engine
configuration, resulting in inability to measure pressure
in all cylinders.
CDI +300 V (electrode bias)
ct
je
In
on
Ionic current
Injection
Combustion
Electron e
H3O+
C3H3+
Fig. 2
CHO Radical ion
Ionic current measurement principle
Ignition noise
High
Engine speed
Ignition noise
Throttle opening
100
%
Low
0
Thermal ion
Chemical ion
Lean
A/F = 14.5
Rich
Flame surface
e
e
Electron
C3H3+
2000 - 3000 K
NO2+
Thermal dissociation
H3O+
+ Radical ion
CHO
CmHn + O2
+
H2O + CO2
+
C3H3 , CHO , H3O
Fig. 1
A/F = 13.1
Fire
A/F R/L bank

(best power setting)
Misfire
MDS output
(lean setting)
Fire
N2 + O2
NOx
+
NO2
A/F R/L bank

(lean setting)
MDS output
(best power setting)
Misfire
Time (s)
Fig. 3
Mechanism of ionic current
102
Detected misfire
(2) Low level of durability and reliability

Sensors damaged by a high level of vibration due to high
engine speeds.
Given this situation, a combustion pressure sensor
integrated with a spark plug (termed a PS plug below)
was developed as a technology for conducting pressure
measurements in all cylinders.
Figure 4 shows the shape of the PS plug, and a
measured combustion pressure waveform.
Comparison of combustion pressure measurements
obtained using the PS plugs in test bed tests with
pressure measurements taken using existing sensors
showed that the plugs performed well.
The use of the PS plugs made it possible to bypass
the limitation on sensor mounting positions and measure
pressure in all cylinders, and thus to understand the
differences in combustion pressure between cylinders.
Figure 5 shows the results of combustion pressure
measurements taken during circuit driving and results for
engine power calculated from the indicated mean
effective pressure (Pmi) from the measurement results.
The indicated power calculated from the Pmi on the
circuit displays identical values to those obtained from
a torque meter mounted on the gearbox, indicating that
the PS plug accurately measures combustion pressure
even during circuit driving.
Combustion pressure (kPa)
2.2. Crankshaft Torsional Vibration

Normally, the excitation of torsional vibration of the
crankshaft is determined by the sum of the inertia of the
reciprocating parts such as the pistons and the conrods,
Std sensor
PS plug
Crystal part
Combustion pressure
Seating noise
Diaphragm part
(sensing part)
Center of electrode
0
0
120
Fig. 4
240
360
480
Crank angle (deg)
600
and the rotational torque resulting from combustion

pressure.
Because Formula One engines operate at high speeds,
the inertia of the reciprocating parts is dominant in the
load on the crankshaft pin journals, and the load exceeds
50 kN at its maximum.
The crankshaft has a specific vibration frequency
determined by its specific torsional stiffness and
rotational inertia. However, resonance is generated by the
excitation force mentioned above, and this can result in
large torsional vibrations, the amplitude of which
exceeds 1 deg.
Because Formula One engines use a central oil
supply system, as shown in Fig. 6, part of the crankshaft
is hollow, resulting in a decline in strength against a
solid crankshaft. For this reason, torsional resonance
might cause breakage of the crankshaft.
In the 2002 Formula One engine, torsional vibration
resonating in three engine speed orders at high speeds
under a motored condition became an issue. In terms of
circuit driving conditions, this corresponded to closedthrottle in the straight ends; the frequency of such use
is high, and this has an effect on the durability of the
engine. In fact, crankshafts broke twice on circuits.
Because torsional vibration occurred at high engine
speeds, it represented an impediment to increasing the
speed of the engine in order to increase power.
The analysis system configuration shown in Fig. 7
was used to measure the torsional vibration of the
crankshaft on the bench and on the track, and the origins
of the phenomenon were analyzed.
Because high-speed sampling was necessary, a data
logger able to measure up to 200 kHz was used to record
pulse signals for the detection of crankshaft rotation
signals. These signals were converted into frequencies
and the changing component of the signal was isolated.
A Fourier transform was further applied to the changing
signal component, helping to enable the torsional
vibration of the crankshaft to be understood.
720
Hydraulic pathway
in crankshaft
Combustion pressure force and shape of PS

plug
(%)
100
50
0
20
(x 1000 rpm)
Throttle opening angle
18
Engine speed
16
14
12
(kPa) (km/h)
10
300
200
100
50
Car speed
Ram pressure
0
500
(kW)
300
100
Engine power of
PS plug (R + L)
(kW)
-100
300
100
Fig. 6
Engine power of gear
box sensor value
Crank pulse
from ECU
Section of Formula One crankshaft

Data logger
XR5000-WB
Sampling rate: 150 kHz
FV converter
KAZ-723A
Frequency
filter
FFT analyzer
P-84
L side power of PS plug
Data logger
R side power of PS plug
DR-R1 Mk2
Sampling rate: 200 kHz
-100
Time
Fig. 5
Fig. 7
Circuit running data
103
On dyno
On circuit
Analysis system configuration
DS2000
The results showed that the increase in torsional

vibration originated in the bank angle used to lower the
center of gravity. In the 2003 Formula One engine, the
bank angle was reduced from 94 to 90, reducing
torsional vibration of the crankshaft.
Results were also verified using the vibration
simulation Blicks, manufactured by AVL Japan, K.K.,
in parallel with the measurements.
Figure 8 shows the simulation results obtained using
Blicks and the measurement results.
The level of vibration at high engine speeds under a
motored condition, which was an issue in the 2002
Formula One engine, was alleviated in the 2003 engine.
The good correlation between the simulation results
and the results of vibration measurements in an actual
vehicle demonstrated the effectiveness of the simulation.
As indicated by the discussion above, understanding
the torsional vibration of the crankshaft when operating
at high speeds necessitated high-speed sampling, and
advances in data logger technology made these
measurements possible.
2.3. Hydraulic Pressure Measurements
The major issues for the durability and reliability of
the conrod bearings are wear resistance and seizing
resistance. To address these issues, changes were made
in the bearing materials, and the viscosity, amount,
pressure and aeration of the lubricating oil were
modified.
Formula One engines employ a central oil supply
system, in which the centrifugal force of the crankshaft
is used to supply oil to the conrod bearings.
Because of this, only the pressure in the hydraulic
pathway of the crankshaft could be measured in order
to measure actual hydraulic pressure. This necessitated
a technology to output the signal from pressure sensors
in the hydraulic pathway to the exterior of the engine.
Simulation data
Measurement data
1.4
1.2
1.0
90 deg V-Angle
94 deg V-Angle
0.8
0.6
0.4
0.2
Twist vibration (ddg)
1.2
94 deg V-Angle
1.0
0.8
0.6
90 deg V-Angle
0.4
2.4. Linking Method

The use of the linking method to conduct
measurements is not a new technology, and is used in
mass production engines. However, in high-speed
Formula One engines, the acceleration component of the
inertial force on the links is 10000 G at 19000 rpm,
approximately twice the force in a mass production
engine. This difference necessitates high-stiffness and
ultra-low weight links, and therefore significantly affects
the design of the linking mechanisms.
0.2
0.0
0.0
10 11 12 13 14 15 16 17 18
Ne 1000 (rpm)
Fig. 8
10 11 12 13 14 15 16 17 18
Ne 1000 (rpm)
15000 rpm
2500
Simulation results and measurement results

Oil pressure (kPa)
Twist vibration (ddg)
1.4
Figure 9 shows the configuration of the hydraulic

pressure measurement system.
Previously, a contact-type slip ring using mercury as
a physical signal path had been employed to guide
signals from the crankshaft. However, engine speeds
continued to increase, and the range of speeds at which
this method could be used was limited. This was a result
of the limits of the capacity of seals to deal with the heat
generated and the fluid leaks occurring when mercury
flowed at high speeds in the cylinders.
To replace this method, wireless technology was
employed to transmit the signal outside the crankshaft.
By contrast with the use of the slip ring, the wireless
method involved no contact, and therefore resolved the
associated issues. In addition, increases in the
computational speed of the CPU of the receiver helped
to enable measurements within the actual range of engine
speeds. The noise resistance of this method was low
compared to the contact method, but insulating materials
were used to protect the receiver.
Figure 10 shows an example of measurements taken
in a hollow crankshaft.
Because in a hollow crankshaft the hydraulic pathway
does not have to pass the center of rotation of the
crankshaft, the decline in hydraulic pressure downstream
(due to pressure loss in internal pipes and pressure loss
due to Coriolis force), a constant issue in central oil
supply systems, was minimized and the average
hydraulic pressure increased. The effect of the hollow
crankshaft was estimated in simulations, but pressure
measurements in the internal hydraulic pathway of the
crankshaft verified the effect. In addition, the effect of
noise was minimized, and a good understanding of the
differences between specifications was obtained.
#2Rr
#1Fr
#5Fr
#4Rr
Transmitter
Receiving set
Fig. 9
2000
1500
1000
500
STD crankshaft
0
Hydraulic pathway
Data analysis
Supply oil pressure 950 kPa
Hollow crankshaft
-500
Data recorder
180
360
540
Crank angle (deg)
Fig. 10
Analysis system configuration
104
Pressure in hydraulic pathway
720
Figure 11 shows a linking mechanism used in a

Formula One engine.
For Formula One engines, the linking mechanisms
were incorporated in the single cylinder engines used for
analyses.
To reduce weight, the linking mechanisms were cut
from aluminum alloys. The covers used to hold the lead
wires were all produced from CFRP in order to further
reduce weight. In addition, to maintain stiffness, a box
section was used for the sectional shape of the linking,
helping to ensure a secondary moment acting on the
section. Small NTN bearings with an internal diameter
of 7 mm and an external diameter of 11 mm were
employed in the connections to help ensure a smooth
sliding motion.
In conrods, the end connected to the piston, which
displays a linear reciprocal motion, is called the small
end, and the end connected to the crankshaft, which
displays a rotary motion, is called the big end. The two
ends are joined by the shaft. Because the shafts of the
conrods are constantly subject to complex forces
representing a combination of linear and rotary motion,
stress concentrates in this section, and the magnitude of
this stress can exceed the breakage limit of the shaft
material. Conrod shaft breakages occurred during
Hondas third Formula One era, and the reliability of the
conrods formed an important development agenda.
Figure 12 shows an example of a conrod breakage,
the result of a CAE analysis, and the set-up used in the
linking method.
Link mechanism
Single-cylinder engine
Fig. 11
Link mechanism for Formula One engines
The point of breakage of the conrods was close to

the small ends, where strength was lower. To analyze
these breakages, CAE analysis was employed to suggest
potential reasons for the concentration of stress.
Following this, the development of a measurement
method using the linking method proceeded, and analysis
of the phenomenon in single cylinder engines
commenced in 2004.
Figure 13 shows the results of measurements of
conrod torsional resonance using the linking method. The
measurement results showed that torsional resonance was
produced in the conrods at full engine loads. In addition,
the fact that resonance did not occur when no load was
acting and the inertial force was therefore dominant
indicated that the conrods were twisting due to excitation
by gas pressure and bending of the crank pins.
As this demonstrates, the linking method as used in
Formula One analyses demonstrated a sufficient level of
accuracy even in measurements at 19000 rpm, and the
link mechanism also displayed sufficient durability. This
made it possible to understand internal engine
phenomena such as reciprocating motion, and
technologies for the quantification of these phenomena
contributed in numerous ways to engine development.
2.5. Friction Measurement Using the Floating Liner
Method
The friction generated in the various parts of the
engine is one of the factors that impedes the achievement
of increased power, and the accurate measurement of
friction is therefore an important issue. Friction in the
reciprocating parts represents approximately 60% of total
engine friction, and its mechanism has still not been
clarified. This section will discuss the floating liner
method, a new method of measuring friction in the
reciprocating parts.
Broadly speaking, friction in the reciprocating parts
can be divided into the following three types:
(1) Sliding friction between the pistons and the sleeves
(2) Sliding friction between the conrod bearings and the
crank pins
(3) Friction due to agitation of the oil in the crankcase
Of these, the floating liner method was developed to
measure (1), sliding friction between the pistons and the
sleeves.
Linking system type
350 Hz
2
425 Hz
500 Hz
1
0
18000 rpm
-1
16500 rpm
-2
Twists vibration (deg)
CAE analysis
Twists vibration (deg)
Failure example
18000 rpm
0
-1
16500 rpm
-2
14000 rpm
Strain gauge
Fig. 12
Failure example, CAE result, linking system

type
-3
-3
-4
10000
-4
12000
Fig. 13
105
14000
16000
Ne (rpm)
18000
14000 rpm
180
360
540
Crank angle (deg)
Twist vibration measurement results for

connecting rod
720
Figure 14 shows a cutaway of the test apparatus

used in the floating liner method.
The feature that sets this apparatus apart from standard
floating liner test apparatuses is the fact that it has been
designed for use at high speeds. Standard test apparatuses
are constructed of iron alloys or aluminum alloys, and they
have a range of measurement of up to 5000 rpm. However,
at high engine speeds the liner resonates in the same
direction as the motion of the pistons, and measurement
accuracy declines. To resolve this issue, the liner in the
new apparatus was constructed from beryllium in order to
reduce its weight and control the occurrence of resonance.
In addition, the load sensor was positioned at the top of
the liner to help ensure stiffness in the mounts. This
positioning prevented the mechanism from being used for
the measurement of firing, but in a later specification,
combustion pressure was artificially produced to simulate
firing conditions. Using a mechanism to introduce air to
the combustion chamber, the pistons could be used to
compress the air, making it possible to freely increase
pressure up to 10 MPa.
Figure 15 shows an example of measurements using
the initial apparatus specifications.
The graph on the left shows figures from the load
sensor in the up-down direction (the direction of piston
motion). Adjustment of the crank angle made it possible
for the floating liner method to be used to analyze
friction in each stroke. The integral of each rotation
corresponded to the sliding friction between the piston
and the sleeve in each cylinder, making it possible to
quantify sliding friction.
The graph on the right shows figures from the load
sensor in the left-right direction (the direction of the
rotating surface of the crankshaft, orthogonal to the
direction of piston motion). These results show the
Cylinder pressure control system
Load sensor
(3-axis type)
Outer sleeve
(beryllium)
Inner sleeve
(aluminum alloy)
Initial spec. (no firing)
Fig. 14
Later spec. (dummy firing)
Floating liner mechanism
800
changes in thrust and anti-thrust force for each stroke.

As the discussion above indicates, it became possible
to measure the friction in each stroke and use the results
in analyses of the form of the piston skirt and other
parameters.
2.6. Visualization Technologies
2.6.1. Fuel spray visualization technology
Due to the latent heat of vaporization, the atomization
of the fuel cools the intake air, increases filling
efficiency, and advances combustion. These effects result
from the droplet diameter and the penetration, which
change with the fuel injection pressure, but a
measurement method for quantifying these parameters
was previously unavailable.
The fuel injection pressure of the electronicallycontrolled fuel injection system (PGM-FI) used in the
turbocharged engines employed during Hondas second
Formula One era was 0.25 MPa. Following this, a 1.2
MPa high-pressure system was developed in 1987 and
used for 17 years until 2004, in Hondas third Formula
One era. From 2005, the increase in pressurization was
accelerated and a 5 MPa system was used, with a 10
MPa system following in 2006. This section will discuss
the fuel spray visualization technology used in the
analysis of droplet diameter and penetration in this
process of increasing pressure.
The fuel spray has a three-dimensional form that
changes with the elapse of time. The following
parameters were focused on:
(1) Droplet diameter
(2) Diffusion
(3) Penetration
(4) Form angle and form
Of these, the form angle and form could be controlled
to a certain extent by means of the direction of the
nozzle, but droplet diameter, diffusion, and penetration
are affected by the shape of each injector hole, and a
number of nozzle shapes was therefore used in
measurements.
Figure 16 shows an external view of the
measurement device, a phase Doppler particle analyzer
(PDPA), and the measurement principle.
The PDPA bombarded the fuel spray with a
continuous laser beam, and measured the droplet
diameter and flow speed by applying signal processing
to the scattered light. The flow speed was measured at
the point where the two laser beams intersected.
Fuel particles entered this point, and the scattered
T.D.C.
Integral calculus value
700
Friction load (N)
600
Measurement system
Laser
2.7 kW/18500 rpm
500
400
Anti-thrust force
300
Interference
fringes
Thrust force
200
VZ
100
Laser
0
-100
-200
-300
0
180
360
540
Crank angle (deg)
720
B.D.C.
-1500 -1000 -500
500
1000
Measurement
area
1500
Y
X
Thrust force (N)
Fig. 16
Fig. 15
Sliding friction measurement result
106
External view of PDPA hardware and

measurement principle
light was captured by a receiver and transformed into an

electrical signal. Using a fast Fourier transform (FFT),
the frequency was analyzed and the speed of the particles
was calculated.
Droplet diameter was measured using the phase
difference between burst signals monitored by multiple
photomultiplier tubes. The droplet diameter was
calculated by correcting the frequencies of the light that
was diffracted differently, depending on the size and
shape of the droplets.
Figure 17 shows the distribution of droplet diameters
in the fuel spray, as measured using the PDPA.
To help enable detailed analysis of the PDPA
measurement results, the form of the fuel spray directly
underneath the injector holes was visualized using the
SprayMaster magnified imaging apparatus manufactured
by LaVision GmbH.
Figure 18 shows a fuel spray as imaged using
SprayMaster.
Using SprayMaster, transmission images of the crosssections of sprays were captured using a laser, and
employed in parameter comparisons of fuel injection
pressure and injector holes, and comparisons of the
characteristics of the sprays produced by injectors with
different numbers of holes.
The SprayMaster apparatus contributed to the
achievement of increased efficiency in the development
of fuel systems, assisting in tasks including the
determination of specifications in injector development
and the simulation of in-cylinder fuel distribution.
important to the realization of rapid combustion. The

optical engines were used to enable measurement of flow
and turbulence in the combustion chambers of Formula
One engines.
This section will discuss the engine visualization
method, measurement results, and validation results of
simulation.
Figure 19 shows a cross section of the optical
engine.
As in the case of the linking and floating liner
methods discussed above, a single cylinder engine was
used for the optical engine from considerations of ease
of modification and versatility. A special sleeve made
from silica glass was fitted to part of the cylinder block.
A glass sleeve height of 40 mm made it possible to
visualize the entire cylinder stroke.
Particle image velocimetry (PIV) was used for the
measurement of the flow motion of the gas in the
combustion chamber. The system utilized a YAG laser
with a second-order harmonic of 532 mm, and a crosscorrelation CCD camera for imaging. A synchronizer
was used to synchronize the two devices.
Ultra-lightweight tracers, which are hollow resins
with a diameter of 40 m or less, were used for the
visualization of flows. FFT cross-correlation was used
in the analyses.
Special piston with extended crown was used, and
measurements were performed at engine speeds of
10000 rpm.
Figure 20 shows PIV measurement results and
simulation results.
2.6.2. Optical engine technology

Understanding the flow and turbulence phenomena of
the air-fuel mixture in the combustion chamber is
Volumetric flow distribution
Particle size distribution
4.E-07
2.E-07
Cylinder
head
Quartz sleeve
Flow velocity distribution

30
Flow velocity (m/sec)
20
Particle size (um)
Volumetric flow (mm3/st)
6.E-07
15
10
Quartz
sleeve
25
Fig. 17
0
-90 -60 -30 0 30 60 90
Position from center (mm)
Visual field
of stroke
Head extension
piston
20
15
10
Cylinder
block
0.E+00
-90 -60 -30 0 30 60 90
0
-90 -60 -30 0 30 60 90
Fig. 19
Visualization engine
View
Fuel spray measurement results

[Measurement section]
Direction of Fr-Rr
Laser sheet
Laser
sheet
[Measurement section]
Direction of IN-EX
View
PIV
Fuel spray
Magnification proximity
photography
Simulation (Vectis)
Cross-sectional
photography
1 mm
0
Fig. 18
Macrophotography using SprayMaster
150 m/s
Fig. 20
107
150 m/s
Measurement result and simulation result
PIV measurements helped to enable the phenomenon

of gas flow in the combustion chamber to be understood.
Based on the results of visualization, validation of the
gas flow simulation software VECTIS made by Ricardo
plc was performed. Thereby, it became possible to
accurately predict the flow motion of gas in the
combustion chamber using the simulation. In addition,
by performing similar validations, analysis of the
behavior of fuel spray and the process of combustion
(flame propagation) in the combustion chamber can be
conducted using VECTIS, and the efficiency of
development enhanced.
3. Conclusion
This paper has discussed measurement technologies
employed in the engine development program in Hondas
third Formula One era. The vast majority of these
technologies were unavailable during Hondas second
Formula One era, and it is no exaggeration to say that
they provided the basis for the evolution of Hondas
Formula One engines during the third era. The authors
believe that the fusion of the measurement technologies
fostered by the Formula One development process with
the measurement technologies employed in the
development of mass production engines, will assist in
the development of environmental technologies and in
increasing the efficiency of the development process.
Reference
(1) Shimasaki, Y., Sakaguchi, J., Maki, H., Suzuki, H., Kato,
A., Nishizawa, K.: Study on Combustion Monitoring
System for formula One Engines Using Ionic Current
Measurement, Honda R&D Technical Review, Vol. 17,
No. 1, p. 56-64
Author
Nagao YANAGISAWA
Naoki HANADA
Kazushi OGIYAMA
Tetsuo GOTO
108
Takeru HAMAKAWA
Development of High-Pressure Fuel Supply

System for Formula One Engine
Tetsuya TANAHASHI*
Yosuke SAWADA*
Kazuji ONO*
Atsushi SHIMIZU*
Masanori HAYAFUNE*
ABSTRACT
Important factors in boosting the performance of todays Formula One engines include: the realization of the
formation of ideal air-fuel mixtures and the achievement of greater combustion efficiency, through the use of shorter
fuel injection periods and increased spray atomization resulting from higher fuel pressures; and, in addition to this,
the achievement of stable combustion in the low-load operating range.
A comprehensive analysis of injector spray characteristics was conducted, leading to the development of a Hondamade high-efficiency, high-pressure fuel supply system. This enabled the achievement of a 15 kW increase in engine
power.
1. Introduction
2. Mechanism of Power Increase with Highpressure Fuel Injection

During the development of fuel systems for Formula
One engines, a single cylinder engine was used to
conduct analyses of intake air and fuel phenomena.
Based on the results of these analyses, the highpressure fuel systems were in use in race engines until
2006, and further modifications of the spray morphology
from 2006 onwards helped to enable an increase in
power of approximately 15 kW.
Figure 1 shows changes in the fuel pressure and
power of Honda race engines, and Fig. 2 shows an

109
16
Enhancement in Pse (kW)
The importance of development programs for fuel

systems in automotive engines is increasing as a
means of resolving technological issues related to
increasing fuel efficiency and reducing exhaust
emissions. Given this, automakers, parts makers, and
research organizations are pushing ahead with the
development of more advanced technologies, and are
conducting more sophisticated analyses of relevant
phenomena.
In Formula One engines, the fuel spray has a
particularly significant effect on engine power, and
optimum spray morphologies exist for specific
combustion chamber shapes and intake port shapes.
When specifications are changed in these areas, the fuel
spray is also studied and redesigned.
Against this background, a basic principle of
Hondas Formula One engine development program
was the in-house production of fuel supply systems, in
order to boost competitiveness through the ability to
trace the development process and produce unique
technologies.
This paper will provide a technological overview of
a fuel pump for the supply of high-pressure fuel, a
regulator to control fuel pressure, and an injector to
inject the fuel and form a spray.
2008 (V8)
14
12
Evolution of fuel system
+15 kW
10
8
2000 (V10)
Equivalent in V8
6
4
2006 (V8)
2005 (V10)
Equivalent in V8
2
0
0
Fig. 1
6
8
10
Fuel pressure (MPa)
Evolution of fuel system
12
14
overview of the components of a fuel supply system.

The following three mechanisms can broadly be
indicated as those responsible for the increase in power
achieved when the fuel pressure was increased from its
initial value of 1.2 MPa to 10 MPa and the spray
morphology was modified:
(1) Increased intake air cooling efficiency due to increased
supply flow rate (shorter injection period)
(2) Enhanced combustion due to greater atomization of the
spray
(3) Increased intake air cooling efficiency due to
modification of spray morphology
2.1. Increased Intake Air Cooling Efficiency with
Increased Supply Flow Rate
Under the initial specifications of a fuel pressure of
1.2 MPa and a supply flow rate of 43 L/h, the maximum
injector supply flow rate was low because the fuel
pressure was low, and a long injection period was
necessary in each cycle. By contrast, the use of highpressure fuel increased the supply flow rate and
consequently reduced the injection period, helping to
enable injection for an optimal injection period.
Figure 3 shows the relationship between the intake
air velocity at the funnel tip and the injection period, as
determined using simulations.
The results of this analysis show that the intake air
cooling effect is maximized by ending the injection
period by the timing of maximum blow at the funnel tip.
Fuel pressure
regulator
Injector
Air box
A decline in the temperature of the intake air can be

determined from changes in the intake air pulsation. As
Fig. 4 shows, a short injection period results in a phase
delay in the intake air pulsation, and this can be judged
as indicating that filling efficiency has increased due to
a decline in intake air temperature.
2.2. Enhanced Combustion Efficiency with Greater Fuel
Atomization
The diameter of the fuel droplets, which had been
approximately 50 m at a fuel pressure of 1.2 MPa, was
reduced to 20 m or less at a pressure of 10 MPa. As
Fig. 5 shows, power was increased as a result of this
refinement of the droplets. The mechanism behind this
increase in power is thought to be increased mixing of
the intake air and the fuel, and a reduction in the
adherence of fuel to the port walls.
2.3. Increased Intake Air Cooling Efficiency with
Modified Spray Morphology
As indicated above, the use of increased fuel pressure
resulted in increased cooling air efficiency and increased
combustion efficiency.
The fuel spray morphology was also studied in order
to further boost these effects. Figure 6 shows changes
in shaft output in a single cylinder test engine.
The increase in engine power that occurs when an
increased number of injection holes are employed is due
to increased atomization of the spray. The delay in the
phase of intake air pulsation shown in Fig. 7 indicates
that a decline in the temperature of the intake air has
increased filling efficiency.
Funnel
Difference in phase of intake air
pulsation (deg)
Collector tank
Throttle
valve
Gas bag
Fuel
Engine
Electric fuel pump
Mechanical fuel pump
4
10 MPa - 100 L/h_10 holes
1.2 MPa - 43 L/h_pintle
3
2
1
0
-1
Base value : 100 L/h - 10 holes
-2
-3
-4
1.0
Fig. 2
Components of fuel supply system

Fig. 4
160
In valve open
1.5
Velocity of intake air
80
40
Difference in Psi (kW)
Velocity of intake air (m/s)

(at tip of funnel)
Inflow
0
-40
-80
Outflow
-120
-160
0
Fig. 3
90
4.0
Effect of injection time on difference in phase

of intake air pulsation
Ex valve open
120
1.5
2.0
2.5
3.0
3.5
Period of effective fuel injection (msec)
Period of fuel injection

1.2 MPa - 43 L/h
10 MPa - 78 L/h
10 MPa - 100 L/h
10 MPa - 120 L/h
180
270
360
450
Crank angle (deg)
540
630
720
1.0
2000
2006
0.5
2008
2005
2007
0.0
-0.5
-1.0
Base value : 2006 100 L/h - 10 holes
-1.5
10
Relationship between intake air velocity and

injection timing
Fig. 5
110
Sauter mean diameter ( m)
Effect of droplet refinement on power
100
The relationship between penetration and the phase

of intake air pulsation was also studied. As Fig. 8
shows, there is a peak in the intake air phase delay,
indicating the existence of an optimum penetration for
the achievement of increased intake air cooling
efficiency. If the level of penetration is low, the spray
will approach the center of the port due to the effect
of the intake air. If the fuel is carried too high, the
amount of fuel adhering to the walls of the port will
increase.
A technological overview of the high-pressure fuel
system based on the results of the analyses discussed
above will be provided below.
1.5
Difference in Pse (kW)
Static flow rate 100 L/h

1.0
0.5
0.0
-0.5
Base value : 10 holes
-1.0
0
Fig. 6
10
15
20
25
Number of holes
30
35
40
Effect of number of holes on power
Difference in phase of intake air

pulsation (deg)
2.0
1.0
0.0
-1.0
-2.0
0
Fig. 7
10
15
20
25
Number of holes
30
35
40
Effect of number of holes on difference in

phase of intake air pulsation
Difference in phase of intake

air pulsation (deg)
4
3
2
1
6 holes
10 holes
16 holes
In 6 holes + Out 10 holes
24 holes
0
-1
-2
-3
-4
60
Fig. 8
70
80
90
Penetration (mm)
100
110
Effect of penetration on difference in phase of

intake air pulsation
3. Development of High-pressure Fuel Pump

In Hondas second Formula One era, the engine
fuel supply system was composed of a gear pump
producing a fuel pressure (PF) of 1.2 MPa, a fuel
pressure regulator, and a side-feed injector, which
supplied fuel from its side. At the commencement of
engine development for Hondas third Formula One
era, this basic configuration was carried over, with
only the form of the injector being changed, to a topfeed type that supplied fuel from the upper section of
the engine. Development and deployment of parts
continued with consideration of their suitability to
each model year engine.
Later, from 2005, the focus was shifted to the
development of a system that helped to enable the
achievement of increased intake efficiency and enhanced
combustion by means of refining the fuel spray droplets
from the injector and increasing atomization, and
reducing the injection period, by increasing the pressure
of the fuel supply system.
The achievement of a fuel pressure of 10 MPa was
established as a target for the opening race of 2006. In
the 2005 season, a commercial 5 MPa system had been
used, and development time was concentrated on the
achievement of 10 MPa.
In order to realize increased engine power by
achieving a PF of 10 MPa, it would be essential to
discard the previous gear pump and to conduct an inhouse development of a high-efficiency and lightweight
mechanical fuel pump that would be able to generate the
necessary supply flow rate at high pressure. The reason
for this was that at high fuel pressures, the efficiency
of the previously used gear pump type would decline due
to fuel leaks from the tips of the gear teeth and the gaps
between the gear side faces, and it would be unable to
transport the necessary amount of fuel.
To resolve this issue, the goal was initially to develop
a gear pump configuration employing movable side
plates to maintain fixed gaps between the gear side faces
and moderate the volume of fuel leakage. However, this
represented a technological challenge from the
perspective of durability, raising issues such as the
seizing of the sliding surfaces of the gear side faces.
With the achievement of high fuel pressure and high
efficiency as targets, efforts were therefore focused on
the development of a new plunger piston-type fuel pump.
3.1. Overview of High-pressure Fuel Supply System
Figure 9 shows a diagram of the fuel system fitted
in the fuel tank, incorporating a mechanical high-pressure
fuel pump, a fuel pressure regulator, and primary and
secondary electric fuel pumps. Formula One vehicles use
explosion-proof fuel tanks (called gas bags)
manufactured from plastic liners. These tanks are fitted
in the monocoque, and each year they are manufactured
in conjunction with the monocoque, to dimensions that
match the capacity of the monocoque. The collector tank
helped to prevent abnormalities of suction of the gas bag
fuel from arising in the electric pumps due to turning
111
or acceleration and deceleration G-forces, and

pressurized the fuel to feed pressure for the mechanical
high-pressure fuel pump. The collector tank was divided
into two parts, a high-pressure collector tank and a lowpressure collector tank. Fuel from the gas bag was sent
to the low-pressure collector tank by the primary electric
fuel pumps positioned to the left and right on the floor
of the gas bag. The fuel was then pumped from the lowpressure collector tank to the high-pressure collector tank
by the secondary electric fuel pumps positioned on the
floor of the low-pressure collector tank. The highpressure collector tank employed a bladder, which was
a variable volume pressurizing device using air pressure,
to help ensure that the feed pressure of the fuel did not
fall below the necessary level for supply to the
mechanical high-pressure fuel pump. When no fuel was
present in the high-pressure collector tank, the bladder
swelled to fill the entire tank; when the tank was filled
with fuel, the pressure of the fuel compressed the
bladder. By means of this mechanism, the high-pressure
collector tank could be constantly filled with fuel. The
high-pressure collector tank was also provided with a
pressure relief valve (PRV), which helped to ensure that
the pressure in the tank remained at a constant upper
limit. The bladder and the PRV stabilized the feed
pressure to the mechanical high-pressure fuel pump, and
helped to prevent cavitation when the high-pressure
pump was taking in fuel at high temperatures.
the pump by 30 mm and its weight by 340 g. The

reduction ratios of the reduction gearbox and the
reduction gear pump were optimized against engine
speed in order to balance durability with good flow
performance.
3.3. Selection of Pump Configuration
The pump was provided with a gear pump to help
ensure stable feed pressure to the plunger pump. It was
essential for the gear pump both to be small and
lightweight and to function to prevent cavitation of the
transported fuel when operating at high fuel temperatures
in a circuit driving. A shape was designed that reduced
pressure loss in the pump inlet port and an optimal gear
pump thickness was set in order to reconcile the
achievement of the desired supply flow rate with the
realization of cavitation toughness. In addition, the
optimum plunger piston size was selected and the
resistance of the fuel channels was reduced in order to
increase fuel transport efficiency, and a variety of
modifications were made, including optimization of the
stiffness of the pump body. As Fig. 11 shows, these
measures resulted in the achievement of a level of total
efficiency of 80% in the pump. As Fig. 12 shows, this
total efficiency level was well beyond that of mass
production pumps.
3.2. Mechanical High-pressure Fuel Pump

Configuration
Figure 10 shows an external view of the fuel pump
produced by this development project. The initial
specifications for the mechanical high-pressure fuel
pump featured three units: a reduction gearbox to
decelerate the pump against engine input, a gear pump
to maintain the feed pressure for the plunger pump, and
a plunger pump for high-pressure fuel supply. Later, in
order to reduce weight, the reduction gearbox and the
gear pump were integrated to produce a reduction gear
pump. The rotational input from the engine was
decelerated by the drive gears, while the gear pump
supplied fuel. This helped to reduce the total length of
Gear pump
Reduction
gear box
Reduction
gear pump
Plunger
pump
Plunger
pump
Final spec
Initial spec
Fig. 10
High-pressure fuel pump
100
90
High-pressure
collector tank
Total efficiency (%)
Bladder
Variable fuel pressure

regulator
Gas bag
PRV
Filter
Low-pressure
collector tank
To engine
Primary
electric fuel pump
Secondary
electric fuel pump
70
60
50
40
30
Drive shaft
20
Mechanical fuel pump
Fig. 9
80
Fuel system for Formula One engine
112
Lowpressure
drop
Optimized
body
stiffness
Small
plunger
bore
Large
plunger
bore
Fig. 11
Fuel pump characteristics
Modified
gear
thickness
difference of the diaphragm, as generated using spring

power alone. When the regulator was switched to the
high-pressure side, the solenoid valve was closed and
fuel was supplied to the back pressure chamber.
80
70
60
50
40
30
Mass
production A
Fig. 12
Mass
production B
Mass
production
2005
Honda
2006
Fuel pump characteristic comparison
4. Development of Variable Fuel Pressure

Regulator
The aim of fuel supply system development for the
2006 season was to increase power by realizing a shorter
fuel injection period through an increased injector supply
flow rate. This presented the technological issue of the
challenge of controlling the injector supply flow rate at
high fuel pressures in the low load and idling ranges,
ranges in which ultra-low supply flow rates are used.
This led to the development of a variable pressure
regulator using diaphragm back pressure control, which
was able to reduce fuel pressure in the low load and
idling ranges, and to increase fuel pressure in the high
load range in order to maximized engine power, as a
means of controlling the supply flow rate of the highfuel-pressure, high-supply-flow-rate injectors.
4.2. Variable Pressure Regulator Performance

The main performance demand on the variable
pressure regulator was responsiveness when switching
fuel pressure between high and low pressures. If fuel
pressure switching performance was unsatisfactory, the
actual injector supply flow rate would become unstable,
the engine combustion state would vary and air-fuel
control would become challenging, resulting in unstable
engine output. To avoid these issues, the responsiveness
for switching from the low-pressure to the high-pressure
side was set to within 500 ms, and the responsiveness
of switching from the high-pressure to the low-pressure
side was set to within 20 ms.
Switching responsiveness from the low-pressure to
the high-pressure side was set via the size of the
orifices in the fuel channels of the variable pressure
regulator. Responsiveness from the high-pressure to the
low-pressure side was set via the supply flow rate of
the electronically-controlled solenoid valve used for
pressure switching. The desired responsiveness was
achieved through optimization of these specifications,
helping to enable the achievement of a stable air-fuel
ratio (Fig. 14).
4.1. Variable Pressure Regulator Configuration

Figure 13 shows the configuration and the operating
principle of the variable pressure regulator. The variable
pressure regulator was based on a pressure regulator
using a diaphragm for return. The unit was provided with
a back pressure spring chamber (back pressure
chamber below) that applied back pressure to the
regulator diaphragm. Forced application of fuel pressure
to the back pressure chamber helped to enable the
pressure difference of the diaphragm to be maintained
at a constant level while the pressure adjustment function
responded to higher fuel pressure. Fuel was supplied to
the back pressure chamber via orifices in the fuel supply
channels, and a fixed supply flow rate was maintained.
A small needle valve regulator was used to hold the back
pressure chamber fuel pressure constant. Because the
small needle valve regulator controlled only an ultra-low
supply flow rate, an orifice was positioned in front of it
to help ensure that it was able to adjust pressure.
Pressure control in the variable pressure regulator
operated as follows: When the regulator was switched
to the low-pressure side, a solenoid valve was opened
and the pressure was released from the back pressure
chamber; pressure was adjusted using the pressure
113
Solenoid
valve
Collector
tank
Engine
Needle
valve
regulator
Mechanical
pump
Diaphragm
regulator
FILTER
90
Orifice
Orifice
Spring
Back pressure chamber
Low-pressure mode
Solenoid
valve
Collector
tank
Engine
Needle
Valve
regulator
Mechanical
pump
Diaphragm
regulator
FILTER
100
Total efficiency (%)
Orifice
Orifice
Spring
Back pressure chamber
High-pressure mode
Fig. 13
Variable fuel pressure regulator system

Table 1
40
Phase 1
Phase 2
Phase 3
Phase 4
00 - 03
04
05
06 - 07
08
Inj layout
Top
Top + NI
Top + NI
Top
Top
Fuel pressure (MPa)
1.2
1.2
Static flow (L/h)
43
43 (NI:33) 55 (NI:55)
Pintle
35
Year
Response (ms)
30
25
20
Hole type
Number of holes
15
Injector type
Phase 5
10
10
78, 100
100
Pintle (NI:Multi)
Multi
Multi
(NI:6)
6, 10, 24
24
10
two injectors per cylinder. Because the NI were

positioned close to the valves, the use of multi-hole
nozzles helped to enable the realization of sprays from
two directions directly aimed at each port. Because it
was necessary to accurately control an ultra-small flow,
the supply flow rate for the NI was set lower than that
of the top injectors.
Response from 10 MPa to 2 MPa

Target
5
0
0
50
100
150
200
Flow rate in solenoid (L/h)
Fig. 14
Fuel pressure response characteristics
5. Evaluation of Durability of High-pressure

Fuel System
The durability of the high-pressure fuel pump and the
variable pressure regulator under actual operating
conditions was verified. Unit tests were conducted to test
durability at the pump speed frequency during mode
operation. Dedicated pump, regulator, and injector unit
test equipment that was able to reproduce circuit driving
modes using an identical engine control computer to the
computers employed in actual vehicles was used to
conduct the evaluation.
The important points of focus in terms of fuel pump
durability included body stiffness, lubrication
performance, and the toughness of each constituent part;
for the regulator they included body stiffness and the
toughness of the plastic internal parts against swelling.
Focusing on these factors, sufficient durability was
achieved to realize a maintenance interval making it
possible to use the pump for multiple races.
6. Injector Development
6.1.3. Phase 3 (2005)

In this phase, a system using a fuel pressure of 5
MPa was introduced in order to boost engine
performance by increasing the atomization of the spray
through the use of high-pressure fuel injection. The
Honda-made NI was modified for high fuel pressure use,
but commercial units were employed for the top
injectors, pump, and regulator.
6.1.4. Phase 4 (2006-2007)
The Phase 3 NI was modified for use with a fuel
pressure of 10 MPa, and was employed in races. From
2006, regulations set an upper limit of 10 MPa for fuel
pressure, and also stipulated that only one injector could
be used per cylinder. Given this, development was
conducted to achieve the effect of the NI using only a
top injector, and the effect of the spray on engine
response was added as spray selection criteria in the
injector spray development process.
To boost engine performance, a development program
was conducted during the season to help enable the use
of a shorter injection period by increasing the injector
supply flow rate and generating further atomization of
the spray by increasing the number of injector holes. The
6.1. Injector Specifications

As Table 1 shows, the injectors employed by Honda
in races until 2008 can be grouped into five phases. This
section will follow these groupings in discussing injector
development.
Top
6.1.1. Phase 1 (2000-2003)

Figure 15 shows the injector layout. The injectors
were positioned at the top of the intake funnels (these
will be termed top injectors below). Fuel was injected
at a pressure of 1.2 MPa, and pintle-type single-hole
nozzles produced a conical spray.
6.1.2. Phase 2 (2004)
To increase intake efficiency and responsiveness in
the transient range in this phase, injectors were
positioned close to the intake valves (near valve
injectors (NI) below), to provide, with the top injectors,
114
NI
Fig. 15
Injector layout 2

Table 2
final spray configuration in Phase 4 was a conical spray

using 24 holes. In addition, because the fuel pressure
regulations had increased the importance of the spray
characteristics, research on new methods of evaluation
of these characteristics and study of next-generation
injector specifications was commenced.
6.1.5. Phase 5 (2008)
Using spray measurement and evaluation methods
developed in the previous year, the effect of differences
in spray on actual engine performance was analyzed, and
a concept of the optimum spray was formulated.
This concept was reflected in a new injector design,
and the injector was employed in races from the opening
race of the 2008 season. Regulations had prohibited any
changes to injector specifications that would boost output
performance, but modifications were made as needed
during the season to increase reliability.
The next section will discuss the Phase 5
specifications, the final injector specifications used by
Honda in Formula One.
6.2. Injector Development Concept
6.2.1. Reduction of size and weight
A reduction in the size of the injectors would reduce
their weight, increase the degree of freedom of layout,
and also help to enable the delivery pipes and other parts
to be reduced in weight. The achievement of weight
savings in the injectors positioned at the top of the
engine and in fuel system-related parts would contribute
to lowering the center of gravity of the engine.
6.2.2. Modification of supply flow rate characteristic
In the high engine speed environment (maximum:
19000 rpm) of a Formula One engine, it is necessary to
promote the dispersion of the fuel spray and achieve an
intake air cooling effect by injecting a large quantity of
fuel in a short time period.
In addition, stable injection is also necessary when
ultra-low quantities of fuel are injected, when the engine
is idling or off-throttle. The achievement of increased
responsiveness in the injector needle valves was focused
on in order to achieve a fuel spray characteristic that
satisfied these performance demands.

Injector specifications 1
Year
2007
2008
Length (mm)
57
48
Diameter (mm)
20
16
Weight (g)
65
38
Number of coil windings
268
100
Coil resistance ()
2.8
1.0
Peak: 2.4
Hold: 0.6
6.0
Operating current (A)
pressure below) was focused on (Fig. 16). The

reduction of fuel pressure loss in the injectors increased
the plate feed pressure, helping to enable the realization
of a spray characteristic equivalent to that of an actual
increase in fuel pressure.
6.3. Developed Injector Technologies and Performance
6.3.1. Reduction of size and weight
Figure 17 shows an external view of an injector. A
new magnetic material developed for use in the injector
(cobalt steel) was employed to reduce the size and
weight of the units. In addition, the structure of the
needle valves was modified, helping to enable the total
length of the valves to be reduced. Because the size of
the injectors was determined by the number of coil
windings, the number of windings was also reduced.
However, reducing the number of coil windings
would result in a decline in magnetic attraction. The
magnetic circuits were therefore optimized and drive
current increased in order to increase attraction.
Table 2 shows the size of the injectors and the coil
specifications.
6.2.3. Modification of spray characteristic

A balance between the form angle, droplet diameter,
dispersion, and penetration is necessary for the spray
characteristic. However, increased atomization is in a
trade-off relationship with dispersion and penetration,
and an increase in atomization will cause a decline in
dispersion and penetration, and vice versa. The previous
injectors had reached the limit values for each parameter,
making the realization of enhanced performance an
urgent issue.
Against this background, the regulations had
established an upper limit for fuel pressure. In order to
efficiently utilize the energy of the fuel pressure in the
spray characteristic, the actual fuel pressure on the holes
(the hole plates) in the injectors (termed plate feed
115
Fuel pressure
Needle valve
Valve seat
Seat pressure loss
Plate feed pressure
(Upstream pressure of hole)
Hole plate
Fig. 16
Injector internal fuel flow (2007 model)
2007
INJ
Fig. 17
2008
INJ
Injector body shape
6.3.2. Supply flow rate characteristic

Without variable fuel pressure control, the 2007
injectors were not able to conduct stable fuel injection
from idling to wide-open throttle. A development project
was therefore initiated to produce a new injector that
would be able to inject fuel across the operating range
under high pressure. In order to realize this goal, the
weight of the needle valve (a movable part) was reduced;
the magnetic characteristic was enhanced by modifying
the magnetic materials and magnetic paths; the rise in
magnetomotive force was increased by the increase in
the rise in current achieved by reducing the number of
coil windings (reducing the resistance value); and the
responsiveness of the needle valve rise was enhanced by
means of the increase in the magnetic attraction of the
valves achieved by increasing the drive current. The
spring set load could also be increased by the amount
of the increase in magnetic attraction, helping to enable
an increase in needle valve fall responsiveness.
In addition, the drive current waveform was
reexamined and the linearity of the fuel supply flow rate
against current was enhanced. As a result, when the
injector spray characteristic was equivalent to the 2007
model, the injector was able to inject fuel at high
pressure across the entire operating range.
However, the needle valve stroke and the diameter
of the valve seats were increased in order to boost the
plate feed pressure, with the aim of enhancing the spray
to increase power. This made control of the ultra-low
supply flow rates at low loads and in the idling range a
challenge, and as a result, the use of variable pressure
control was continued.
This was the result of a comparative examination of
the effect on engine performance of enhancing the spray
characteristic and the effect on actual vehicle
performance (weight savings) of discarding the variable
pressure control system. It was determined that the
24 holes
(concentric circles)
Fig. 19
Year
Static flow (L/h)
Injector internal structure (2008 model)
Atomization achieved by 24-hole injector
2007
2008
100
100
Angle of spray (deg)
53
53
Seat diameter (mm)
1.30
1.48
0.10
0.15
0.4
0.2
Valve stroke (mm)

Plate thickness (mm)
Number of holes
Hole diameter (mm)
24
24
0.100
0.090
enhancement of the spray characteristic was a more

important consideration.
6.3.3. Spray characteristic
The needle valve stroke and the diameter of the valve
seats were increased in order to reduce the pressure loss
resulting from the squeezing of the valves, resulting in
increased plate feed pressure. Table 3 shows the points
that were modified, and Fig. 18 shows the configuration
of the injectors.
In addition, the fuel flow was modified by optimizing
the shape of the internal fuel channels and employing
spherical valve seats. These modifications increased plate
feed pressure from 4.7 to 8.9 MPa at a fuel pressure
setting of 10 MPa.
The enhancement of spray performance (droplet
diameter, penetration, diffusion) helped to enable the
achievement of a good balance between these
parameters, which are in a trade-off relationship. The 24hole conical spray configuration that was judged to be
optimal in terms of vehicle performance based on the
analysis in Section 2.1. was adopted (Fig. 19).
With regard to atomization, an average droplet
diameter (SMD) of approximately 17 m was achieved.
7. Conclusion
Spherical shape seat
Fuel flow passage in needle valve
Fig. 18
Table 3 Injector specifications 2
(1) Development of high-pressure pump and variable

pressure regulator
A pump efficiency of 80% was achieved through a variety
of measures, including reduction of pressure loss in the
pump inlet port and optimization of the size of the plunger
pistons.
In addition, the employment of a variable pressure
regulator helped to enable the realization of high-supplyflow-rate injection, resulting in increased engine output
and stable injection of ultra-low fuel supply flow rates
at low loads.
(2) Development of high-pressure injectors
The use of new magnetic materials and the modification
of the injector configuration and the drive current helped
to enable the injectors to be reduced in size.
With regulations stipulating an upper limit on fuel
pressure, the development aimed to increase combustion
efficiency by means of atomization of the spray;
reduction of internal pressure loss in the injectors helped
to enable the achievement of a spray characteristic that
116
produced a performance enhancement equivalent to that

of an actual increase in fuel pressure.
(3) Engine performance
Increasing fuel pressure helped to enable an increase in
the maximum injector supply flow rate; the resulting
shorter injection period promoted dispersion of the fuel
spray, and an intake air cooling effect was obtained. In
addition, increased atomization promoted greater mixing
of the intake air and the fuel, enhancing combustion
efficiency. As a result of these measures, fuel system
development produced an increase in power of
approximately 15 kW against previous levels at the
commencement of Hondas third Formula One era.
Author
Tetsuya TANAHASHI
Kazuji ONO
Yosuke SAWADA
Atsushi SHIMIZU
117
Masanori HAYAFUNE
Descriptions of Gearbox Technologies
Development of Seamless Shift

for Formula One Car
Takashi YOSHIOKA*
Katsumi KUBO*
Takeshi UCHIYAMA*
Ryo MATSUI*
ABSTRACT
Honda focused on gearbox development during its third Formula One era. The reduction of shift time is an effective
means of maximally increasing race competitiveness within the constant-mesh 7-speed gearbox regulations. In the
standard shift process, the current gear was disengaged, the system went into a neutral state, and the following gear
was engaged. Hondas seamless shift realized up-shift with a torque loss time of zero, by engaging the following
gear and then disengaging the current gear. Normally, this process would lead to damage due to double engagement,
but in the developed system double engagement was prevented and transmission of deceleration torque was enabled
by adding one-way clutches with a locking function to the conventional shift mechanism. The selective use of these
one-way clutches, positioned between the gear hubs and the mainshaft, and the use of cooperative control with the
engine, enabled the realization of seamless shift across the entire shift range. As a result, lap time was reduced by
0.4 sec per lap, and the system was used in races from 2005 as the first shift mechanism of its type in the Formula
One world.
This paper will also discuss the removal of the shift forks and shift rings as well as the fitting of the gear selection
mechanism inside the mainshaft in order to reduce the total length and weight of the mechanism while maintaining
its seamless shift performance.
1. Introduction
The shift mechanisms used in Formula One vehicles
are constant-mesh parallel twin shaft types, in which gear
stages are changed using dog clutches of the type
frequently employed in motorcycles. In the project
discussed in this paper, a shift mechanism was modified
at the initial stage of gearbox development in order to
boost the competitiveness of Formula One vehicles. The
development aim established to realize this goal was to
reconcile rapid shift (reduced shift time) with secure shift
(durability and reliability). In order to realize rapid shift,
acute chamfer angles were employed in the dog clutches,
barrel inertia was reduced, and the speed of operation
of the shift system was increased. The increase in shift
speed generated a number of issues, including
operational irregularities originating in shift fork
overshoot and inclination of the shift rings. In order to
prevent these issues, the forms of the barrel cams and
the shift forks were optimized to stabilize shift operation.
This helped to realize an equivalent level of
competitiveness with the vehicles of other teams, but
innovative technologies are essential to achieving victory
in the fast-evolving world of Formula One racing. In
conventional shift mechanisms, it was necessary to

reduce engine torque to close to zero during up-shift, and
the time of deceleration produced by the air resistance
of the vehicle had a significant effect on performance.
However, further reductions in shift time could not be
expected using this type of mechanism. The ultimate
goal for attempts to reduce shift time is the realization
of shift with no torque loss. It was assumed that
achieving this within the regulations would involve a
variety of issues, but it was also considered to represent
an excellent opportunity to increase Hondas
competitiveness against other teams, and a development
project was therefore commenced.
2. Development Aims
Shifts using one-way clutches, as are often employed
in automatic transmissions in mass-production vehicles
for shift from 1st to 2nd gears, enable at least one shift
with no torque loss. The development aim established for
the project was to realize seamless shifts by developing
a shift mechanism in which this function could be
applied to all the gear stages in a Formula One vehicle
gearbox.

120
3. Methods of Achieving Development Aim

3.1. Alteration of Dog Clutch Engagement Timing
In a conventional shift mechanism, when torque is
acting on the dog clutches, friction on the dog clutch
coupling sections prevents shift (engagement and
disengagement of the dog clutches). Because of this,
during shift the engine torque was reduced, and the dog
clutch of the current gear stage was disengaged. When
this operation was completed, the dog clutch of the next
gear stage was then engaged simultaneously with the
recovery of engine torque. However, the engagement of
the dog clutch of the next gear stage while the current
gear stage was still transmitting torque, and the
disengagement of the dog clutch of the current gear stage
after the next gear stage commences transmitting torque
(when the current gear stage has ceased transmitting
torque) would enable shift with no drop in engine torque.
Figure 1 shows the difference in dog clutch engagement
timings for shifts from 5th to 6th gears in the conventional
and seamless shift mechanisms as an example. As the
figure shows, in conventional shift there is a neutral state
between the stages, while in the seamless shift, the dog
clutches for both 5th and 6th gears are engaged.
3.2. Prevention of Double Engagement during Up-shift
As Fig. 1 shows, if this shift timing is applied using
the conventional shift mechanism, there will be a period
during which the current gear stage and the next gear
stage will be engaged simultaneously, i.e., double
engagement. It is widely known that double engagement
may cause issues that make the vehicle inoperable. A
mechanism enabling double engagement to be prevented
was therefore developed for the new system. Figure 2
shows the configuration of the parts of the seamless shift
mechanism. Figures 3 and 4 show the movement and
function of each of the parts during up-shift.
Figure 3 shows the parts during acceleration, as seen
from the front. The direction of rotation is
counterclockwise. Torque from the engine is transmitted
to the clutch, the lay shaft gear, the mainshaft gear, the
shift ring, the gear hub, the strut, and the mainshaft, and
is output to the tires. In a conventional shift, the gear
hubs and the mainshaft are connected by splines, but the
seamless shift mechanism employs struts that function
as one-way clutches between the gear hubs and the
mainshaft.
Figure 4 shows a state of double engagement, when
the next gear stage has commenced transmitting torque
and the dog clutch of the current gear stage is still
engaged, during up-shift using the seamless shift
mechanism.
When the dog clutch of the next gear engages and
the gear commences transmitting torque, the rotation of
the next gear stage becomes faster than that of the
current gear stage, and the strut of the current gear stage
is taken into the mainshaft pocket inside the gear hub
where it functions as the idling side of a one-way clutch,
enabling double engagement to be prevented. Because
Mainshaft gear
Shift ring
Gear hub
Strut
Mainshaft
Release bearing
Ball
Push rod
Fig. 2
Parts configuration of seamless shift mechanism

Mainshaft gear
Shift ring
Gear hub
Conventional shift
Drive by 5th
Strut
Neutral state
Drive by 6th
Mainshaft
Fig. 3
5th gear
Up-shift state
6th shift ring
5th shift ring
6th gear
Gear hub
Seamless shift
Drive by 5th
Double engagement
Strut
Drive by 6th
Mainshaft
Ball
Release
bearing
Fig. 1
Dog engagement timing
Fig. 4
121
Ratcheting state
3.4. Downshift
The new system features the following major
differences from a conventional shift mechanism to
enable the realization of seamless shift:
(1) A shift barrel profile enabling a double engagement
timing to be obtained
(2) A mechanism enabling double engagement to be
prevented
To enable downshift using mechanisms that focus on
up-shift, when the release bearing has been released and
the system has downshifted from the current gear to the
next gear, the release bearing is locked again. In order
to release the release bearing, it is necessary to reduce
torque. This method increases the length of time that
deceleration torque is reduced against a conventional
shift mechanism. A shift barrel profile enabling
downshift without releasing the release bearing was
therefore proposed in order to achieve downshift in the
same short period as a conventional shift mechanism
while realizing seamless up-shift. Figure 6 shows the
shift barrel profile during up-shift. The shift rings and
the dog clutches engage when the shift fork stroke
reaches 39% or more, and the diagonally-shaded area in
the figure therefore represents a state of double
engagement.
4. Effects
The time-chart of engine speed, engine torque, and
wheel speed during up-shift were compared in order to
verify the effects of the seamless shift mechanism
developed in this research. Figures 8 and 9 show data
for conventional shift and seamless shift respectively.
The seamless shift realizes up-shift while maintaining
engine torque, enabling shift to be completed with no
loss of drive power. In a bench test simulating the
Silverstone Circuit, figures of 4 km/h when converted for
speed and 7.6 m when converted for distance at the end
of the home straight were obtained. The new mechanism
was employed in races from the 1st race of 2005.
Shift barrel profile

100
Shift fork stroke [%]
3.3. Engine Brake

This simple one-way clutch mechanism using the
struts enables torque transmission during acceleration,
but it is unable to transmit torque during braking. In
order to resolve this issue, a ball is inserted beneath the
strut in order to lock it, enabling deceleration torque to
be transmitted. Figure 5 shows the system during
application of the engine brake. The path of deceleration
torque is opposite to the torque path during acceleration.
The input from the tires is transmitted to the driveshaft,
the differential gear, the bevel gear, the mainshaft, the
ball, the strut, the gear hub, the shift ring, and the
mainshaft gear, and is finally input to the engine. The
release bearing is locked or released by a pushrod.
Figure 7 shows the shift barrel profile during

downshift. The diagonally-shaded area in the figure
shows the system in a neutral state. This neutral state
enables downshift with the release bearings locked.
This shift barrel profile also features free areas, in
which the shift forks move freely. These are shown as
the vertically-shaded areas in the figures. Unwanted
movement of the shift forks has been prevented by
positioning dtente springs between the selector rails and
the shift forks. The engagement and disengagement of
the dog clutches on the up-shift and downshift sides at
different timings has enabled the reconciliation of the
desired up-shift performance with downshift in the same
time period as a conventional shift.
the current gear stage is no longer transmitting torque,

the shift ring can be released from the dog clutch without
causing engine torque to drop, enabling the realization
of seamless shift.
The strut of the next gear stage engages automatically
with the gear hub due to centrifugal force, and functions
as the engaging side of a one-way clutch.
Previous gear
80
100
Up-Shift
60
40
Next gear
20
100
0
0
20
40
60
80
Shift barrel angle [%]
Fig. 6
0
100
100
Up-shift shift barrel profile
Shift barrel profile
Strut
Mainshaft
Ball
80
60
Down-shift
40
Previous gear
20
Overrun state
20
40
60
80
Fig. 7
122
0
100
0
100
0
Release
bearing
Fig. 5
Next gear
Gear hub
100
100
100
Down-shift shift barrel profile
5. Evolution
gearbox to be reduced. At the same time, the weight of

the gearbox itself was reduced (Figs. 10 and 11).
5.1. Further Development Aims

Honda led the world in introducing the seamless shift
mechanism, but by the end of the following season,
almost all other teams were adopting systems that
offered the same benefits, and Hondas advantage from
the perspective of shift performance had weakened. The
system had maximally increased performance in terms
of preventing torque loss during acceleration, but
Hondas gearbox fell behind those used by the other
teams from the perspectives of weight and compactness.
A development program was therefore conducted with
the aims of making the system more lightweight and
compact while maintaining the same level of shift
performance.
5.2. Mechanism
5.2.1. Salient features
An in-shaft shift mechanism was developed that
positioned one-way clutches able to control torque
transmission and idling between all the shift gears and
the mainshaft, and that was fitted within the mainshaft.
Doing away with the shift forks, shift rings, gear hubs,
and other equipment, which were conventionally
positioned between the shift gears, and using only shift
gears arrayed in a line enabled the total length of the
Engine speed
1000 rpm
Gear position
5th
5.2.2. Configuration
The in-shaft shift mechanism features an operating
range in which double engagement is used, and the
principle of preventing torque loss is the same as that
used by the seamless shift mechanism discussed above.
However, the configurations of the shift mechanisms
differ significantly. This section will discuss the parts
forming the mechanism and the roles of those parts.
Figure 12 shows the shift mechanism, including shift
gears. Each strut positioned between a shift gear and the
mainshaft moves in a seesaw fashion by means of small
and large balls, producing the following three essential
states for shift:
(1) An in-gear state, in which a strut simultaneously engages
and is locked in place, enabling acceleration and
deceleration torque to be transmitted (Fig. 13)
(2) A neutral state, in which the angle between the struts
and the mainshaft in the circumferential direction is
smaller than it is in the in-gear state and the gears and
6th
100 Nm
Input shaft torque
5 km/h
Vehicle speed
50 ms
Fig. 10
Fig. 8
Conventional seamless shift
Data for conventional shift
1000 rpm
6th
Engine speed
Gear position
5th
100 Nm
Input shaft torque
5 km/h
Vehicle speed
50 ms
Fig. 9
Data for seamless shift
Fig. 11
123
In-shaft seamless shift
mainshaft idle during both acceleration and deceleration

(Fig. 14)
(3) A one-way state, in which only acceleration torque is
transmitted (Fig. 15)
The state of the struts is controlled by the positions
of two large and small balls in the radial direction. The
positions of the balls are adjusted by the movement in
the axial direction of a slide cam that is provided with
a cam groove in the axial direction designed specifically
for use with the balls. The movement of the shift
bearings in the axial direction adjusts the position of the
slide cam in the axial direction by means of a spring.
The position of the shift bearings in the axial direction
is adjusted by the movement in the axial direction of a
pin integrated with the shift bearings, which follows a
barrel cam groove formed on the inside of the shift
bearings in the circumferential direction. As a result, it
is possible to selectively control the three states of the
struts by means of barrel rotation.
Figure 16 shows the timing of in-gear states for upshift and downshift. a shows the disengagement timing
of the current gear during downshift, b shows the
engagement timing of the next gear during up-shift and
downshift, and c shows the disengagement timing of
the current gear during up-shift. Different timings are
necessary for the disengagement of the current gear
during up-shift and downshift. This is necessary for the
same reason as was the case for the seamless shift
discussed above, but the method of achieving it differs
5.2.3. Operation
The process of shifting from the current gear to the
next gear is as follows:
(1) When the current gear is driving, the current gear is in
an in-gear state, and all the other gears are in neutral
states.
(2) Shift is commenced by the rotation of the barrel; the struts
of the current gear and the next gear are put in one-way
states simultaneously.
(3) When torque transmission shifts from the current gear
to the next gear, the strut of the current gear, which has
ceased transmitting torque, is put in a neutral state.
(4) The next gear is put in an in-gear state, and shift is
completed.
Fig. 13
In-gear state
Fig. 14
Neutral state
Idler bearing
Strut
Gear
Ball (big)
Fig. 15
Slide cam
Spring
One-way state
100
Barrel
Slide cam stroke [%]
Engage
(Up & Down)
Ball (small)
Engage area
Disengage
(Down)
Disengage
(Up)
Shift bearing
Mainshaft
Guide tube
Oil pipe
Fig. 12
Disengage area
0
10
20
ab
30
40
50
60
Barrel angle [%]
70
c
80
Shift
Parts configuration of in-shaft seamless shift

mechanism
Fig. 16
124
Shift timings
90
100
for reasons of space. In the new system, the different

timings are enabled by positioning springs between the
slide cams and the shift bearings. During up-shift, the
strut of the current gear cannot be disengaged until the
current gear has ceased transmitting torque. By means
of compressing the spring even when the barrel is
rotating, the movement of the slide cam is delayed, and
as a result the transition to a neutral state is delayed, thus
creating a range in which double engagement occurs.
During downshift, because a load is applied in the
direction of disengagement of the strut, the transition to
a neutral state can be made with no delay in the rotation
of the barrel against the movement of the slide cam, thus
enabling double engagement to be prevented.
5.2.4. Effects
The employment of the newly developed in-shaft
seamless shift mechanism has reduced the total length
of the mechanism by 19% (from 192.7 mm to 155.1
mm) and the weight of the mechanism by 12% (from
10.4 kg to 9.1 kg) against a conventional shift
mechanism.
5.3. Verification of Performance and Reliability
Shift performance figures for the in-shaft seamless
shift mechanism were identical to those for the seamless
shift mechanism, shown in Fig. 9. The new system was
being used under the four-race gearbox regulation
introduced by the FIA in 2008, making it important to
verify reliability. Bench tests were commenced on the
system, but Honda then announced its withdrawal from
Formula One racing, and the development project was
discontinued when the tests had reached the 1300 km
mark of a projected 2500 km.
6. Conclusion
A shift mechanism that has a modified shift sequence
to prevent torque loss during up-shift and selectively
uses one-way clutch mechanisms to prevent double
engagement was realized within the scope of Formula
One regulations. This development produced the
following outcomes:
(1) As the first shift mechanism of its kind in the Formula
One world, the new system enabled lap times to be
reduced by 0.4 sec per lap, and was used in races from
2005.
(2) The realization of up-shift with no torque loss and
minimal torque fluctuation during shifting enabled
shifting in situations of low tire grip, such as when
cornering or during rainy conditions.
The quest for increased compactness while
maintaining seamless shift performance by removing the
shift forks and dog rings and fitting the mechanism
inside the shaft enabled the total length to be reduced
by 19% and weight to be reduced by 12%.
Author
Takashi YOSHIOKA
Ryo MATSUI
125
Takeshi UCHIYAMA
Katsumi KUBO
Development of Honda Gears for Formula One

Gearbox
Koichi KONISHI*
Teruoki NAGANUMA*
Hiroaki NOJIMA*
Hiroshi KIMURA*
ABSTRACT
Development programs were conducted to enable the development of Honda-made gears that balanced low weight
and compactness with reliability. Development of a technology for the formulation of gear specifications using FEM
analysis and a method of predicting the lifespan of the gears based on S-N curves, which formed the basis for FEM
analysis, helped enable continuous short-term development efforts. As gears would sometimes fail in actual vehicle
tests, a method of evaluation of lightweight gears optimized for high-load, short-lifespan race use was established
with the plastic deformation of the gears as an index. The lightweight, compact, and high-reliability gears developed
in this process were employed in races from 2003, and were entirely trouble-free.
In addition, a diamond-like carbon (DLC) coating optimized for race use was applied to all the gear tooth surfaces
to boost the performance of the powertrain, helping to achieve a transmission efficiency of 97%. The coating was
applied to different gear sets in sequence from 2007.
1. Introduction
2. Layout of Formula One Gearbox Internals
At the beginning of Hondas third Formula One era,

gears manufactured by a specialist race-gearbox maker
were used. However, the top teams were using lighter
and more compact in-house gears to increase their
competitiveness.
Formula One gearboxes are positioned close to the
rear-ends of the vehicles, and their weight makes a
significant contribution to dynamic performance. A wide
variety of ratio gears is necessary to help enable drive
force to be set for each race circuit. In addition, gearbox
issues will directly result in retirement from the race.
These factors necessitated methods of determining
specifications that would reduce the weight of the gears
as well as helping to enable accurate prediction of gear
life. Regulations freezing engine development also
increased the importance of reducing loss in the
drivetrain.
In response to these demands, Formula One gear
design methods using FEM analysis and standards for
quantitative evaluation were formulated, and methods of
dealing with exceptional overloads were developed,
helping to enable the development of gears that
reconciled low weight with reliability. In addition, the
transmission efficiency of the gears was identified as the
major factor in drivetrain loss, and a DLC coating
adapted to race conditions was developed for the gears.
Given considerations of vehicle dimensions,

aerodynamic packaging, and the stipulations of
regulations, a longitudinal two-parallel-axis configuration
in which the intermediate shaft (cross-shaft) was
positioned one stage upstream from the final reduction
gears, was employed for the gearbox internals. Figure
1 shows this gearbox configuration.

126
Final gear and

differential
Ratio gear set

(7-speed)
Bevel gear and

cross shaft
Shift system
Carbon clutch
Fig. 1
3D model of Formula One gearbox internals
3.1. Issues of Applying Mass-Production Gear Design

Methods to Formula One
The conventional gear design process involves the
use of in-house gear specification simulation tools to
study gear specifications and evaluate the stresses
generated on gear-tooth bottoms. For mass-production
gears, it is normal for a specific thickness of the geartooth bottoms to be determined by the gear module, and
this is used as a precondition for the above-mentioned
simulation tools. Formula One gears must balance
resistance to extremely high loads and weight saving,
even in short lives, in comparison to the gears used in
mass-production vehicles, and the concepts used in
determining specifications are consequently different.
Tooth root strength and pitting strength are necessary in
Formula One gears, and the gear modules are therefore
increased in size. The thickness of the tooth roots is
reduced in order to save weight. The resulting decline
in stiffness must be taken into consideration, so
conventional design methods were not sufficient to the
task. FEM analyses that considered contact between gear
teeth were therefore conducted.
Figure 2 shows the difference in shape between a
mass-production gear and a Formula One gear.
conditions the condition in which the stress on the tooth

roots was greatest, and the reflection of this result in the
FEM analysis, helped to reduce in length the period
necessary for analysis.
3.3. Optimization of the Degree of Tip Relief
Tip relief refers to optimizing the form of the geartooth tip in relation to an involute curve with
consideration of elastic deformation when a load is
placed on the tooth. In Formula One gears, the reduction
in stiffness mentioned above results in a high level of
elastic deformation, which changes tooth contact and has
a significant effect on tooth strength and contact face
pressure. The level of elastic deformation of the gear
teeth tips was therefore predicted using FEM analysis.
The degree of tip relief was optimized based on the
results.
Driven
gear
(a) 1 tooth meshing
Fig. 3
3.2. FEM Analysis Conditions

Analysis conditions are important in FEM analyses,
and in the FEM analyses of spur gears in particular, it
is necessary to reproduce the worst-case conditions for
a variety of meshing states. The contact gear ratio for
the spur gears alternates between 1 and 2, and the gear
meshing condition is therefore a continuous sequence of
one tooth meshing two teeth meshing one tooth
meshing. FEM analysis showed that of these, one tooth
meshing, and in particular the meshing condition in
which the meshing point was highest in the direction of
the depth of the gear tooth, placed the greatest stress on
the gear tooth root.
Figure 3 shows differences in gear meshing
conditions and stress distribution, and Fig. 4 shows
changes in the stress on the tooth roots when the
meshing phase was altered.
In one tooth meshing, stress concentrates at a single
meshing point (Fig. 3), while in two teeth meshing there
are two contact points, and the stress concentration is
therefore relieved (Fig. 4).
The use of the gear specification simulation to
determine from among a variety of tooth meshing
Mass production
gear shape
Fig. 2
Formula One
gear shape
Gear shape comparison
Rotation
Drive
gear
(b) 2 teeth meshing
Gear meshing condition

2500
1 tooth
meshing
2 teeth
meshing
1 tooth
meshing
2 teeth
meshing
2000
1500
Stress [MPa]
3. Formula One Gearbox Design Method

using FEM Analysis
1000
500
Tooth meshing angle
Fig. 4
Change in stress with gear meshing condition
4. Quantitative Evaluation Method for

Formula One Gears
4.1. Gear Fatigue Strength Tests
Up to this point, it had been normal to evaluate the
durability of race gears using vehicle tests alone. For the
development of in-house Formula One gears, however,
rig tests of gear units using S-N curves (Wohler curves)
and torque frequency measured in actual vehicles were
employed in gear evaluation. The reliability of high-load,
low-cycle fatigue S-N characteristics represented a
concern, but cross-referencing with the results of vehicle
127
test evaluations conducted previously indicated that unit

rig tests functioned adequately for evaluations. The
determination of gear specifications on this basis helped
to enable weight savings to be reconciled with high
reliability (Fig. 5). The value of stress at the tooth roots
was calculated from input torque, and the S-N curve was
formulated based on test results.
4.2. Early Breakage of Gears
The technology for prediction of gear lifespan using
rig tests reduced gear breakages, but breakages
sporadically occurred earlier than the end of the
predicted lifespan of the gears during vehicle tests.
Figure 6 shows a representative example of breakage.
When gears that had been used in races were
measured, plastic deformation was observed in from one
to several teeth (Fig. 7). Given that the plastic
deformation only affected part of the gears, it was
conjectured that it resulted from instantaneous overloads,
and that this resulted in the early breakage of the gears.
The phenomenon was therefore analyzed.
Because the amount of plastic deformation was
minimal, criteria for judgment of tooth profile data were
set, and a difference in the degree of pressure angle
correction of 0.2 m or more was defined as plastic
deformation.
Plastically deformed gears were put through fatigue
strength tests. The results showed in all cases that the
life of the gears was reduced, with the lifespan until

breakage only 11.6% of the projected figure in one case
(Fig. 8).
4.3. Measures against Overload
In Formula One gearboxes, control is applied to
produce an optimum clutch clamp load in order to
protect the drivetrain components from instantaneous
overloads when the vehicle takes off or shifting is
engaged.
Despite this, instantaneous overloads due to inertia
and other forces in the gearbox resulting from variations
in the clutch friction materials, tire traction on uncertain
road surfaces, and irregular steering operation cannot be
entirely prevented. However, if gear design incorporated
allowances for occasional overloads, increases in weight
would be unavoidable, and competitiveness would
decline.
In the Formula One gearbox, torque input during
vehicle operation was monitored by a torque sensor
directly connected to the engine input shaft. The input
torque on plastically deformed gears was studied from
torque sensor log data. The results, when converted to
figures for stress at the gear teeth roots in a simulation,
showed that plastic deformation of the gears occurred when
the torque input exceeded a threshold of 2300 MPa.
Gear durability tests also showed that plastic
deformation occurred at levels of tooth root stress above
3000
2000
Reference line
1500
Tooth
deformation
1000
500
Fig. 7
0
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07
Tooth profile data of deformed gear
Cycle
3000
Fig. 5
Fatigue failure test results

Stress of tooth root [MPa]
Stress at tooth root [MPa]
2500
Fatigue failure of narrow and lightweight gear

S-N curve of narrow and lightweight gear (fatigue failure)
Fatigue failure of previous gear
S-N curve of previous gear (fatigue failure)
2500
S-N curve (fatigue failure)

Life of undeformed gear
Deformation : 4.2 micron
2000
1500
1000
11.6%
500
0
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07
Cycle
Fig. 6
Broken gear
Fig. 8
128
Fatigue failure test result of deformed gear
The following three methods were available to reduce

the coefficient of friction of the gear surfaces:
(1) Reducing the roughness of the gear surfaces
(2) Applying a coating to the gear surfaces
(3) Using a low-friction gear oil
With regard to (1), the gears used in Formula One
engines receive a polishing finish and barrel polishing,
and no further roughness reduction could be expected.
Transmission loss was therefore reduced through (2), the
application of a coating to the gear surfaces, and (3), the
development of a low-friction oil.
2.0
Plastic deformation
Gear plastic deformation [m]
2.5
1.5
1.0
0.5
0.0
0
500
1000
1500
2000
2500
3000
3500
Stress at gear tooth root [MPa]
Fig. 9
Gear strength test results
2300 MPa (Fig. 9).

Based on these results, a gear torque limit
corresponding to a level of gear-tooth root stress of 2300
MPa was set.
In addition, telemetry data from the torque sensor
was monitored in real time, and in the event that a torque
input exceeded the torque limit, the relevant gear was
replaced during gearbox maintenance.
The measures discussed above helped to reconcile the
achievement of low weight with high durability, and the
gears were trouble-free during races.
5. Achievement of Increased Efficiency in

Formula One Gearboxes
5.1. Analysis of Status of Transmission Loss
A situation analysis of factors in transmission loss
was conducted. These factors were isolated by measuring
friction in an unloaded state and efficiency in a loaded
state (i.e., during transmission). The following factors
were identified:
(1) Loss due to oil churning
(2) Loss due to dragging of the oil seals and other parts
(3) Loss due to torque transmission
(4) Oil pump drive loss
Of these, loss due to torque transmission was the
reason that loss increased or decreased with the
magnitude of torque input. Bearing loss as a factor in
transmission loss in a loaded state was calculated using
a theoretical formula to separate it from gear
transmission loss. Figure 10 shows the rate of
contribution of each power loss factor.
These results showed that gear transmission loss was
the major factor in gearbox loss, representing 62% of the
total figure.
5.3. Selection of Coating Film Type

Under actual Formula One use conditions, the
velocity of gear surface slip reaches 20 m/s or more, and
gear surface pressure reaches 2 GPa or more. It was
necessary to employ a coating which would resist
wearing even under these conditions. The gear coating
film was selected from among the following candidates:
(1) Metal DLC, which displays good wear resistance, and is
used on race engine components
(2) An increased-hardness metal DLC
The following two procedures, which are used to help
prevent pitting of the gears in mass-production
transmissions and could be expected to reduce the
coefficient of friction by increasing slidability, were also
included in the comparative study:
(3) Solid-film lubrication (Molybdenum disulfide resin
coating)
(4) Sulfurizing
These four coating types were studied for
effectiveness in reducing loss and levels of durability.
Gears to which only a carburizing treatment had been
applied, which were previously standard, were used as
a benchmark for comparison.
5.4. Verification of Effects of Coatings and Durability
Figure 11 shows the results of verification of loss
reduction in unit rig tests, and Fig. 12 shows the surface
condition of gears following durability tests.
Loss was reduced by 2 kW in the gears with the DLC
coatings. Some damage occurred to the metal DLC
coating, but only minute scuffing of the tips of the gear
teeth was observed in the gears with the hardened DLC
5.2. Methods of Increasing Efficiency

Transmission loss resulted from slipping of gear
surfaces against each other during meshing, and could
be reduced by reducing the coefficient of friction of the
gear surfaces.
129
Churning
Non-loaded
Seals
40%
Bearings
Gears
6%
Pump
50%
4%
Loaded
14% 2% 20%
0%
Fig. 10
20%
62%
40%
60%
2%
80%
Effective rate of power loss factors
100%

150
Power loss (left scale)
Rising time (right scale)
Power loss [kW]
34
140
33
130
32
120
2 kW
31
110
30
100
29
90
28
80
27
Delta 10 degC oil temperature

rising time [sec]
35
70
Carburizing
Fig. 11
Metal DLC
DLC
MoS2 spray
Sulfurizing
5.6. Low-friction Gear Oil

A low-friction oil was developed in cooperation with
an oil manufacturer. At the initial stage of the
development, reductions of loss of 1 kW or more were
obtained in gears without DLC coatings, but when DLC
coatings were applied, no loss reduction effects was
obtained.
The compatibility between the DLC coating and the
oil was therefore examined, helping to enable the
development of an oil that reduced transmission loss by
0.4 kW. This oil was used in races from 2008.
Confirmation of loss reduction effect
6. Conclusion
Gear surface
condition
Coating
Carburizing
Teeth touch
mark only
Fig. 12
Metal DLC
DLC
MoS2 spray
Sulfurizing
DLC peeling Minute scuffing Coating peeling Gear surface

DLC wear
Gear surface
wear & scoring
wear & scoring
Surface condition after durability test
coating, and this did not affect the actual performance

of the gears.
Solid-film lubrication and sulfurizing demonstrated
no transmission loss reduction effect, and wearing of the
coatings was also observed. This is because it is
necessary for the surfaces of the gears to be made
rougher than the bases in order to apply these coating
treatments, and the consequent increase in roughness
results in a decline in transmission efficiency.
5.5. Use of DLC Coating in Races
DLC-coated shift gears and final gears were
employed in races from the middle of the 2007 race
season, and DLC-coated bevel gears were employed
from the 2008 season. Optimization of tip relief and the
coating pre-treatment helped to enable the achievement
of a sufficient level of durability to comply with the
regulation that gears must be used for a full four races
(Fig. 13).
The following results were obtained from the

Formula One gear development program:
(1) A Formula One gear design method using FEM was
developed.
(2) Methods of predicting the lifespans of gears in rig tests
and of quantitatively evaluating exceptional overloads
were developed. These methods formed the basis of a
method implemented in gearbox management during
races.
(3) A gear coating and an oil were developed that helped to
achieve a gearbox transmission efficiency of 97%.
The development results listed above helped to
achieve a good balance between function and reliability,
and the developed gears demonstrated increased
transmission efficiency and produced no gear-related
issues during races.
Acknowledgments
The authors wish to take this opportunity to express
their gratitude to the staff members of Nippon Oil
Corporation, who generously assisted in the development
of a low-friction oil.
Author
Koichi KONISHI
Fig. 13
Final gear condition after 4 races
Hiroshi KIMURA
130
Hiroaki NOJIMA
Teruoki NAGANUMA
Development of Lightweight and Compact

Differential for Formula One Car
Toshio HIYOSHI*
Haruki YOKOYAMA*
Yuichi SUENAGA*
Yoshikazu KATSUMASA*
ABSTRACT
The differentials used in Formula One vehicles comprise a differential mechanism and a differential restriction
device (LSD), which together form a bias-adjusting mechanism, and a final reduction gear set. Because the differential
is positioned close to the rear end of the vehicle, it contributes to dynamic performance. Reducing the weight of the
differential helps to centralize the mass of the vehicle, and increasing its compactness boosts aerodynamic performance.
Development efforts were concentrated on the achievement of a lightweight and compact differential, with the focus
in the initial stage on the centralization of mass. Previously mounted on the final driven shaft, the bias-adjusting
mechanism was positioned on the final drive shaft, helping to reduce the center of gravity and the yaw moment of
inertia. A full pinion engagement planetary gear was employed as the differential gear. A study of the merits and
drawbacks of the mechanism resulted in the development of an ultra-short differential (USD) employing a full pinion
engagement double pinion planetary gear positioned on the final driven shaft as the differential gear. It was predicted
that this would increase the compactness of the unit and reduce its weight by 1.2 kg.
1. Introduction
2. Development Aims
Normally, the bias-adjusting mechanisms of

Formula One differentials are positioned on the final
driven shaft, and transmit the drive force that has been
reduced by the transmission to the left and right drive
wheels. This necessitates a large number of
transmission parts and a high level of torque
transmission, and the differential therefore represents a
significant percentage of the weight of the drivetrain.
In addition, due to restrictions on the operating angle
of the drive shaft, the final driven shaft is positioned
close to the rear end of the vehicle, at approximately
the same height as the centers of the wheels. This
means that a heavy component is located in a high
position, far from the vehicles center of gravity.
Reducing the weight and size of the differential would
contribute to lowering the vehicles center of gravity
and reducing yaw moment of inertia; reductions in
these parameters help to boost cornering performance.
In addition, the bias-adjusting mechanism also
incorporates a limited-slip differential (LSD) that
controls the right-left differential restriction torque, and
the performance of this unit is also an important factor.
This paper will discuss the development of a
lightweight and low center of gravity differential
reconciling the satisfaction of these performance
demands with durability.
The following aim was established in the initial

development project in order to realize a low center of
gravity and a low yaw moment of inertia.
(1) Positioning weight as low and as close to the center of
gravity as possible
The development project aimed to achieve a low
center of gravity and a low yaw moment by shifting the
bias-adjusting mechanism from the final driven shaft to
the final drive shaft, closer to the vehicles center of
gravity, and to a lower position.
Furthermore, the following aim was established for
the development of the ultra-short differential, based on
knowledge gained from the initial development and
consideration of advantages and disadvantages:
(2) Reducing the weight of parts located far from the center
of gravity and in a high position
This development aimed to contribute to the
realization of a low center of gravity and a low yaw
moment by reducing the size of the differential in order
to reduce its weight.
Regulations stipulate the total weight of Formula One
vehicles, meaning that a weight reduction would be
compensated for by using ballasts. By this means, it
would be possible to realize an even lower center of
gravity and yaw moment. Figure 1 shows an image of
the aims of the developments discussed here.

131
Final driven gear and

differential gears
significant structural changes to the gearbox case, and

the consequent increase in weight would nullify the merit
of lowering the center of gravity.
This development project demonstrated that the biasadjusting mechanism could be reduced in size through
the use of a full pinion engagement planetary gear
mechanism.
(2)
Front
(1)
Fig. 1
Method of achievement of low center gravity

and low yaw moment
3. Methods of Achievement of Development

Aims
3.1. Efforts to Centralize Mass
The first stage of the development project focused on
centralizing mass that would be effective in lowering the
center of gravity and reducing yaw moment of the
vehicle. A mechanism was developed that shifted the
differential, the heavy component of the gearbox, from
the final driven shaft to the final drive shaft, closer to
the vehicles center of gravity.
The transfer of the differential from the final driven
shaft to the final drive shaft necessitated the use of two
pairs of final gears. With respect to the width of the
gears, because it was sufficient that one of the gears was
able to transmit torque to the left or right wheel, the
tooth width was reduced by approximately half against
a conventional configuration.
The transfer of the bias-adjusting mechanism from
the driven shaft side to the drive shaft side, which has
a lower reduction ratio, helped to enable torque capacity
to be reduced and the unit to consequently be reduced
in size. Because the differential gears would be
incorporated in the final drive bevel gears, a compact
configuration was essential. A full pinion engagement
planetary gear configuration that reduced the shared load
through meshing of all contiguous pinion gears was
adopted.
The cases of Formula One gearboxes must be made
narrow, to help prevent interference with the flow of air
to the rear of the vehicle. To help enable this, a
configuration was adopted in which the LSD pistons
were positioned opposite the clutch, and rods were
introduced to connect them.
Figure 2 shows this bias-adjusting mechanism.
This configuration ultimately transferred
approximately 2.5 kg of mass to the final drive shaft,
reducing the yaw moment of inertia by 0.19% and
lowering the center of gravity by 0.4 mm.
Figure 3 shows an image of this transfer of mass.
Rig tests showed that this mechanism achieved the
necessary level of durability. However, the configuration
was not employed in a vehicle, because the transfer of
the differential to the final drive shaft would necessitate
3.2. Efforts to Reduce Weight (USD)

It was proposed that the bias-adjusting mechanism
should be positioned in the interior of the final driven
shaft between the right and left drive shaft joints in order
to achieve weight savings by reducing the size of the
final gears and the differential.
Figure 4 shows the layout of a standard planetary
bias-adjusting mechanism. This is a single-pinion
planetary configuration that is composed of four pairs of
mutually meshing single pinion gears, as differential
gears, and left and right drive shaft drive gears
positioned in parallel that mesh with the single pinion
gears. The configuration is positioned on the final shaft
outer circumference side and the inner side of the final
132
LSD piston/rod
Planetary gears
LSD clutch
Fig. 2
Final gears
Bias-adjusting mechanism on final drive shaft
Forward: 100 mm
Downward: 83 mm
Fig. 3
Image of movement of weight
Diff area
LSD area
Drive shaft joint area
Fig. 4
Ring gear
Standard differential
Sun gear
Pinion gear
Clearance
Fig. 5 Standard double

pinion planetary
gear
Fig. 6 Full pinion

engagement
double pinion
planetary gear
driven gears (the differential area) to drive the pinion

gears. The unit is also provided with an LSD that
produces left-right differential restriction torque using
hydraulic control (LSD area).
From considerations of strength, the existing external
diameters of the drive shaft joints could not be reduced.
In addition, an increase in the total width of the
differential was unavoidable if the existing mechanism
was to be made more compact (the diameter of the final
gears reduced), as it is fitted inside the final driven gear.
The configuration of the LSD had also become complex,
and no weight saving benefit could be obtained.
In order to resolve these issues, a double pinion
planetary bias-adjusting mechanism was adopted to help
enable the mechanism to be positioned between the drive
shaft joints (in the drive shaft joint area) and reduced
in size.
In the conventional single pinion planetary
configuration, three gear meshing positions left sun
gear-pinion gear, pinion gear-pinion gear, and pinion
gear-right sun gear, were arrayed in parallel. By contrast,
the gear meshing positions in the double pinion planetary
gear ring gear-pinion gear, pinion gear-pinion gear, and
pinion gear-sun gear - were arrayed vertically, and inline engagement was obtained in parallel. This helped to
enable the width of the bias-adjusting mechanism to be
reduced. In addition, in order to further reduce the width
and diameter of the double pinion planetary gears, the
number of pinion gears was increased to six pairs from
the former four pairs, reducing the shared load. However,
increasing the number of pinion gears made it necessary
to respond to interference between contiguous gears. If
the diameters of the pinion gears were reduced in order
to help prevent interference, it would be necessary to
increase the width of the teeth to maintain strength.
Again, if the angles of the interior and exterior pinion
gears were increased to help prevent interference, the
diameter of the ring gears would increase (Fig. 5). A
full pinion engagement pinion gear configuration was
therefore adopted to help enable the size of the ring
gears and the final gears positioned outside the ring gears
to be reduced. Figure 6 shows this configuration.
As a result, the distance between the final gear shafts
was reduced from 125 mm to 100 mm, and reductions
in size and weight were realized while the width of the
total assembly remained unchanged (Fig. 7).
4. Verification of Performance and Durability
Fig. 7
USD
4.1. Differential Friction

The purpose of a hydraulically controlled differential
mechanism is to restrict the differential motion between
the left and right wheels in response to the operating
status. If differential restriction torque exists during
steady cornering with no drive force acting, then
understeer is generated and this impedes drivability.
Because friction exists in the differential limiting
mechanism itself, even if oil pressure is entirely
removed, differential restriction occurs corresponding to
the amount of this friction. Given this, the development
133
aimed to achieve the same differential friction as that of

the proven 2008 specifications, and achieved almost
identical performance.
4.2. LSD Control Performance
The differential restriction torque demanded from the
differential during a race changes moment by moment
in response to the operating status of the vehicle. A
sufficiently short response time from the torque
command to the generation of differential restriction
torque is essential to controlling vehicle behavior as
desired.
Figure 8 compares the level of displacement of the
clutch assembly and torque response time for two
differential specifications with different levels of clutch
assembly stiffness. When engaged, the time from
commencement of hydraulic pressure rise to 80% of
maximum demand torque was evaluated, and when
disengaged, the time from the commencement of
hydraulic pressure drop to 5% of the same torque figure
was evaluated. The vertical axis of the graph shows the
level of axial displacement of the clutch assembly under
a specific clamp load. The higher this figure, the lower
the stiffness of the clutch assembly. Response time when
the clutch was engaged satisfied the target, bettering the
response time of the 2008 specifications. Neither
specification satisfied targets when the clutch was
disengaged, but a correlation with stiffness was verified.
The clarification of stiffness targets guided development.
The destruction rate of damaged final gears is 100%.

Figure 10 shows an S-N curve for the ring gears in this
development estimated based on this fact from the rig
test results for the final gears up to this point.
Structurally speaking, minimal benefit would be
obtained from reducing the width of the teeth of the final
gears in the USD against the previous final gears, and
some margin was thought to exist in terms of tooth root
bending stress. Easy-to-manufacture shapes were
therefore selected for the tooth roots, but target durability
could not be achieved due to a decline in strength.
The effect of the single radius produced by the
modification of the tooth root shape was verified in the
same rig tests that had been conducted previously on the
final gears. As in the case of the single radius, the
research experience confirmed that the application of
shot peening treatment to the tooth root increased the life
of the gears by approximately 15%. Applying these
results to the USD, the destruction rate corresponding to
four races, the period for which regulations stipulate that
the gears must be used, was 2.5%. The use of a single
radius for the tooth roots and the shot peening treatment
applied to the tooth roots would therefore increase
durability. These results are shown in Fig. 11.
The effectiveness of these measures was, however,
not verified due to the termination of the development
program.
Large
4.3. Durability
Verification of gearbox durability was essential, given
the regulation stipulates that gearboxes must be used for
four consecutive races. Rig tests were therefore
conducted using loads corresponding to four races on the
Monaco circuit, on which the frequency of use of the
differential is highest. In the initial durability test, fatigue
breakage occurred in the ring gears at approximately
60% of the target load. The following measures were
effected before the next durability test:
(1) The tooth root shape was modified.
(2) Shot peening treatment was applied.
The modification of the shape of the tooth roots
resulted in gear teeth being connected with a single
radius at the bottom. An image is shown in Fig. 9.
Previous shape
Single radius
Displacement
2008 spec line
Engage
Disengage
Response time
Fig. 8
Modified shape
Long
Torque response time
Fig. 9
134
Modification of root shape
S-N
3500
Frequency
S-N without shot peening
Stress amplitude [MPa]
3000
2500
2000
1500
1000
500
0
1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
N [Cycle]
Fig. 10
Until fatigue destruction
3500
Stress amplitude [MPa]
+15% 3000
with
shot peening
2500
S-N
Frequency
S-N without shot peening
S-N with shot peening
2000
1500
1000
500
0
1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
N [Cycle]
Fig. 11
4 races with large root R and shot peening
5. Conclusion
A full pinion engagement double planetary gear was
employed in a differential mechanism in order to reduce
the weight of the mechanism, lower the center of gravity,
and reduce yaw moment of the vehicle. The following
results were obtained:
(1) Compared to the 2008 specifications, the weight of the
unit was reduced by 1.2 kg (-14% in terms of weight
ratio), and the vehicles inertial moment was reduced by
0.30% and its center of gravity lowered by 1.1 mm.
(2) The distance between the final drive shafts was reduced
from 125 mm to 100 mm.
(3) Tests showed that the performance of the LSD could
match the performance of the 2008 specifications.
(4) Tests showed that a sufficient level of reliability could
be achieved to satisfy the regulation that gearboxes were
not to be replaced for a four-race period.
Author
Toshio HIYOSHI
Yoshikazu KATSUMASA
135
Yuichi SUENAGA
Haruki YOKOYAMA
Development of Clutch System

for Formula One Vehicle
Koichiro INUKAI*
Ryuhei KATAOKA*
Yasuo NAKAGAWA*
Yoshiki NAGATOSHI*
Hiroyuki KUSHIYAMA*
Keiji SHIOTA*
ABSTRACT
A direct push clutch (DPC) was developed to enhance the controllability of clutch torque in a Formula One clutch
system. A direct push mechanism in which clamp force was generated by a plate with the function as lever and a
hydraulic actuator was employed in place of the previous mechanism, in which the clamp force was generated by a
diaphragm spring. The transition from stroke control to direct control of the clamp force did away with elements that
formerly produced changes in system characteristics, such as variations in the diaphragm spring characteristic and
spring hysteresis, and thermal expansion throughout the system. In combination with other enhanced control
technologies, this enabled the development targets for the system to be realized. This helped to enable half-clutch
starts (slip starts) to be employed from 2006, even after regulations prohibited the use of feedback control. As a result,
a maximum reduction in time from 0-100 km/h of 0.46 sec (corresponding to a distance of approximately 11 m) was
obtained, giving the vehicle a competitive advantage. The developed system was used in races from the first race of
the 2006 season.
1. Introduction
One of the important functions demanded from the
clutch systems used in Formula One vehicles is the
control of torque at the start of the race in order to
maximize tire performance. In 2003, feedback control
using the tire slip rate as a parameter was commonly
used to control clutch torque, enabling optimal clutch
torque to be realized. However, changes to the Formula
One regulations in 2004 prohibited the use of feedback
control, necessitating the development of a clutch system
in which clutch torque could be predicted with a high
degree of accuracy.
The friction coefficient, , of the clutch friction
material and the clamp force acting on the clutch friction
material (clamp load) are factors that affect the accuracy
of prediction of clutch torque. Of these, the clamp force
can be controlled in a clutch system.
The conventional Formula One clutch system was
termed pull clutch due to its mode of operation, and
it originated in the clutches used in mass-production
manual transmissions. A diaphragm spring was used to
apply a clamp load to the dry carbon clutch friction
material. A hydraulic actuator controlled the stroke in the
direction that would detach the diaphragm spring, and
clutch torque was controlled with the stroke as a
parameter. Factors including variations in the dimensions
of the component parts, variations in the diaphragm

spring characteristics and spring hysteresis, and thermal
expansion of the parts would cause the clamp load to
change, making it challenging to accurately control
clutch torque. As a result, slip start, a start method
necessitating accurate clutch torque prediction, could not
be used. A race start method known as dump start was
employed to enable these issues to be avoided. This start
method is used in conjunction with engine over-speed
control, in which the clutch is instantaneously engaged
and the tires are intentionally made to slip. However,
when the tires are made to slip, their coefficient of
friction, m, temporarily declines, and ideal acceleration
is not obtained. Development therefore commenced on
a configuration that would do away with unstable
elements and employ a direct clamp load control method,
applying the clamp load directly by means of hydraulic
pressure.
In 2004, Honda worked in collaboration with ZF
SACHS Race Engineering GmbH (SRE below) to
develop a dry multi-plate direct push clutch (DPC)
provided with a lever disk plate with a return load that
helped to ensure zero torque operation and a lever
function that amplified the clamp load on the clutch
friction material. A twin-piston clutch actuator in which
hysteresis was reduced by lowering friction was
simultaneously developed. These developments enhanced

136
the accuracy of clamp load control, and helped to enable

slip start to be employed. The reliability of the developed
system was confirmed in vehicle tests in 2005, and it
was employed on the track from the first race of the
2006 season. This helped to enable the achievement of
a maximum reduction of 0.46 sec, corresponding to a
distance of approximately 11 m, when accelerating from
0-100 km/h at race start.
2. Development Aims
The mechanism of production of clutch torque is
simple, and can be expressed by the following equation:
Tc = FNrmi
(1)
Tc: Clutch torque

FN: Clamp load
: Friction coefficient
rm: Effective radius of friction material
i: Number of disk surfaces
The controllability of the clamp load and the
stability of the friction coefficient, , are therefore
important factors in the accurate prediction of clutch
torque.
In the conventional pull clutch configuration, the
clamp load was generated as the reaction force to
compression of the diaphragm spring by the clutch cover
(Fig. 1).
The clamp load therefore varied with variations in the
characteristics of the diaphragm spring and spring
hysteresis, and with variations in the dimensions of the
component parts and thermal expansion of the parts,
making accurate control a challenge.
The development project discussed in this paper
was conducted with the aim of doing away with
factors introducing uncertainty into clamp load
control and thus of enhancing the accuracy of clutch
torque prediction, in order to allow slip start to be
used in races and to obtain a competitive advantage
over competitors.
3. Methods of Achievement of Development

Aims
3.1. Development of Direct Push Clutch (DPC) with
Lever Plate
Employing a configuration in which the clamp load
was obtained directly from hydraulic pressure, the
developed system replaced the diaphragm spring used to
generate clamp force in the conventional pull clutch with
a lever plate (Fig. 2). The lever plate was fixed in place
between the foot of the clutch basket and the clutch cover
to enable it to produce a return load that would help ensure
zero torque. The hydraulic actuator pushed the lever plate,
and the lever plate made contact with the pressure plate.
The lever plate became a lever with this point of contact
as its working point, and a clamp load was generated on
the clutch friction material. The clamp load was determined
exclusively by the pushing load of the hydraulic actuator
and the lever ratio of the lever plate, and there was thus
no effect from the characteristics of the diaphragm spring
and variations in these characteristics, or variations in
dimension due to thermal expansion of the component
parts. Because the clamp load was directly determined with
the pushing load of the hydraulic actuator as the parameter,
the controllability of the clamp load was enhanced, and as
a result the accuracy of clutch torque prediction was also
enhanced.
3.2. Development of Low-friction Actuator
The reduction of hysteresis in the actuator was an
important factor in enabling high-accuracy control of the
clamp load using hydraulic pressure to be realized. The
major determinant of actuator hysteresis was seal friction,
which can be expressed using the following equation:
Fr = sdpbsP
(2)
Fr: Friction
s: Friction coefficient of the seal
dp: Piston diameter
bs: Seal width
P: Hydraulic pressure
Pressure plate
Lever plate
Diaphragm spring
Fig. 1
Pull clutch mechanical section
Fig. 2
137
Push clutch mechanical section
Seal friction is therefore proportional to piston

diameter (the seal length), and it can easily be seen that
plunger-type pistons present a greater advantage than
annular pistons. Taking the friction reduction effect and
the potential for prevention of stick by inclining of the
piston into consideration, a twin-piston actuator was
therefore adopted, reducing seal length by approximately
80% against that of an annular piston (Fig. 3, Fig. 4).
3.3. Clutch Engage System
Engine start was one of concern in using DPC.
Because the clutch was fully engaged when the engine
was stopped (zero hydraulic pressure) in the conventional
clutch system, the engine could be started by externally
rotating the layshaft on the gearbox input side, which is
positioned coaxially with the crankshaft. However, in the
case of a DPC, because no clamp load is generated when
the engine is stopped, the crankshaft will not rotate even
if the layshaft is rotated.
A clutch engage system (CES) was therefore
employed. In this case, a hydraulic pressure storage unit
known as a clutch disengage system (CDS), which had
previously been used to temporarily disengage the pull
clutch to enable a vehicle with a stopped engine to reach
a safe stopping point on the track, was used as a system
to engage the clutch before engine start.
3.4. Organization of Development Process and

Allocation of Roles
The development team discussed whether the
development process, from clutch to actuator, should be
conducted 100% in-house, or whether the development
should be conducted in collaboration with a specialist
clutch manufacturer. For the reasons listed below, it was
decided that the clutch unit would be developed in a
joint development with SRE, and that in the case of the
actuator, only the parts used in bench tests would be
manufactured in Japan and supplied, with the technology
of the parts for vehicle tests and race use to be
transferred to the race team (HRF1).
(1) The clutch is a consumable part needing day-to-day
maintenance. The support from a location close to the
race team was therefore judged to be more efficient.
(2) The lead time for manufacturing the carbon material that
was used in the clutch was long, and the production of
the material in small batches was challenging. In addition,
the level of technological challenge was high, and it was
therefore decided that in view of the development
schedule for the DPC, this aspect of the development
should be separated from the DPC project.
(3) In order to incorporate feedback from the race team
rapidly, it was judged that transfer to HRF1 would be
the most efficient means of maturing the design quickly.
Fig. 3
(Inner diameter)
54 mm
(Outer diameter)
56 mm
4. Results
4.1. Effect of Use of Slip Start for Race Start
The application of the developed clutch system
enabled slip start to be used in the absence of feedback
control. Figure 5 compares time for acceleration from
0-100 km/h using slip start and dump start.
A comparison of acceleration times to 100 km/h
shows that the use of slip start provided a competitive
advantage by helping to enable a maximum time
reduction of 0.46 sec, corresponding to a distance of
approximately 11 m. This would enable the race vehicle
to overtake one lead vehicle by the first corner.
4.2. Characteristics of Clutch Unit
This section will discuss results obtained for the
differences in characteristics between the DPC and the
Push clutch single-piston actuator
Clutch paddle
Clutch pressure
10.5 mm X 2
DPC
Engine speed
Pull
Acceleration
= 0.2G
Rear wheel speed
Car speed
Fig. 4
Push clutch twin-piston actuator
Fig. 5
138
100 kph
= 0.46 sec (11 m)
0 - 100 km/h pull clutch - DPC comparison
conventional pull clutch. In the case of the pull clutch,

the clamp load of close to 150 daN at race start was in
the range for maximum diaphragm spring hysteresis
(range of maximum pulling load), and a hysteresis of
20 daN was generated (Fig. 6). By contrast, the lever
plate used in the DPC was employed in an entirely
insensitive range to the diaphragm spring characteristic,
i.e., a range in which it worked as a lever to amplify
the load, and its hysteresis was controlled to around 10
daN (Fig. 7).
In addition to spring hysteresis, the pull clutch also
incorporated other uncertain elements, including changes
in the clamp load due to thermal expansion of the
constituent parts, variations in the diaphragm spring
characteristic, and stroke sensor offset due to heat, which
increased the uncertainty of controlling the clamp load.
system combining the DPC and the twin-piston actuator.

The graph in Fig. 9 plots clutch torque against actuator
pressure as measured during tests in circuit simulation
mode bench test with an engine. As in the case of the
actuator unit tests, torque was output in a linear
relationship with hydraulic pressure, with minimal effect
from uncertain elements such as wear of the clutch
friction material or thermal expansion. These results
indicated that the system fulfilled the initial development
targets established for it, accurately controlling clamp
load throughout its period of use, and helping to enable
clutch torque prediction.
1000
4.3. Actuator Friction Reduction

Figure 8 shows the results of measurements of clamp
load hysteresis against hydraulic pressure for the annulartype single-piston actuator and the plunger-type twin-piston
actuator. The twin-piston actuator displayed 80% lower
friction in a static test. This is a result of reducing the seal
length of the hydraulic piston by approximately 80%.
Clamp load [daN]
800
4.4. Characteristics of System

Figure 9 shows results for the characteristics of the
Single-piston
600
Twin-piston
400
200
Pull clutch clamp load

500
Clamp load (disk wear = 0 mm) [daN]

Clamp load (disk wear = 0.75 mm) [daN]
50
100
150
Actuator pressure [bar]
400
Load [daN]
Fig. 8
200
Actuator pressure - clamp load
300
Race start area
240
200
: Before race (12 km)

200
0
-2
-1
Fig. 6
1
2
Release travel [mm]
Input shaft torque [Nm]
100
Pull clutch: Stroke - clamp load
: Middle of race (315.3 km)

: After race (551.8 km)
160
120
Disengage
Race start area
80
Engage
40
DPC pressure clamp load characteristics

300
Clamp load [daN]
250
25
Disk wear = 0 mm
Disk wear = 0.6 mm
Fig. 9
200
50
75
100
125
Clutch actuator pressure [bar]
150
Clutch pressure - torque
150
5. Conclusion
100
Race start area
50
0
0
Fig. 7
20
40
60
Push clutch: Pressure - clamp load
80
Factors resulting in variations in the clamp load and

the friction of the seal in the actuator were focused on
in order to enhance the accuracy of torque control in a
carbon dry multi-plate clutch for use in a Formula One
vehicle. The use of a hydraulic direct push mechanism
employing a lever plate in the clutch unit and the use
139
of two low-diameter pistons in the thrust load generator

of the actuator produced the following results.
(1) A hydraulically-controlled DPC system enabling
accurate control of clamp load (clutch torque) was
developed.
(2) In 2006, the use of slip start without feedback control
was realized, enabling the achievement of a maximum
reduction of 0.46 sec (corresponding to approximately
11 m) in time for acceleration from 0-100 km/h.
The following issues for future study can also be
indicated.
(1) Variations in clutch characteristics occurred due to
friction material wear and the load characteristic of the
return spring in the lever plate (Fig. 7). This represents
an uncertain element. A wear correction system for the
hardware to help ensure that there is no change in the air
gaps even when wear occurs would be an ideal solution.
(2) In the range immediately following the generation of the
clamp load, a shelf characteristic appears by the
clearance between the lever plate and the surrounding
parts (Fig. 7). This is a factor that increases the
complexity of control, and correction of this characteristic
is therefore desirable.
(3) While a great deal of research has been conducted on
the subject, the friction coefficient of the carbon disk,
another element that determines clutch torque, is at
present not fully understood. The prediction of the friction
coefficient is a major theme in research aimed at
producing ideal clutch systems.
Acknowledgments
The authors wish to take this opportunity to express
their sincere gratitude to the staff of ZF SACHS Race
Engineering GmbH and technical consultant Leo Ress,
who collaborated in the development and manufacture of
the clutch unit discussed in this paper.
Author
Koichiro INUKAI
Yasuo NAKAGAWA
Hiroyuki KUSHIYAMA
Ryuhei KATAOKA
Yoshiki NAGATOSHI
Keiji SHIOTA
140
Descriptions of Chassis Technologies
Development Methodologies for Formula One

Aerodynamics
Atsushi OGAWA*
Takashi TAKIGUCHI*
Shujiro YANO*
Shinsuke NAKAMURA*
Susumu MASHIO*
Mitsuru SHINGAI*
ABSTRACT
The greater part of aerodynamics development for Formula One involves optimization using objective functions
of downforce and the lift-drag ratio (L/D), and verification of effects using wind tunnels and CFD in the development
process. As examples of development methodologies used to advance this optimization process, this paper will discuss
aerodynamics development tools, including wind tunnels and CFD, and the conventional development indices that
employed these tools. In addition, the paper will introduce new indices for analysis of the effect of tire deformation
on the aerodynamic load and analysis of transient aerodynamic characteristics during deceleration, and will also consider
the findings made regarding air flows using these indices.
1. Introduction
Downforce can be used to increase the side-force
limit on the tires when the vehicle is cornering and the
braking-force limit when the vehicle is decelerating. Air
resistance is an important factor in determining the
acceleration performance of the vehicle. The front-rear
downforce balance also contributes to vehicle stability.
The purpose of aerodynamics development is to
maximize either downforce or lift-drag ratio with
consideration of the trade-off between these three
elements.
During the development process, the vehicle shape
was optimized using a 50% scale model in wind tunnel
tests, following which effects were verified using fullscale wind tunnel tests. Simultaneous analysis of
aerodynamic phenomena using CFD and particle image
velocimetry (PIV) helped to enable the development in
the model-scale wind tunnel to move forward in an
efficient manner. It also became possible to a certain
extent to quantitatively evaluate aerodynamic load using
CFD, making it a tool capable of supporting part of the
optimization process. The importance of CFD is also
increasing as a bridge between wind tunnel tests and the
vehicle actually running on a race track. For example,
using CFD to reproduce the air flow when the tires were
deformed by side force, which could not be reproduced
using an actual vehicle in a wind tunnel, produced new
findings regarding the air flow around the vehicle
running on a race track. Some of these findings were
verified in wind tunnel tests.
In recent years, attention is beginning to be directed

towards unsteady aerodynamic characteristics when the
vehicle is vibrating or decelerating. In Formula One, the
maximum deceleration force reaches 5 G during braking,
and transient aerodynamic characteristics were frequently
an issue during deceleration. It was possible, however,
to elucidate the effects of some transient aerodynamic
phenomena by means of slow-motion tests in water
towing tanks.
2. Aerodynamics Development Tools

This chapter will discuss moving-belt wind tunnel
equipment and CFD as aerodynamics development tools.
2.1. The Wind Tunnel
2.1.1. Wind tunnel configuration
The Formula One aerodynamics development process
generally used a low speed closed circuit wind tunnel.
Open and closed test sections were employed. When
closed sections were employed, adaptive walls were used
to reduce blockage.
When a boundary layer formed on the floor of the
wind tunnel, this boundary layer would sometimes
separate prior to the boundary layer underneath the
vehicle, due to the adverse-pressure gradient underneath
the vehicle. The following countermeasures were put into
place to address this issue (Fig. 1).
First, knife-edge and perforated-plate structures were
used at the inlet of the test section to remove the
boundary layers on the floor. Additionally, downstream

142
from these structures, a moving belt was used to control

the occurrence of boundary layers. As a result, ground
surface boundary layers were not built above the limit
height (approximately 3 mm) of the boundary layer pitot
rake measurement, and the free stream was confirmed
to be maintained.
To help ensure that the belt was not lifted by the low
pressure under the vehicle, suction was applied from
underneath the belt when a plastic belt was used, and a
tensile force was applied between the front and rear drive
shafts when a stainless steel belt was used.
In order to minimize the minute deformation of the
different parts of the model when aerodynamic loads
were applied, the effect of expansion and contraction of
materials due to temperature, and temperature drift in
measurement devices, the temperature at the test section
was maintained at a constant level during tests. For 50%
model tests, the temperature was held at 25 C, and a
wind speed of approximately 45 m/s was used during the
optimization process.
2.1.2. Configuration of wind tunnel models and
measurement instruments
This section will first discuss the basic configuration
of a 50% model (Fig. 2). The model was supported by
a strut from the ceiling. The lower section of the strut
was fitted with a pivot to help enable the orientation of
the model to be changed, an actuator, and a sixcomponent load cell to measure the aerodynamic load on
the model.
The spine (framework) of the model was covered by
a replaceable bodywork. The front and rear of the spine
were fitted with laser ride-height sensors, to control pitch
angle and vehicle height. Three-component load cells
were positioned at the bases of the front wings (FW) and
the rear wings (RW) to measure downforce, drag, and
pitching moment.
The model was also fitted with yaw-, roll-, and steercontrol mechanisms, for the reproduction of complex
vehicle orientations. In addition, the entire floor of the
wind tunnel, including the moving belt, was capable of
yaw rotation. The model was also fitted with
anemometers at the radiator (vane anemometers or
differential pressure gauges), pressure scanners to
measure the static pressure at each part, and single-axis
load cells able to measure the drag on each wheel. The
wheel support method is discussed below.
Two types of wheel support were available: wheeloff and wheel-on. In the wheel-off configuration, the
wheels were supported by wheel support arms from
outside. The wheels were not in contact with the model,

and changes in ride height, therefore, did not generate a
reaction force in the model. However, slight inaccuracies
remained in geometries, including infinitesimal changes
in the camber of the wheels, and the distance between
the inner faces of the wheels and the adjacent vehicle
parts. The wheel-on configuration used a suspension
system with the same geometry as an actual vehicle. The
suspension was specially designed not to constrain the
motion of the pushrod, and could therefore remove a
reaction force from the ground. However, bump and
rebound of the suspension would produce a slight
amount of friction and a spring force which results in a
reaction force to the model and for this reason care was
necessary on corrections of the zero point (tare
measurements) of the measurement devices.
Formerly, rigid tires with trapezoidal cross-sections
(manufactured from aluminum or CFRP) which
reproduced the squashed shape close to the contact patch
around their entire circumferences were employed. More
recently, however, it has become standard practice to use
50% scale rubber tires provided by tire suppliers for
wind tunnel model use.
Tare measurements were normally conducted while
rotating the belt at a low speed, with a wind speed of
zero. Reproducing all the vehicle orientations that were
to be measured and taking tare measurements at each
measurement point during this process would help enable
more accurate measurements. However, this method was
disadvantageous in terms of measurement time. In many
cases therefore, measurement time was reduced by taking
one vehicle orientation as representative for tare
measurements, or multiplying by the linear
approximation of the ride height at this representative
orientation and the highest and lowest ride heights.
2.1.3. Flow visualization techniques
PIV measurements were conducted with the technical
assistance of the Honda Fundamental Technology
Research Center. This section will discuss the set-up for
three-dimensional PIV measurements.
A seeding device was positioned behind a heat
exchanger, in a space between flow-smoothing
honeycombs. The seeding particles were glycol-based,
and had a diameter of 1 micrometer. The laser optical
system, situated outside the path of the wind, illuminated
the periphery of the model. By changing the path of the
optical system and the position of the cameras, it was
FW load cell
6 comp. load cell

RW load cell
Pressure scanners
Knife edge
Suction
Wind speed
Wheel drag
load cells
Blowing
Belt speed
Radiator
anemometers
Laser ride height sensors
Suction
Fig. 1
Moving belt configuration
Fig. 2
143
50% model measurement system
possible to take measurements in the x, y, and z

directions (Fig. 3). However, there were limits to the
ability to take measurements in areas where the laser
could not physically reach, such as the x cross-sections
of the inboard faces of the tires.
2.2. Computational Fluid Dynamics
As indicated above, the importance of CFD in
Formula One development projects has today grown to
rival that of wind tunnels. The point to be observed in
the use of CFD is whether the air flows around vehicles
in wind tunnels and on a race track are qualitatively and
quantitatively reproduced. To provide an example, this
section will consider the method used when obtaining a
correlation between wind tunnel and CFD results for a
wake behind a tire.
The air flow around a rotating tire was elucidated in
a wind tunnel using a tire and PIV measurement
equipment. Figure 4 shows a vector diagram for a tire
center-section obtained using PIV measurements. Here,
the position of the commencement of separation from the
top of the tire and the size of the wake will be focused
on. (a) shows the results of PIV measurements, and (b)
and (c) show CFD results. The surface mesh was
approximately 1 mm, and the realizable k- model was
used as the turbulence model. (b) shows results for y+1, and (c) shows results for y+-10. y+ expresses the state
of the flow as a dimensionless number calculated from
the friction velocity at the nearest wall, the distance to
the nearest wall, and local kinetic viscosity. For the
simulation of separation, it was taken as effective to
place the first mesh on the viscous sub-layer (y+-1). The
height of the first mesh for Y+-1 was 10m.
Considering the results, (b) shows a good separation

position and wake size, but the wake is excessively large
in the case of (c). The result is in accord with the
simulation settings in which the air speed around the tire
was over-estimated.
As in the case of the tires, simulation methods for
other parts were also studied, and it was decided that the
following simulation methods would be used for the fullvehicle simulation:
The tetra mesh was used as the basic elements, and
prisms were created from surface triangle mesh on the
wings. 1 mm surface mesh was used for the airfoil
sections, and 2-4 mm surface meshes were used for other
parts. Figure 5 shows the mesh used for the FW. The
actual first height of the boundary layer mesh was 10 m.
Using a cocoon around the periphery of the vehicle,
the surface mesh size was increased to 10 mm. Figure
6 shows an image of a cocoon. Total pressure on the
cocoon was largely controlled within an overall isosurface pressure of CpT=0.7.
In the case of the volume meshes, the mesh number
differed depending on the growth rate from contiguous
meshes. A growth rate of 1.7 would generate a mesh
number of about 50 million, while a growth rate of 1.1
would produce a mesh number of about 300 million.
Figure 7 shows the difference in space mesh density
produced by a difference in the growth rate. It is clear
Beam from laser
Fig. 5
CCD camera 1
Mirror
Surface mesh of FW and symmetry plane
CCD camera 2
Cylindrical lens
Traverse system
Traverse along x -axis
Fig. 3
(a) PIV
Fig. 4
Fig. 6
PIV setting
(b) CFDy+ - 1
(c) CFDy+ - 10
Cocoon (iso-surface of CpT =0.7)
(a) Propagation 1.7
Velocity vector at tire center section
Fig. 7
144
(b) Propagation 1.1
Mesh propagation
that a low growth rate helps to enable a high mesh

density to be maintained even in areas distant from the
object.
Figure 8 shows the Cp distribution under the vehicle.
It was possible to maintain a higher mesh density around
the vehicle in the case of (b), which was generated using
a cocoon, than was the case for (a), generated without
using a cocoon. This minimized the effect of numerical
diffusion and, as a result, produced strong suction. This
indicates the effectiveness of using a high mesh density
and conducting the simulation within the effective wave
numbers.
A Linux device with a 128 GB memory was used to
run PrePost. The calculation server was equipped with
a 1024-core Xeon 3.0 GHz CPU, using InfiniBand
interconnects. At a growth rate of 1.1 generating 300
million meshes, 24 hours was required to converge
enough for FLUENT 6.3 with 256 parallel computations.
(a) Propagation 1.7
Fig. 8
number of RH points per map varies with the purpose

of the map. The maps used in development feature a
comparatively low number of RH, while the maps used
in vehicle dynamics analyses (termed mapping data)
are selected to incorporate a larger number of RH.
Using these RH maps, vehicle downforce was
optimized in relation to target drag for each circuit.
3.2.2. Adding yaw (cross-winds), steer, and roll
The recent advancement in wind tunnel facilities and
the associated control technologies have enabled yaw,
steer, and roll to be added to the tests in addition to RH
control. This has helped to enable aerodynamics
development using vehicle orientations closer to actual
conditions. Yaw (or cross-winds) and steer have become
indispensable elements in recent aerodynamics
development for Formula One, and developments are
now conducted using complex vehicle orientations in a
single RH map in which yaw, steer, and roll are
interwoven.
3.2.3. Trajectory map and curved flow
Because the aerodynamic load on a vehicle is
proportional to the square of its speed, it is sometimes
thought that the only case in which aerodynamics
dominates vehicle performance is that of high-speed
corners taken at close to 250 km/h. However, today it
is understood that the effect of aerodynamics is not
insignificant even in the case of low- and medium-speed
corners.
(b) Propagation 1.1
Cp contour of underbody, range from -2 to 0.
3. Wind Tunnel Optimization Indices

3.1. Regulations
The major part of aerodynamics development for
Formula One consists of optimization of the form of the
vehicle within the scope allowed by the FIA regulations.
In addition to specifying the regulation box, the scope
within which any aero parts must fit (Fig. 9), the
regulations also contain a flat-bottom rule prohibiting the
three-dimensional shape of the vehicle floor, a shadow
rule prohibiting the use of wings and other projections
that would provide ground effects in certain areas, and
a rule specifying the number of RW vanes.
The regulations are also revised every few years in
order to control vehicle speed increases resulting from
the ongoing evolution of aerodynamics.
145
Rear RH [mm]
3.2. Ride Height Map

3.2.1. The ride height map defined
Using an advanced control logic, a simulation was
conducted of the 50% model in the wind tunnel at the
ride height (RH) of an actual vehicle, and aerodynamic
performance was evaluated as the arithmetic mean of
multiple RH. These multiple vehicle orientations form
what is termed an RH map, which plays an important
role in determining in what direction the vehicle can be
optimized. Figure 10 shows an example of an RH map.
The most basic type of RH map is formed by
selecting representative RH figures that enclose groups
of RH measurements drawn from circuit data. The
Fig. 9
Regulation box and F1 chassis
70
65
60
55
50
45
40
35
30
25
20
10
15
20
25
30
Front RH [mm]
Fig. 10
Ride height map
35
40
Given this, together with the evolution of systems for

controlling changes in vehicle orientation, a method of
discretely reproducing a sequence of vehicle orientations
linked to cornering while braking and accelerating on an
RH map was developed. This is termed a trajectory map.
During actual cornering, a vehicle turns at a specific
angular velocity. Because of this, the air flows at the
front and rear of the vehicle have different inflow angles.
The angular velocity is greater in the case of a Formula
One vehicle than a normal car, and its effect therefore
cannot be ignored. The difference in the front-rear inflow
angle is particularly pronounced during low- and
medium-speed cornering, when the angular velocity is
high in relation to vehicle speed, and viewed from the
perspective of a vehicle, the air flow appears bent. This
is termed a curved flow (Fig. 11).
Curved flow could not, of course, be reproduced in
wind tunnel tests. CFD was therefore used to investigate
its effect, and the phenomenon was responded to by
adding a type of yaw correction to the trajectory map.
3.2.4. Weighting and penalties
Trajectory maps functioned very well to reproduce
the characteristics of specific corners, but lacked
flexibility in terms of evaluation of the overall dynamic
performance of the vehicle attributed to aerodynamics.
In addition, the importance of minimizing shifts in the
vehicles aerodynamic center (CoP) when the vehicle
orientation changes was reconfirmed from the
perspective of vehicle dynamic performance.
Given this, the viable options were weighting of the
RH and the application of penalties in response to shifts
in the CoP. Weighting of the RH involves consideration
of importance of low-, medium-, and high-speed corners,
and employing a weighted means weighted towards RH
corresponding to the corners where the level of
contribution to lap times is higher. The application of
penalties is a method of quantifying and evaluating the
level of robustness to changes in vehicle orientations, in
which shift of the CoP as a result of changes in vehicle
Equivalent air flow condition
At low-speed corners
Body side slip
Body side slip
Yaw rate
(large)
orientation is considered as a reduction in effective

downforce.
Even today, methods of evaluation of aerodynamic
performance and the vehicle dynamic performance it
generates are still being explored at the wind tunnel
test stage.
4. Formulation of New Development Indices

This chapter will discuss the new development
indices to be analyzed, the effect of tire deformation on
aerodynamic load and the transient characteristics of the
aerodynamic load produced on the RW during
deceleration.
4.1. Effect of Tire Deformation on Aerodynamic
Characteristics
When the vehicle is cornering, accelerating, or
decelerating, the tires are constantly deformed due to
vertical, lateral, and longitudinal forces (Fig. 12).
However, the only loads that can be applied to the tires
during full-scale wind tunnel tests are the vertical force
from the weight of the vehicle and the aerodynamic
downforce. Similarly to model-scale wind tunnel tests,
it is only possible to apply vertical loads from the wheels
and tires, and some side force from the belt. In other
words, the deformation of the tires due to side force or
longitudinal force when the vehicle is cornering,
accelerating or decelerating cannot be accurately
reproduced in a wind tunnel. It was therefore necessary
to study the effect of tire deformation on the
aerodynamic performance of the vehicle when it is
actually running on a race track, and use the data
obtained to formulate new indices for aerodynamics
development, such as methods of reproducing tire
deformation effect in a wind tunnel. This new index then
enables an aerodynamically robust vehicle against tire
deformation to be developed.
First, the forces acting on the tires when the vehicle
is running on a track were reproduced in rig tests, and
the shapes of deformed tires were obtained using threedimensional shape measurement systems. Using these
shapes, CFD was employed to analyze air flows, helping
to clarify the effects of the deformed tires to the
aerodynamic performance of the vehicle. Next, a method
of reproducing these flows in a wind tunnel was
formulated.
This chapter will focus on tire deformation due to
side force, which displayed the highest level of
contribution to vehicle aerodynamics.
Equivalent
Vertical force
Side force
Fig. 11
Curved flow
Fig. 12
146
Forces applied at cornering
4.1.1. Circuit data

The forces acting on the tires when the vehicle is in
operation were calculated from the pushrod load and
other parameters. As an example, Fig. 13 shows the side
force on the left front tire during a single lap of a circuit
obtained from circuit data, and Fig. 14 shows the
vertical position of the tire axis resulting from the force
acting on the tire. The maximum side force acting on
the outer tire reaches a maximum of 10000 N, and the
vertical displacement of the tire axis, i.e., the amount of
vertical squash displaced by the tire, is as high as 25 mm
due also to the vertical force from the downforce acting
at the moment. In addition, during deceleration, a load
of approximately 5000 N acts on the tire in a rearward
direction.
4.1.2. Shape measurement of deformed tires
The loads acting on tires for vehicles running on
track were reproduced in rig tests (Fig. 15), and the
shapes of the tires were measured using threedimensional measurement systems, including FARO,
Vectron, and T-Scan. However, because these were static
loads, the deformation of the tire due to centrifugal force
during rotation was not reproduced. As an example, Fig.
16 shows the shape of a tire subjected to a side force
of 7000 N. The side wall of the tire is deflected by

approximately 20 mm in the lateral direction at close
proximity to the road surface.
4.1.3 Analysis of flow when front tire is deformed by
side force
The measured front tire shape was analyzed using
CFD. This analysis showed that, in comparison to a tire
with no side force acting on it, approximately 5% of the
vehicles downforce was lost when a side force of 9000
N acted on the tire. Figure 17 shows the total pressure
distribution close to the road surface with and without
a side force acting on the tire. The results show that the
position of the separation point on the outboard-side wall
of the tire moves back significantly when a side force
acts on the tire.
This backwards shift of the separation point changes
the circulation around the tire in the XY planes, and the
tire wake which previously flowed to the outboard of the
vehicle now flows under the vehicle (Fig. 18). The fact
that this reduces the dynamic pressure underneath the
15000
Side force [N]
10000
5000
(a) Baseline
-5000
120
Fig. 16
130
140
150
160
170
180
190
(b) Fy: 7000 N
Snapshots of scanned tire
200
Time [s]
Tire loaded radius [mm]
Fig. 13
350
345
340
335
330
325
320
315
310
305
300
295
290
120
Front tire side force on track
(a) Baseline
130
140
150
160
170
180
190
200
(b) Fy: 9000 N
Time [s]
Fig. 14
Fig. 17
Front tire loaded radius on track
Total pressure distribution
Baseline
Fy: 9000 N
Fig. 15
Tire test rig
Fig. 18
147
Tire wake direction
vehicle, resulting in a decline in downforce, can be seen

from the change in the static pressure underneath the
vehicle when the tire goes from a state of no side force
to one in which side force is acting (Fig. 19).
Next, the reason for the shift of the separation point
was analyzed. First, a simulation was conducted to apply
the surface form of a tire with no side force acting on
it to outboard side of a tire with a side force applied
(Fig. 20). The resulting forward shift of the separation
point resulted in a recovery of 4% of the 5% loss in
downforce.
Next, the shape of the tire contact patch was focused
on, and a simulation was conducted with an edge applied
to the corner of the leading edge of the outboard side
of the tire with a side force applied (Fig. 21). In this
case, 4.5% of the downforce was recovered as well,
producing a result almost identical to that for a tire with
no side force acting on it.
Next, this phenomenon was reproduced using a 50%
model in a wind tunnel. However, because side force
cannot be applied to the tires in a wind tunnel, at first
only the vertical displacement was reproduced.
Figure 22 shows the results of measurement taken
using two-component PIV. As in the case of the CFD
results, the deformation of the tire resulted in the
backwards shift of the separation point on the outboardside wall of the tire. It also showed 3% loss of
downforce, which agrees with the CFD results

qualitatively as well.
Despite differences in the exact shape of tire
deformation, both CFD and wind tunnel results showed
that the deformation of the tires produces a backwards
shift of the separation point, which causes the loss of
downforce on the vehicle.
4.1.4. Analysis of flow when rear tire is deformed by
side force
CFD was used to conduct an analysis for the rear
tires in the same way as previously conducted for the
front tires. The results showed that approximately 4% of
the vehicles downforce was lost when a side force of
6500 N was applied, as compared to a state in which no
side force was applied. Figure 23 compares the total
(a) Baseline
dCp
1.000e-01
5.000e-02
0.000e+00
-5.000e-02
-1.000e-01
Fig. 19
(b) 10mm squashed
Deformed tire effect (delta Cp)
Fig. 22
PIV velocity vector
With side force

Baseline packer
Fig. 20
Outboard
Result of baseline packer

Inboard
Flow direction
(a) Baseline
Fig. 21
Result of edged contact patch corner
Fig. 23
148
(b) Fy: 6500 N
Tire wake comparison (total pressure)
pressure distribution close to the diffuser for tires with

no side force and a side force of 6500 N. When the side
force is applied, the separation area of the inner side of
the tire becomes large (shown by the solid circle in the
figure), simultaneously, as was the case for the front tire,
the separation point shifts to the rear (shown by the
broken circle in the figure), and the wake of the tire
flows to the center of the vehicle (shown by the solid
square in the figure). The fact that this becomes a
blockage and results in the decline of the diffuser flow,
with a consequent decline in the vehicles downforce,
can be judged from the change in the static pressure
underneath the vehicle when the tire goes from a state
of no side force to one in which side force is applied
(Fig. 24).
4.1.5. Verification of effect using full-scale wind tunnel
tests
The results obtained from CFD and model-scale wind
tunnel tests were verified using an actual vehicle in the
wind tunnel. Due to the system limitations of the wind
tunnel, it is impossible to apply any loads to the tires
other than the weight of the vehicle and the downforce.
Therefore, changes in the downforce produced by
changes in wind speed were used to vary the amount of
squash of the tires in the vertical direction, the effect of
which was analyzed. At this time, RH was adjusted by
varying the length of the push rods.
Figure 25 shows the amount of squashing of the
front tires and the rate of change of normalized
downforce against wind speed. The downforce at a wind
speed of 30 m/s was employed as the standard for the
rate of change of downforce. Normalized downforce
declines with the squashing of the tires as the downforce
increases with increasing wind speed. This matches the
tendency of results from CFD and the model-scale wind
tunnel tests, as discussed in Section 4.1.3. In addition,

PIV confirmed that this change in the normalized
downforce occurred with the shift of the separation point
on the outboard side of the tire.
Figure 26 shows the shape of a front tire when the
downforces generated at vehicle speeds of 40 m/s and
70 m/s were statically loaded on it. The results show that
the increase in load relaxes the apex of the outboard side
of the contact patch, and increases the degree of
deformation of the bottom of the tire.
However, in addition to the effect of the deformation
of the tires, the effects of deflection of bodywork due
to wind and variations in the position of the suspension
arms are also included in the changes in the downforce,
and these effects, however small they may be, need to
be taken into account.
4.1.6. Effect of differences between tire manufacturers
The shape of the deflection of the side of the tire and
the shape of the contact patch also differ between tire
manufacturers, and the effects of these parameters
therefore also differ. For example, if tires manufactured
by maker A and maker B which display an identical
amount of squashing are compared using CFD, the
vehicle downforce when maker Bs tires are used will
be approximately 5% lower. As Fig. 27 shows, this is
because the separation point on the outboard of maker
Bs front tires shifts rearward, and the wakes from the
inner sides of the makers rear tires are large. This
tendency matches the changes in magnitude of the
normalized downforce in the full-scale wind tunnel tests
dCp
1.000e-01
5.000e-02
0.000e+00
-5.000e-02
-1.000e-01
Fig. 24
Wind speed: 40 m/s

Wind speed: 70 m/s
Outbd
Inbd
Outbd
Fig. 26
Snapshot of scanned tire
Fig. 27
Total pressure under floor
Deformed tire effect (delta Cp)
0.0
Downforce
Front squash
Rear squash
-2.0
-4.0
-6.0
-8.0
-10.0
12.0
11.0
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
Squash amount [mm]
(CL-CL_30 ms)/CL_30 ms [%]
Maker A
2.0
Maker B
30 35 40 45 50 55 60 65 70 75 80
Wind speed [m/s]
Fig. 25
Squash and downforce change
149
Inbd
(Fig. 28). However, because the shape of the contact

patch also changes with the camber and other
parameters, a conclusion cannot be drawn until these
other factors are taken into consideration.
4.1.7. Methods of reproducing tire deflection in wind
tunnel
In order to obtain stable aerodynamic performance on
the track, it was necessary to increase robustness against
tire deformation. This made it necessary to reproduce
worst-case scenarios in wind tunnel development. It was
considered that using the pushrods to apply vertical loads
to the tire, or, as shown in Fig. 29, using a roller to push
the side wall from the inner side of the tire, would be
effective methods of doing so.
4.2. Transient Aerodynamic Characteristics during
Deceleration
Data obtained during track test had indicated the
possibility that normalized downforce declined when the
vehicle was decelerating. In particular, it was possible
that the normalized downforce was declining because of
the development of a boundary layer at the RW due to
deceleration and the fact that the RW wake overtook it.
An actual RW was therefore used in slow-motion tests
in a water towing tank, and the occurrence or nonoccurrence of separation and the changes in load when
the vehicle was decelerating were analyzed.
q=
1
U 2 = const
2
where : density and U: speed,

then at an atmospheric temperature of 25C and a
water temperature of 15C, the speed in water will be
air
Uwater =
Uair = 0.034Uair
water
i.e., the speed in water will be one-thirtieth of that

in air. The Reynolds number at this time is given by
Rewater =
Uwaterair
Reair = 0.47 Reair
Uairwater
where : dynamic viscosity coefficient. The Reynolds

number in water is half its value in air, but considering
changes in vehicle speed and the results of model-scale
wind tunnel tests, this was in an allowable range.
When the Froude number is made constant,
U2
= const
gL
where g: acceleration and L: reference length, the
acceleration in water is expressed by
4.6
Fr =
4.5
4.4
CL
4.2.1. Test conditions

First, in order to make non-aerodynamic factors such
as deformation of the RW under loads correspond to
conditions when the vehicle was actually running on a
race track, the aerodynamic loads in a towing tank and
that on a track were coordinated. If the forces in air and
in water are coordinated, i.e., the dynamic pressure is
made constant,
4.3
4.2
gwater =
Maker A
4.1
i.e., acceleration in water is 1/1000 of that in air.

Given this, it is possible to conduct slow-motion tests
in water, and transient aerodynamic loads during
deceleration, which are challenging to measure when the
vehicle is actually running on a race track due to error
factors, can be measured with a high degree of accuracy.
During this program of tests, deceleration tests from
0.005-0.05 G and fixed-speed tests from 1.02-2.94 m/s
(corresponding to 106-307 km/h in air) were conducted.
Maker B
4.0
3.9
30
35
40
45
50
55
60
65
70
75
80
Wind speed [m/s]
Fig. 28
U 2water
gair = 0.0012gair
U 2air
Normalized downforce comparison
4.2.2. Towing tank and test rig

The towing tank employed in the tests was 200 m
long, 10 m wide, and 5 m deep (Fig. 30). A six-
Fig. 29
Method of deforming rubber tire in wind tunnel
150
Fig. 30
Towing tank
component load cell and a mounting rig were attached

to the towing train, and the RW was positioned upsidedown in 1 m of water. To help prevent waves from being
generated on the surface of the water, an acrylic plate
was positioned close to the surface (Fig. 31).
4.2.3. Test results
In order to investigate the existence of transient
aerodynamic characteristics during deceleration, the RW
downforce when decelerating from 2.94 m/s
(corresponding to 307 km/h in air) at 0.005 G
(corresponding to 4.2 G in air) was compared with
values for driving at a constant speed (Fig. 32). The
same tests were conducted with braking G-force varied
down to 0.05 G. No transient characteristics during
deceleration were observed during these tests, verifying
that the RW was fulfilling its function as a unit
according to design targets even when the vehicle was
decelerating.
The result indicated that towing tank tests were an
effective method of verifying performance on a track
Acrylic plate
when, for example, dealing with higher-load RW or

studying aeroelasticity. In addition, the towing tank was
an effective means of reproducing new fluid phenomena,
such as when studying the possibility of a decline in the
normalized RW downforce during deceleration due to the
rear-tire wakes overtaking the RW.
5. Afterword
This paper has discussed aerodynamics development
indices used until 2008, and new indices such as the
analysis of the effect of tire deformation on aerodynamic
loads and the analysis of transient aerodynamic
characteristics during deceleration. These efforts
deepened understanding of the air flow around the
vehicle actually running on a race track.
Acknowledgments
The authors wish to take this opportunity to offer
their sincere thanks to all the staff members of the
former BAR and HRF1 for their many years of
assistance in aerodynamics developments, and also to the
staff of IHI Corporation who generously offered their
advice and assistance in the towing tank tests.
Author
Fig. 31
RW assembly
6000
Downforce [N]
5000
4000
Deceleration: 0.005 G
Constant speed
Constant speed
Constant speed
3000
Atsushi OGAWA
Shujiro YANO
Susumu MASHIO
Takashi TAKIGUCHI
Shinsuke NAKAMURA
Mitsuru SHINGAI
2000
1000
0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Speed [m/s]
Fig. 32
Hydrodynamic force comparison between

deceleration and constant speed modes
151
Atsushi OGAWA*
Yasutaka MASUMITSU*
Susumu MASHIO*
Masayuki MINAGAWA*
Daisuke NAKAMURA*
Yusuke NAKAI*
ABSTRACT
Formula One vehicles are fitted with a variety of aerodynamic devices. This produces complex mutual interference
in the air flows around the vehicles, generating highly nonlinear flows. The clarification of these aerodynamic
phenomena helps to enable efficient optimization of aerodynamic devices. This paper will provide some examples of
findings regarding the air flows around Formula One vehicles obtained using wind tunnels and CFD.
1. Introduction
The shapes of the aero parts employed on Honda
Formula One vehicles were optimized using 50% model
wind tunnel tests and CFD(1). The chief consideration
during this process was the tradeoff between downforce,
the front-rear balance of downforce, and drag. Mutual
interference between the air flows around these aero
parts produced a strong non-linearity in the air flows
around the vehicles, and in practice it was not possible
to clarify the mechanism of all aerodynamic phenomena.
However, the use of CFD to analyze the core
aerodynamic phenomena did generate new findings
concerning the air flows around Formula One vehicles.
This paper will discuss some of these findings.
2.2. The Entire Flow around the Vehicle

For the reasons discussed in Section 2.1., in addition
to generating downforce close to the ground, it is also
important to actively generate lift, which may initially
appear disadvantageous, in areas more distant from the
ground. In parts in which the influence of ground effects
is minimal, a significant downwash can be generated by
a small lift, and this can be used to increase the
downforce of the rear parts. To do so, lift-generating
parts called bunny ears and fox ears (see Chapter
w/ ground effect
w/o ground effect
Inflow
Upwash
2. The Entire Air Flow around the Vehicle
Ground
2.1. Ground Effects

This section will provide a simple discussion of twodimensional and three-dimensional ground effects. The
chief two-dimensional ground effect is an increase in
dynamic pressure and a decrease in upwash (a change
in the angle of the wake in an upward direction) due to
mirror images of the wings (Fig. 1). Because the upwash
of the wing is the angle of attack for aero parts behind
it, it plays a particularly important role in aerodynamics
involving multiple bodies, as in Formula One vehicles.
A reduction in upwash and a reduction in induced drag
due to mirror images of tip vortices can be indicated as
three-dimensional ground effects (Fig. 2).
152
Mirror
image
Fig. 1
2D ground effects
Actual inflow
Original inflow
Tip vortex
upwash of
mirror image
Tip vortex
upwash
Induced drag
Ground
Fig. 2
3D ground effects
4) were fitted on the upper part of the vehicle in order

to direct the upwash from the front wing (FW)
downwards and boost the performance of the floor, the
diffuser, and the rear wing (RW) (Fig. 3). The front
suspension arm and the nose also play a role in
generating downwash. This can be seen from the static
pressure distribution on the upper and lower surfaces of
the vehicle (Fig. 4).
This section has offered a rough picture of the air
flow around the vehicle, but the actual flow is composed
of countless small and large flow phenomena. The
following pages will present a discussion of some of
these phenomena.
LMD + LMD = (MD + MD)l

Mid area
MD + MD
Rear area
Front area
FR C1MD
LFR + LFR = (FR C1MD)C2l
Fig. 3
RR
LRR = (RR)C3l
Ideal chassis circulation
Low
and tip vortex it produces have a significant effect on

flows at the rear of the vehicle, and they also display a
high degree of sensitivity to changes in vehicle
orientation produced by steer, pitch, and other factors.
These characteristics, in addition to the fact that the
range of adjustment of the CoP must be considered,
make the FW the device on which most time is expended
in wind tunnel development.
3.1. Effect of FW Downforce and Consequent Upwash
The ground effect of the FW has a beneficial effect
on the lift-drag ratio, and also controls upwash to a
certain extent. However, if the downforce generated by
the FW is too strong, the limit of the lift generated by
the suspension arm and other parts behind the FW can
be reached, and upwash from the FW can reduce the
angle of attack on the devices at the rear of the vehicle,
and consequently reduce the downforce of the vehicle.
In addition, the wake from the FW and the suspension
arm can also reduce the dynamic pressure close to the
RW, resulting in a decline in the RW downforce (Fig.
6). Because of this, it was necessary to control the
downforce of the FW to a specific value or below, rather
than maximizing it. The surplus that this produces in the
regulation box (1) was used for a variety of purposes,
including the reduction of steer sensitivity.
In concrete terms, the height and shape of the FW
were adjusted in the Y direction in order to help improve
the upwash distribution and the steer sensitivity. It was
necessary to optimize the wing tips in coordination with
the front wing end plate (FWEP), which will be
discussed below.
High
Fig. 4
Cone
Surface static pressure
Foot plate
Strake
Fig. 5
3. Front Wings (FW)

The downforce produced by the FW (Fig. 5)
represents approximately 20-25% of the downforce of
the entire vehicle. In addition, the aerodynamic balance
(CoP) of the vehicle is extremely sensitive to the FW
downforce, and it therefore plays a role in the adjustment
of the CoP of the entire vehicle, through alteration of
the flap angle.
Because the FW is positioned upstream, the upwash
153
Front wing
Suspension
FW
Fig. 6
Suspension stall (CpT, Y=300 mm)
3.2. FW Tip Vortex and Tires

Another issue in FW development was the treatment
of the tip vortex shed by the FWEP. Until 2008, the
FWEP vortex was shed to the lower sections of the
inboard faces of the tires. A separation vortex generating
a significant total pressure loss was shed from the
leading edges of the lower sections of the inboard faces
of the tires (Fig. 7). One measure considered was to
merge the separation vortex and the FWEP vortex; the
merged vortices would remain in the same position (the
YZ position)as they go downstream, which would
prevent the vortices from going into the inboard side of
the vehicle and help to prevent the loss of dynamic
pressure underneath the vehicle floor (Fig. 8). It was
necessary for a stable FWEP vortex to be shed at an
appropriate position and at a strength at which vortex
breakdown would not occur.
In addition, the effect of steer could not be ignored
in FW development. Steer accompanies yaw angles in
the flows, and FW was optimized to help ensure that it
would not result in large variations between steer case
and no-steer case in the relationship between the
positions of the separation vortex shed from the inner
lower section of the windward tire and the FWEP vortex.
The FWEP tip vortex could be controlled by adding
appendages to the FWEP, or using a flat plate called a
foot at the bottom of the FWEP and a semi-conical plate
called a cone at the rear-end of the FWEP. It was also
possible to adjust the strength and position of the tip
vortex by means of the longitudinal vortex produced by
the strake, a perpendicular plate positioned in the center
of the FW (Fig. 5).
Cp_Total
0.00
Fig. 7
0.25
0.50
0.75
1.00
Tire LE separation wake (CpT, Z=0 mm)
2
FWEP vortex
Strake vortex
In addition to this, a variety of other methods were

employed, including raising the mounting position of the
outboard sides of the lower front suspension arm in order
to help prevent changes in air flow around the front
suspension due to changes in the vehicle ride-height
from affecting the position of the FWEP vortex.
4. Chassis Upper Devices

The chassis upper refers to the area extending from
the nose and the upper section of the monocoque to its
sides. While there are virtually no stipulations in the
regulations regarding limits on the vehicles in the
direction of height, the height of the front rollhoop is
the effective maximum height based on considerations
of visibility for the driver and safety around the cockpit.
Chassis upper devices mainly contribute to increasing
downforce under the floor of the vehicle and from rear
components such as the RW by means of the downwash
generated by lift.
4.1. Front Suspension Arm
The front suspension arm is the part that plays the
greatest role in deflecting the upwash flow from the FW
downwards by means of lift (Fig. 9). If no front
suspension arm was used, control of the position of the
barge-board vortex (see section 5.2.) would cease to
function entirely.
However, as the suspension is not allowed to be used
as an aero part, there are restrictions on the aspect ratio
of the cross-sectional thickness to the chord length of the
suspension members and their angle of attack. Because
of this, the upper section of the suspension arm can be
stalled by the upwash from the FW. The cross section
of the suspension arm and the suspension geometry were
therefore adjusted, among other measures, in order to
maximize the aerodynamic effect.
4.2. Bunny Ears and Fox Ears
Bunny ears that attach to the upper section of the
nose (Fig. 10) are another device that creates a
downwash. However, the bunny ears have more effect
in increasing the angle of attack to the RW than in
increasing the suction underneath the vehicle.
Consequently, while they increase downforce, the liftdrag ratio of the vehicle, being dependent on that of the
RW, does not improve. For this reason, bunny ears can
be indicated as a device that is particularly effective on
Tire vortex
4
Merged vortex
Merged vortex
Fig. 8
2 3
FWEP and tire vortices (CpT, velocity vector)
Fig. 9
154
FW upwash deflection by suspension arm
high-downforce circuits. In addition, because there are

few aero parts on the upper section of the chassis, the
position of the tip vortex from the bunny ears remains
stable to vehicle orientations and disturbances, which
makes the bunny ears effective in crosswinds (Fig. 11).
The use of bunny ears resulted in an increase of
approximately 2% in maximum downforce.
Fox ears (Fig. 12) are fin-shaped devices mounted
in front of the mirrors. Like bunny ears, fox ears increase

the angle of attack to the RW, in addition to which they
contribute to increased downwash to the vehicle floor.
This downwash also affects the position of the
longitudinal vortex of the barge board. The Formula One
regulations state that aero parts in this area can only be
positioned above parts fitted to the floor of the vehicle
(the shadow rule). This added the limitation that the fox
ears must not extend past the projection on the XY plane
of the water wings (WW below; Fig. 13) directly
below them. In addition, while the effect of the fox ears
would increase the higher the outer tips were positioned,
the potential obstruction of the drivers field of vision
significantly limited the degree of freedom of the shape
of the device.
5. Chassis Lower Devices
Fig. 10
Bunny ears
Blue: low pressure
Side wind
Fig. 11
Effect of bunny ears (delta Cp)
Fig. 12
Vane
Vertical fence
The chassis lower area refers to the area from the

front axle to the front ends of the side pods (Fig. 13).
There are many regulation stipulations regarding
minimum heights and maximum widths in this area. The
shadow rule also applies across almost the entire area,
and there are many regulations concerning the shapes of
parts. In addition to the increase in downwash by means
of lift, as achieved by chassis upper devices, downwash
is also generated by longitudinal vortices. By this means,
the mass flow under the floor is increased, and
downforce is generated even by devices in front of the
floor. Because of this, it is necessary to consider
interference between almost all devices from the FW to
the diffuser, and the level of contribution of this area to
vehicle performance is high, making it important in
terms of aerodynamic performance.
5.1. Barge Board Longitudinal Vortex Control
The barge board (Fig. 14) functions to generate a
downwash in front of the floor of the vehicle by
controlling the longitudinal vortices that it produces. The
shape of the barge board varies between teams, but its
role in each case is generally identical.
The cambered barge board produces upper and lower
tip vortices (Fig. 15). Directly behind the barge board,
the flow induced between the vortices is directed
outboard of the vehicle. As it moves downstream, the
upper vortex moves outboard and downwards by the
lower vortex, the vanes, the suspension, the side pods,
Fox ear
Water wing
Barge board
Hammer head
Fig. 13
Barge board
Devices on chassis lower section
Fig. 14
155
Barge board
and other parts. The lower vortex proceeds downstream,

increasing in intensity as it induces longitudinal vortices
at the vertical fence and the WW. Because of this, and
also due to ground effects, the lower vortex, shed from
the barge board, remains in the same YZ position as was
shed from the barge board. In this way, a displacement
in the Y direction is produced in the upper and lower
vortices, and the flow induced between the vortices is
directed downwards. This is a downwash produced by
the barge-board longitudinal vortices. This downwash
increases the angle of attack towards the underfloor, and
increases suction at the leading edge of the floor. In
addition, while the effect is small, downforce is also
increased by the suction of the lower vortex itself that
flows under the floor of the vehicle. However, excess
suction at the leading edge of the floor and total pressure
loss at the center of the lower vortex can promote growth
of the boundary layer on the floor and result in a
decrease in diffuser performance. Because of this, the
sequence of aero parts to be optimized and setting
priorities in optimization are important issues.
The barge board does not produce any other effects
as significant as the downwash effect produced by the
longitudinal vortices, but it does play another role.
Figure 16 shows a comparison of streamlines with and
without a barge board. The outward flow created by the
barge board pushes the separation wake of the leading
edge of the inboard lower section of the front tire
outwards, helping to prevent a decrease in diffuser
1
Vane vortex
2
Upper vortex
Upper vortex
performance. However, very little change is observed in

the large separation flow at the back of the tires whether
or not a barge board is used. Along with the fact that
the separation wake from the leading edges of the tire
interacts with the FW tip vortex, as indicated before,
these results show how strongly the barge board interacts
with front and rear aero parts.
5.2. Development of O-nose Fence
Adding to the effect of the barge board in controlling
longitudinal vortices, the O-nose fences (ONF) (Fig. 17)
increase downforce by means of the suction they
produce.
The ONF generates strong suction below the nose by
means of a Venturi effect produced between the left and
right fences extending from the nose towards the ground
(Fig. 18). In addition, the bottom edges of the fences
are close enough to the ground to seal the flow through
the gap between the ground and the fences, which
increases this effect further. However, if strong suction
is generated at the lower section of the nose, an upwash
is also generated as a consequent reaction, and this can
result in a decrease in underfloor performance. Further,
as in the case of the barge board, as a result of the strong
suction generated by the fences, the separation vortices
from the leading edges of the inboard lower sections of
the tires are caught up in the lower vortices of the
fences, and this can also result in a decrease in
underfloor performance.
To resolve these issues, the downwash was
strengthened through the use of a high-lift vane, a
strong longitudinal vortex was generated through the
use of biplanar ONF, the use of locally high camber
profiles on the ONF, and the addition of a shadow
Lower vortex
Vane
HH
WW
Flap
Shadow plate
Fig. 15
BBD vortices
Fig. 17
O-nose fence
Fig. 18
ONF suction
Without barge board

LE wake
TE wake
With barge board
Fig. 16
Tire wake direction
156
plate, and the position in which the upper vortex was

shed was controlled by adjusting the relationship
between the ONF flap height and the position of the
vanes. These measures helped to control the generation
of upwash. The optimal placement of the cambers also
alleviated the drawing-in of the separation vortices from
the leading edges of the tires.
The application of these measures helped to enable
the ONF to generate the same level of downwash as the
barge board with no loss of suction at the lower section
of the nose (Fig. 19). As a result, the ONF used at the
time increased the lift-drag ratio of the vehicle by 1.5%.
5.3. Effective Use of Downwash of HH and WW
Hammer heads (HH) are parts attached to the vehicle
to maximize the use of the downwash of the barge board
and other parts. To obtain an effective aspect ratio from
the hammer heads, they are extended forwards in the X
direction, rather than the Y direction, in which
regulations stipulate the maximum vehicle width. The
fact that the downwash flows outboard is another reason
that the HH, with their leading edges positioned towards
inboard, are effective. In addition, gurneys and other
parts are attached to the upper section of the trailing
edge in order to increase circulation.
The role of the WW, touched on above, will now be
discussed in a little more detail. In passing the lower
section of the WW, the lower vortex generated by the
barge board produces a pressure difference between the
upper and lower sections of the WW, which generates
the tip vortex from the WW and increases the intensity
of the barge board lower vortex. The WW also play
another important role. As Fig. 20 shows, the flows at
1
Vane vortex
the leading edges of the HH are accelerated and the

suction at the HH is increased by narrowing the gap
between the WW and the HH. However, if too much
downforce is generated, the downforce of the rear aero
parts will be reduced as a result of upwash and the
growth of underfloor boundary layers.
5.4. Downwash produced by SPLEF
The side-pod leading-edge flickups (SPLEF) are
vertical plates with L-shaped YZ sections that are
positioned on the exterior of the side pods (Fig. 21). The
role of the SPLEF is to intensify the downwash to the
HH by providing a blockage and producing circulation.
The use of the SPLEF increases the downforce of the
HH, but they also generate high levels of induced drag
and pressure drag acting on them. However, it was
possible to increase the lift-drag ratio through the
application of treatments to the leading edges of the
vertical plates and the optimization of their shape.
5.5. Ideal Flow
Figure 22 shows the surface streamlines from the
nose to the chassis. The figure shows that the use of
chassis upper and lower aero parts has resulted in the
FW upwash being rapidly directed towards the
underfloor the ideal flow for a Formula One vehicle,
as described in section 2.2.
Vane vortex
Upper vortex
Fig. 21
SPLEF
Lower vortex
1
Fig. 19
ONF vortices
Fig. 22
Surface streamlines
6. Bodywork
WW
HH
The bodywork refers to the areas around the side

pods (SP) and the cowl.
1
Fig. 20
Channel flow
6.1. Effect of SP Undercut

From 2000, most teams began using a curved shape
called an undercut for the lower sections of the anterior
halves of the SP [Fig. 23(b)].
157
The SP undercuts can be indicated as having three

main effects. The first of these is an underfloor seal
effect. The circle in Figure 24(b) shows a hypothetical
side wall connecting the side edges of the floor and the
ground vertically. Without such a side wall, the flows
from the leading edges of the SP flow out from under
the floor and flow back under the floor and towards the
diffuser, but part of the flows also flow out towards the
rear tires [Fig. 24(a)]. By contrast, with side walls in
place, the underfloor flows are sealed in and flow in
straight lines to the diffuser. This accelerates the
underfloor flows, boosting suction and increasing
downforce [Fig. 24(b)]. However, because the flows
towards the rear tires are also blocked, the suction at the
leading edges of the SP tends to become weak. In other
words, if only the middle and the rear section of the
floor was sealed, the maximum benefit of the seal effect
could be obtained.
The SP undercuts help to enable this effect to be
realized within the scope of the regulations. First, as a
result of the undercuts, the influence of the downwash
from the front half of the vehicle is extended to the
center of the floor (Fig. 25). There, the downwash plays
the role of side walls, producing a floor seal effect and
increasing underfloor suction (Fig. 26). An increase in
the suction at the sides of the SP also accelerates the
downwash, further intensifying the seal effect.

The front of the floor, where no seal effect is
necessary, could be used to extend the HH, and it was
possible to increase suction on the lower sections of the
HH through the use of devices such as gurneys on the
trailing edges. This was the second effect of the
undercuts.
Depending on the type of barge board employed, the
barge-board upper vortex might directly pass the
undercuts (Fig. 27). This upper vortex also produces a
seal effect, but has a greater effect in optimizing the
suction on the lower section of the HH in the direction
of the span, by means of the position at which it passes
the HH. This is the third effect of the undercuts.
However, the effects are not uniform, due to the fact that
the intensity and height of the upper vortex differ with
different types of barge boards, and its optimum spanwise position to pass the HH also differs. The large
barge boards employed by almost all teams generally
function to promote the effects of the undercuts.
The use of undercuts rather than linear side pods
produced a 1.5% increase in downforce in model-scale
wind tunnel tests. The undercuts also helped to increase
the underfloor suction in line with predictions. It is also
confirmed that the undercut shape was effective even
when a barge board was not used.
6.2. Analysis of Chimney Exhaust Air Flow
The flows that pass the radiators and oil coolers
located inside the SP are exhausted through openings in
the sides and rear end of the cowl. Cooling performance
is adjusted by changing the area of the side openings,
but the exhaust flows have a considerable effect on
aerodynamic performance of the vehicle.
(a) Flat SP
(b) Undercut SP
Fig. 23
Side pod variations

Streamline
PATCH: dp
2.000e - 02
1.000e - 02
0.000e+00
-1.000e - 02
-2.000e - 02
Cp
(a) Without side fence
Fig. 24
Fig. 26
Undercut SP effect (delta Cp)
(b) With side fence
Floor seal effect
Upper vortex
Fig. 25
Downwash flow
Fig. 27
158
BBD upper vortex path
From the beginning of the 2000s, chimney-shaped

exhaust holes (termed chimneys) began to be
employed (Fig. 28). Cooling performance was adjusted
by means of the openings at the top of the chimneys.
Wind tunnel tests using a model showed that fully
opening the openings from a fully closed state increased
vehicle downforce by 4% and vehicle drag by 5%. The
exhaust flow affects the RW load, and half of these
increases were due to the change in the load on the RW.
Figure 29 shows the total pressure distribution when
the openings of the chimneys are fully closed and when
they are fully open. The areas surrounded by the broken
circles are wakes from the back of a front tire, and
include areas of high-pressure loss (Cpt=0.3). This wake
passes above the openings of the chimneys, and part of
the wake flows close to the tips of the RW (shown
surrounded by the solid squares in the figure). When the
chimney openings are fully open, the wake from the
front tire is shifted upwards and towards the outboard
of the vehicle by the exhaust flow from the chimneys
(shown surrounded by the solid circles in the figure). By
this means, the area of pressure loss at the RW is
reduced in size, and the load on the RW is increased.
Because this change in load is due to a change in
dynamic pressure, there is almost no change in the liftdrag ratio of the RW.
Chimney
Upper flick up
Fig. 28
It was also necessary to consider the increase in

downforce and induced drag due to the upwash from the
chimneys, and the increase in pressure drag due to the
increase in the cooling flow. In addition, because the
chimneys generated suction on the sides of the SP, like
the SP undercuts they produced an underfloor seal effect
(Fig. 30).
CFD was employed to study the effect of the
blockage, and it was determined that the direction of the
wakes from the front tires was affected by the size of
the blockage, and the degree of change in the load on
the RW due to the opening of the chimneys was affected
by the blockage. It was necessary to give sufficient
attention to this point in wind tunnel tests.
7. Diffuser
The Formula One regulations strictly determine the
dimensions of the diffuser (Fig. 31). Within the scope
of these regulations, the suction of the diffuser and the
underfloor area was increased by maximizing the
effective sectional area and minimizing the static
pressure at the exit.
7.1. Maximization of Effective Sectional Area of Diffuser
Exit
Separation in the areas in which pressure is
recovered, the in-flow from the upper section at the
diffuser tips, and the entry of the leading-edge separation
vortices of the rear tires can be indicated as factors that
reduce the effective sectional area of the diffuser exit
[Fig. 32(a)].
Pressure recovery is optimized by the kick-up shape,
the section shape, and the fence to produce as uniform
Chimney and upper flick-up

PATCH: dp
5.000e - 02
2.500e - 02
0.000e+00
-2.500e - 02
-5.000e - 02
Fig. 30
(a) Closed
Fig. 29
Chimney effect (delta Cp)
(b) Open
Total pressure distribution in section around

chimney trailing edge and RW leading edge
Fig. 31
159
Diffuser
a pressure distribution as possible in the Y direction

without generating separation locally (Fig. 33). The flow
from the upper section and the entry of the tireseparation vortices are controlled to a certain extent
chiefly by the shape of the foot [Fig. 32(b)].
7.2. Minimization of Static Pressure of Exits
Figure 34 shows the static pressure distribution at
the diffuser exit. The static pressure at the diffuser exit
is affected by the position of the suspension members
and the lower rear wings. The static pressure at the exit
can be reduced, and the diffuser flow-rate increased, by
adjusting these positions and shapes and optimizing the
suction peak location of the upper rear wings.
7700
Normalized rear downforce
7.3. Effect of Front Half of Vehicle

The diffuser, positioned at the rear of the vehicle, is
affected by the flows from the front half of the vehicle.
For example, if the suction at the inlet of the floor is
increased, or the lower vortex of the barge board is
intensified, boundary-layer thickness will increase under
the floor, and the separation toughness of the diffuser
will decline. Because of this, the choice of which area
aerodynamic development proceeds from, and at what
timing, have a significant effect on the aerodynamic
characteristics of the vehicle.
7.4. Aerodynamic Characteristics produced by

Changes in Ride Height
The area ratio of the entrance and exit of the diffuser
change with the ride height of the vehicle, and the
relative position of the diffuser in relation to the
suspension also changes. Because of these changes, the
downforce generated also changes.
The downforce characteristic produced by ride
height is determined based on a variety of
considerations, including the CoP characteristic at lowand high-speed cornering and during braking. In most
cases, the CoP during braking is shifted to the rear of
the vehicle for the sake of braking stability. This means
that aerodynamic development is conducted to produce
an increase in the dimensionless rear downforce as the
rear ride height increases due to braking (Fig. 35,
broken line).
In order to realize this characteristic, vehicle design
attempts to ensure that at a low rear ride height there is
partial separation from the diffuser and the downforce
is reduced. However, caution is necessary, because if the
level of separation is too great, hysteresis may prevent
the downforce from being recovered when the vehicle
is braking (Fig. 35, solid line).
Tire wake
(a) Velocity vector
(b) Cp
7600
7500
With separation
(acceleration)
7400
7300
With separation
(braking)
7200
Without
separation
7100
7000
0
20
40
60
80
Rear RH (mm)
Fig. 32
Fig. 35
Flow around diffuser
Ride height aero characteristics
8. Rear Wings
Fig. 33
Fig. 34
Diffuser Cp
Diffuser exit static pressure
The RW are positioned in an area specified by

regulations, behind the rear axle and 300 mm higher in
the Z direction than the reference flat section of the
vehicle bottom. From 2004, the regulations allowed two
elements in the upper section and one element in the
lower section of the RW regulation box.
Because the RW have low aspect ratios, an upwash
distribution is formed by the tip vortices in the direction
of flow (Fig. 36), and massive separation does not occur
even when an airfoil with a high camber is used.
Induced drag represents the major component of RW
drag, and an increase in drag necessarily follows from
an increase in downforce. In addition, the downforce and
drag generated by the RW represents a considerable
percentage of the downforce and drag for the vehicle as
a whole, and is therefore used in the adjustment of the
160
balance between maximum speed and cornering

performance. This is to say that it is necessary to use
several RW generating different levels of downforce in
order to cover the speed characteristics of all circuits.
The purpose of RW development is to help increase the
lift-drag ratio for the vehicle as a whole by developing,
among other considerations, efficient airfoils, load
distribution in the Y direction, and RW end plates for
each of these various downforce levels.
8.1. RW with Raised Tips
Figure 37 shows the distribution of the angle of
attack when the upper RW is not in place, as obtained
by CFD. These results show that there is a significant
upwash at the tips of the RW.
Figure 38 shows a YZ section of the total pressure
distribution close to the tips of the RW. The tips of the
RW are exposed to the wake from the front half of the
vehicle, and the total pressure declines. Because of this,
the load on the tips declines. An upwash is generated
close to Y 300 mm, and a downwash close to the wing
tips, alleviating this sudden change in load in the Y
direction (Fig. 39). This phenomenon can also be seen

from the fact that at the tips the pressure distribution in
the direction of flow (Fig. 40) shows characteristic of
high angles of attack such as a strong peak at the leading
edge (Y = 450 mm) though the angle of attack forms
an upwash at the tips when no upper RW is placed.
The reduction of the camber of the tips in order to
reduce the load could therefore be expected to reduce
pressure drag and increase the lift-drag ratio. Other than
this, optimization was conducted in the RW development
with consideration of changes in induced drag.
8.2. Increasing Robustness by means of Pillar Shape
In most cases, pillars are used to support the centers
of RW, and have the function of reducing the weight of
the entire structure. However, the air flow undergoes a
sudden acceleration between the left and right pillars,
necessitating strong pressure recovery. In rearward-
Y300
Inflow
Upwash
distribution
Fig. 36
Fig. 39
Secondary vortex due to RW load distribution
Upwash by tip vortices

Upper rear wing - 2D slices
1.00
Pressure coefficient
0.00
Red: Upwash
Blue: Downwash
Fig. 37
-1.00
-2.00
Y = -1 mm
Y = -50 mm
Y = -100 mm
Y = -150 mm
Y = -200 mm
Y = -250 mm
Y = -300 mm
Y = -350 mm
Y = -400 mm
Y = -450 mm
Y = -490 mm
-3.00
-4.00
Inflow angle for RW
-5.00
3.40
Total pressure
3.45
Fig. 40
3.50
3.55
3.60
X-coordinate (m)
3.65
Sectional Cp distribution
Front tire wake, etc
Fig. 38
Cp total around RW
Fig. 41
161
RW with raised tip
3.70
Reference
(1) Ogawa, A., Yano, S., Mashio, S., Takiguchi, T.,
Nakamura, S., Shingai, M.: Development
Methodologies for Formula One Aerodynamics, Honda
RA107
Fig. 42
RA108
Boundary layer shear direction on pillars
leaning pillars like those used in the RA 107 (Fig. 42),

the spanwise suction distribution produced by the rear
inclination directs the flow in the boundary-layer
upwards, extending the distance at which boundary
layers grows; at the same time, a crossflow occurs
between the boundary layer and the outer flow. As the
boundary-layer of the pillars merges with that of RW,
these are thought to be the factors behind a decline in
the separation stability of the RW.
The following changes were made in order to resolve
the instability produced by the pillars in the wings
developed for the RA 108. First, the space between the
pillars was increased to the regulation limit and the
sectional shape of the pillars was modified in order to
ease pressure recovery. The pillars were also moved to
the leading edges of the wings in order to separate the
pressure recovery areas of the wings and the pillars. In
addition, small wings were positioned between the pillars
in order to ease pressure recovery on the RW. Finally,
the rearward inclination of the pillars was minimized in
order to control the upward inclination of the pillar
boundary layer flow.
9. Afterword
The air flows around multiple bodies such as the
chassis of Formula One vehicles display a strong nonlinearity, and it is not possible in practice to understand
the detailed mechanisms of all the aerodynamic
phenomena involved. However, aerodynamics
developments were conducted efficiently by the use of
CFD for qualitative analysis of the core aerodynamic
phenomena, backed up by quantitative data obtained in
wind tunnel tests. These methods also enabled the
accuracy of predicting the aerodynamic performance of
the vehicle when it is actually running on a race track
to be increased.
Author
Atsushi OGAWA
Susumu MASHIO
Daisuke NAKAMURA
Yasutaka MASUMITSU
Masayuki MINAGAWA
Yusuke NAKAI
10. Acknowledgments
The authors would like to take this opportunity to
offer their sincere thanks to all the staff members of the
former BAR and HRF1 for their many years of shared
developments in aerodynamics, and to the other members
of the Honda team and staff members of other companies
who generously assisted in a wide variety of ways in
development projects.
162
Technologies for Enhancement of Dynamic

Performance of Formula One Vehicle
Hideaki SHIBUE*
Kazuhiro TANEDA*
Yasutaka KITAKI*
ABSTRACT
No matter how superb the performance of the powerplant, unless that power can be transmitted efficiently and
effectively through the tires to the track surface, a race vehicle will not have a competitive edge. The most recent
technologies for the enhancement of the dynamic performance of Formula One vehicles are developed with a focus
on maximizing tire performance. The tires used in Formula One are designed for good performance only within an
extremely narrow range of conditions in terms of parameters such as tire contact state and tire temperature, in order
to enhance the performance of the tires to the limit. Therefore, understanding and controlling these conditions is an
important issue in the development of enhancement technologies for the dynamic performance of Formula One vehicles.
This paper will discuss suspension design and vehicle setup, two factors that affect tire performance, and the
development of a tire model to function as their theoretical basis.
1. Introduction
2. Suspension Design
Development programs to enhance the dynamic

performance of Formula One vehicles are conducted in
two stages. The first is the design stage for vehicle
performance. At this stage, the vehicle concept and
targets for the various elements of dynamic performance
are established based on performance analyses of Honda
vehicles and other teams vehicles, and vehicle
components enabling these targets to be realized are
designed.
The second stage is vehicle setup on an actual circuit.
The dynamic performance of Formula One vehicles
varies with the layout of the circuit, track conditions, and
atmospheric conditions, and a vehicle will therefore not
always display identical performance. Accordingly, at
this stage, element parts with varying specifications and
mechanisms that enable vehicle characteristics to be
varied are prepared, and the vehicles dynamic
performance in actual circuit driving is optimized by
means of their application and coordination.
While vehicle design determines the potential
dynamic performance of the vehicle, setup on the circuit
enables the vehicle to reach that potential in response to
circuit conditions and race strategy. The use of these two
stages enables the vehicle to display a high level of
dynamic performance in a race.
The paper will discuss the concepts employed in
suspension design and setup, factors that affect tire
performance, and will consider the development of a tire
model to serve as their theoretical basis.
The suspension plays an important role in

maximizing the performance potential of the tires. In
Hondas third Formula One era, suspension development
was conducted making maximal use of techniques
fostered in the development of mass production vehicles.
In particular, specifications were set for the scrub radius,
caster trails, and other aspects of king pin geometry
using identical concepts to those employed in mass
production vehicles, and their usefulness was verified.
Elements specific to Formula One are a tire
characteristic in which the tires perform well under an
extremely narrow range of conditions, and an
aerodynamic characteristic in which vehicle behavior that
is determined by the suspension plays an important role.
Based on these considerations, development was
conducted as follows:
(1) Optimization of initial camber, camber gain, and camber
change with steering angle in accordance with tire
characteristics
(2) Development of tire air temperature control technique
for stabilization of tire pressure
(3) Development of load transfer control for adjustment of
mechanical balance, in order to enhance tire warm-up
performance
(4) Development of geometry enabling maximization of
aerodynamic performance
As one example aspect of suspension development,
this paper will discuss the development of a front
pushrod on upright (FPROU) suspension that is designed

163
to maximize tire performance by means of control of

wheel load for all four wheels.
2.1. Overview of FPROU
The Formula One steering characteristic basically
uses understeer (US) to increase stability during highspeed cornering and oversteer (OS) to enhance turn-in
performance during low-speed cornering. Suspension
design exploited the nonlinearity of the angles of
rotation of the rocker arms in relation to the vertical
motion of the pushrods to induce a high level of
nonlinearity in the wheel rate of the front wheels (a
rising rate). This increased front roll stiffness during
high-speed cornering, when a strong downforce is
acting, thus enabling use of US.
During Hondas third Formula One era, the company
was successful in enhancing turnability during low-speed
cornering by mounting the front pushrods on the uprights
rather than the lower wishbones in order to produce a
characteristic that varied the mechanical balance (the
front-rear allocation of lateral load transfer during
turning) in relation to the steering angle (Fig. 1).
2.2. Load Transfer Mechanism during Turning
When the vehicle is statically steered, the front inner
wheel will normally be subjected to a downward load,
due chiefly to the effect of the angle of the casters.
Because the motion in the front roll direction that this
induces is restricted by the rear suspension, the wheel
loads for the four wheels are transferred diagonally. It
is known that because the load transfer produced by this
geometric motion is added to the load transfer produced
by the front-rear allocation of roll stiffness and the antiforce geometry when the vehicle is turning, load transfer
can be controlled by controlling the trajectory of contact
patch lift when the steering wheel is turned. In
conventional suspensions, increasing the caster angle will
reduce the load transfer at the front of the vehicle.
This section will focus on the inner turning wheels
in order to consider the changes produced when the
pushrods are mounted on the uprights. Because there is
virtually no change in damper stroke when the steering
wheel is turned, pivot N of the pushrod on the rocker
side (Fig. 1) can be considered to be fixed. This means
that pivot M of the pushrod on the upright side is
confined on the sphere surface with pivot N at its center.
When pivot M is positioned to the rear of the kingpin
(Fig. 2), pivot M rotates around the axis of the kingpin

and shifts towards the center of the vehicle when the
steering wheel is turned. Because the pushrod is at an
anhedral angle, it is necessary for the upright to shift in
a downward direction. In the same way, if the outside
wheel is considered, the upright shifts in an upward
direction.
Figure 3 shows the amount of contact patch lift
against toe angle. When the M pivot is offset to the rear,
the same effect can be obtained as when the caster angle
is increased. As a result, there is a comparative reduction
in front load transfer when the vehicle is turning. By
contrast, when the M pivot is offset to the front, the
same effect can be obtained as when the caster angle is
reduced.
The exploitation of the mechanism described above
enabled control of the mechanical balance that is
dependent on toe angle without restriction by geometric
considerations such as the position of the kingpin axis,
and played an important role in helping to inhibit US
during low-speed cornering.
In addition, when the M pivot was offset in the leftright direction (y) using the same mechanism, the upright
rose together with the inner and outer turning wheels
when the steering wheel was turned, lowering the height
of the vehicle (Fig. 4). This effect was particularly
marked in the large steering angle range.
2.3. Dynamic Performance Simulation
Figure 5 shows the results of prediction of changes
in mechanical balance in a vehicle motion simulation
using ADAMS. When a positive offset is introduced, i.e.,
x offset
Fig. 2
Front pushrod on upright suspension
Definition of pushrod offset
Standard
FPROU 10 mm rearward
FPROU 10 mm forward
-5
-10
Fig. 3
164
y offset
10
-10
-20
Pivot M
Fig. 1
Pivot M
Contact patch lift [mm]
Pivot N
Forward
0
Toe angle [deg]
10
Change in lift against toe angle
20
when the M pivot is offset to the rear of the kingpin

axis, the mechanical balance shifts to the rear, and the
steering characteristic tends towards OS. The degree of
change in the mechanical balance increases in proportion
to the level of offset.
2.4. Track Tests and Application in Races
Track tests were conducted on the suspension system
designed and manufactured on the basis of the
simulations in September 2005 on the Jerez circuit. The
tests demonstrated the superiority of the new suspension
system, as it matched the targets set for it in helping to
inhibit US during low-speed cornering and consistently
bettering lap times against the base vehicle. Based on
these results, the suspension system was employed in
races from the Japan Grand Prix in 2005 onwards.
10
differential motion of the differential gears (ultimately,

spools were fitted to mechanically connect the left and
right wheels). Due to this increased US during turn-in,
the mechanical balance was setup to shift towards the
rear wheels in order to enhance turn-in performance.
However, this increased the load on the outer rear wheels
during cornering, generating sudden snap OS resulting
in the rear wheels slipping when the vehicle accelerated
coming out of a corner. This made the vehicles
cornering traction performance fall behind that of other
vehicles (Fig. 6).
In addition, the increased load on the rear tires
brought about a degradation of the tire compound, and
stability and traction performance declined with each lap.
In addition, the setup produced a vicious circle in which
a decline in cornering speed also reduced the temperature
of the front tires, further increasing US and thus resulting
in a decline in dynamic performance.
Contact patch lift [mm]
Standard
FPROU 50 mm inside
-5
-10
-20
Change in mechanical balance [%]
Fig. 4
-10
0
Toe angle [deg]
10
20
Change in lift against toe angle (y-direction offset)

2
1
0
-1
-2
-3
3.1. Method of Formulation of Setup for RA106

The main issue with regard to the setup of Hondas
2005 Formula One vehicle was that engineers had
become process-focused and shortsighted in their quest
for local optimal solutions, resulting in the inability to
exploit the potential for the dynamic performance that
should be expected from the vehicle.
Accordingly, the formulation of a concept to enable
vehicle setup to be conducted in a strategic fashion was
focused on in the development of the RA106. This meant
a rigorous process in which the performance elements
that should be enhanced in setup were clarified, methods
of achieving these targets were quantitatively analyzed,
effects and levels of sensitivity were predicted, taking
in pros and cons, and these predictions were verified in
track tests. The process aimed, by establishing the
orientation for the setup on the desk and using track tests
for verification, to avoid falling into the trap of localized
optimization as had been the case when the setup was
successively altered on the basis of the results of
individual tests.
-4
-5
-10
-5
5
10
x offset [mm]
15
20
300
Apex
Car speed [km/h]

100
Fig. 5
Change in mechanical balance with x-direction

offset
2000
10
Slip ratio [%]
2200
0
-10
2000
3. Concept of Vehicle Setup
2100
2300
RL (outside)
RR (inside)
2100
2200
2300
2100
2200
2300
6000
This chapter will focus on vehicle setup, and will

consider the concept for the setup of the 2006 RA106
vehicle, which became the basis for later vehicle setup.
Having taken second place in the 2004 World
Constructors Championship, Honda commenced the
2005 race season determined to win the title, but the
2005 vehicle displayed low braking stability. The drivers
demanded increased stability during braking, and the
hydraulic control was modified to greatly inhibit the
Tire vertical
load [N]
0
2000
5000
U/S
Tire longitudinal
force [N]
165
O/S
-5000
2000
2100
2200
Distance [m]
Fig. 6
Cornering conditions in 2005
2300
Delta of M/B [%]

-50
10
-40
-30
-20
-10
10
20
30
40
50
-5
-10
Delta of CoP
Delta of W/D
At low-speed corner
-15
-5
-4
-3
-2
-1
Delta of M/B
Delta of CoP, Delta of W/D [%]
Fig. 8
Traction change with setup sweep

Cornering speed dependence
10
-2
Delta of CoP [%]
3.3. Methods of Achieving Targets

A simulation using a tire model was employed in an
analysis to enable engineers to determine how to achieve
the targets listed above using the following elements of
vehicle setup, which determine dynamic performance:
Front axle weight distribution (W/D)
Front axle downforce distribution (CoP)
Front axle mechanical balance (M/B)
Lateral distribution of rear braking and driving force (Diff)
Figure 7 shows the results of an analysis of changes
in traction performance when cornering in each speed
range. The forward shift of M/B has the greatest effect
in enhancing traction performance during low-speed
cornering, with W/D having the next greatest effect. The
CoP makes a high contribution in the high-speed range.
Figure 8 shows the results of an analysis of changes in
traction performance when each of the elements of the
setup is varied. In the case of M/B, traction performance
increases virtually linearly up to a shift of approximately
25% forward. Because increased traction performance
from low speeds upwards would make a significant
contribution to enhanced lap times, forward M/B was
made the main element in the achievement of enhanced
traction performance.
With forward M/B, the overall steering balance
would become US. Figure 9 shows estimates of the
amount of adjustment of W/D and CoP necessary to
achieving this balance. Despite the fact that sensitivity
Difference of traction force [%]
3.2. Concept of Setup of RA106

The dynamic performance of Hondas 2005 Formula
One vehicle was analyzed, and the following
performance elements were selected as targets for the
RA106, based on considerations of maximizing tire
performance by optimizing the total setup of the vehicle:
(1) Enhancement of cornering traction performance
(2) Reduction of degradation of rear tires
(3) Achievement of increased grip force by increasing
temperature of front tires
differs with cornering speed, each parameter displayed

a largely linear response in relation to the degree of
change in M/B. Figure 10 shows the results of an
analysis of the effect of the lateral distribution of
longitudinal forces on the rear wheels produced by the
differential gear on the steering balance. During lowspeed cornering, the US generated by forward M/B
could be compensated by differential control, but this
did not have sufficient effect during high-speed
cornering, and it would therefore be necessary to
compensate for US in combination with another
element of the setup. It was judged that the
combination of forward M/B and rearward W/D would
contribute to enhanced traction performance, and this
was therefore selected as the best strategy.
Delta of W/D [%]
In addition, setting milestones in the formulation of

the setup and designing and implementing track test
plans on this basis helped to enable a reliable verification
process.
-4
-6
Low
Med.
High
-8
-10
0
10
20
Delta of M/B [%]
Fig. 9
Low
Med.
High
6
4
2
0
0
30
10
20
Delta of M/B [%]
30
Relationships in setups for achievement of

equivalent steer balance
Delta of M/B +10[%]

Delta of W/D -1[%]
Delta of CoP -1[%]
3.0
Difference of yawing moment [Nm]
Difference of traction force [%]
3.5
2.5
2.0
1.5
1.0
0.5
0.0
Low
Med.
Turn-in braking
2500
1500
1000
500
0
-500
-1000
-1500
-2000
Delta of M/B +25[%]

Delta of W/D -2.7[%]
Free differential gear
US
-2500
High
Low
Med.
High
Cornering speed
Cornering speed
Fig. 7
OS
2000
Traction enhancement with setup changes
Fig. 10
166
Effect of differential gear on steer balance
3.4. Verification of Effects

The effects of the RA106 setup that had been
formulated on the desk were verified in track tests
conducted on a variety of courses from December 2005
to February 2006.
Figure 12 shows a comparison of traction force
during low-speed cornering in track tests conducted on
the Jerez circuit. For this test, the M/B of the RA106
was increased by 6.4% and its W/D reduced by 1.5%
against the 2005 vehicle in the setup. As a result, the
average traction force of the RA106, averaged over 10
laps, increased by approximately 10%. Figure 13 shows
a comparison of section time for cornering traction

during low-speed cornering, with the RA106 setup
reducing the vehicles average time per corner by 0.05
sec. As a result, the lap time of the RA106 was reduced
by an average of 0.74 sec per lap (Fig. 14). These
results indicated that the cornering traction performance
of the RA106 had been enhanced.
Figure 15 shows the results of an analysis of tire
degradation. Data for which track conditions and
atmospheric conditions could be regarded as identical
was isolated from data collected in multiple track tests
of the 2005 vehicle and the RA106 conducted for the
same periods of time. Lap times for each lap were
averaged from this data. These average lap times were
taken to be representative values for change in lap time,
and the degree of degradation in lap times was compared
for the vehicles. Taking into consideration the difference
in the weight of the vehicles, a simulation was used to
calculate the sensitivity of lap time to vehicle weight,
T2
T14
2
Section time [s]
Figure 11 shows the results of an analysis of the

change in tire surface temperature during cornering when
the M/B was varied. It was estimated that forward M/B
could cause the temperature of the front outside tire,
which is dominant during cornering, to increase, and that
the tires grip force would increase as its temperature
approached the necessary temperature range. In addition,
because the temperature of the rear outside tire would
decrease, this method could also be expected to control
the degradation of the tire compounds due to excessive
temperatures.
Based on the results of the analyses discussed above,
the proposed setup for the RA106 was formulated as
follows:
(1) Forward mechanical balance
(2) Rearward weight distribution
(3) Differential gear control during cornering
2005 Car
1.9
1.9
RA106
1.8
1.8
1.7
1.7
1.6
1.6
1
10
Timed lap
Fig. 13
10
Timed lap
Section time during corner acceleration
85.0
2.0
84.0
1.5
1.0
Lap time [s]
Difference of
tire surface temperature [C]
2.5
Front outside
Front inside
Rear outside
Rear inside
0.5
0.0
-0.5
-1.0
83.0
82.0
2005 Car
81.0
RA106
-1.5
-2.0
80.0
-2.5
-40
-20
20
40
79.0
1
Delta of M/B [%]
Fig. 11
5
6
Timed lap
10
Estimation of change in tire surface temperature

Fig. 14
Difference of lap time average
from 1 T/L [s]
2005 Car
4000
Comparison of lap time
4.0
5000
Traction force [N]
RA106
3000
2000
1000
3.0
2.0
2005 Car
1.0
RA106
0.0
T2
T6
T11
T14
10
Timed lap
Turn
Fig. 15
Fig. 12
Comparison of traction force
167
Analysis of tire degradation from perspective

of lap time
and the lap times were converted to represent vehicles

of the same weight. The results of the comparison
showed that the level of deterioration in lap time was
reduced by an average of approximately 0.48 sec for the
RA106, indicating that the degradation of the tires had
been controlled.
Figure 16 shows the results of a comparison of front
tire temperature conducted in the same way, using data
from track tests with largely identical track and
atmospheric conditions. The tests were conducted on the
Jerez circuit, which features a large number of right
turns. The temperature of the left front tire, which more
frequently worked as the outer tire, increased by an
average of approximately 16C, while the temperature of
the right front tire increased by approximately 5C.
The results of these analyses confirmed that the
RA106 setup had performed as speculated, increasing
cornering traction performance, controlling tire
degradation, and increasing grip force by raising the
temperature of the front tires.
3.5. Summary of Concept of Vehicle Setup
The elements of the RA106 setup which would lead
to performance enhancements were determined through
analysis of the 2005 vehicle, and a simulation employing
a tire model was used to study a setup that would
actualize these elements. The effects of the setup were
verified in track tests, confirming that the setup accorded
with the goals established for it. Following this
verification of the effects of the RA106 setup in tests
conducted during winter, the setup was used as the base
for all vehicle setups from 2006, and contributed to
Hondas victory in the 2006 Hungary Grand Prix.
While vehicle setup had previously relied to a great
extent on comments from drivers and the experience of
the engineers involved, these results demonstrated that
a logical approach, based on a tire model, was effective.
The necessity for the development of a more
sophisticated tire model capable of predicting
performance under all conditions therefore increased.
FR
Tire surface temperature [C]
FL
60
60
55
55
50
50
45
45
40
40
2005 Car
35
35
RA106
30
30
1
3
4
Timed Lap
Fig. 16
3
4
Timed Lap
Front tire surface temperature
4. Development of Tire Model with Coupled

Calculation of Force and Temperature
Tire performance has a considerable effect on the
dynamic performance of Formula One vehicles. The
achievement of enhanced dynamic performance
necessitates a quantitative understanding of tire
performance, and vehicle design and setup that maximize

tire performance under all conditions. Regulations
stipulate that the setup be performed by Saturday
morning in race week, and it is therefore necessary to
perform the setup with consideration of factors including
changes in environmental conditions during the
qualifying sessions on Saturday afternoon and the finals
on Sunday. This necessitates a model capable of
calculating at a sufficient rate of accuracy the force
generated by the tires under a variety of driving
conditions, including load, alignment, and track surface
temperature. The performance of Formula One vehicle
tires is in particular affected by the surface temperature
and the internal temperature of the tires, and this
necessitates the use of a model that considers the effect
of temperature, and is able to perform coupled
calculations of force and temperature in response to
changes in tire input due to environmental temperature
and driving conditions.
The ability to simultaneously analyze tire force and
temperature would represent a significant advantage not
only to considerations of vehicle setup and tire
performance, but also to considerations including race
strategy, such as warm-up performance when a safety car
is on the track, and degradation due to heat.
The Magic Formula is one well-known tire model,
but this model does not include a temperature element,
making it unable to calculate the effect of temperature
changes resulting from driving conditions on tire force.
Tire suppliers also provide tire models, but their insides
are concealed, in practice inhibiting any increase in the
accuracy of the models or modification for use by the
race team.
It was judged that in order to excel over other teams
in the area of vehicle dynamic performance, it would be
necessary to possess, as an in-house developed
technology, a tire model able to perform coupled
calculations of force and temperature, and to use this tire
model as the basis for a technology enabling prediction
and analysis of dynamic performance.
4.1. Structure of the Model
4.1.1. Model concept
Considering the use of the model in vehicle setup at
the circuit in addition to vehicle design, the development
of a model with a good balance between calculation
accuracy and speed was established as a target. The
elements with the greatest effect on the accuracy of
calculations of tire forces were isolated and modeled to
enable this target to be met.
Tire forces are chiefly determined by the structural
deformation of the tire and the friction characteristic
between the road surface and the tire. The friction
coefficient of the tires is changed significantly by
environmental conditions and driving conditions,
including the road surface roughness, dust on the road
surface, tire surface temperature, tire slip speed, rubber
wear, and thermal degradation. Accordingly,
characteristics such as structural deformation and heat
transfer, which could be modeled on the basis of
168
theoretical concepts and the results of bench tests, were

separated in the models structure from elements such
as the friction coefficient, which are dependent on actual
driving conditions, with the parameters for the latter
being identified from actual vehicle data. This helped to
enable the creation of a more realistic and more accurate
model.
4.1.2. Force model
The model treated deformation of the tire contact
patch as divided into a belt section and a tread rubber
section.
Belt deformation was approximated by expression as
a quadratic function in relation to the position of the tire
contact patch in the longitudinal direction. The
deformation obtained in this manner was corrected using
the tire side force, self-aligning torque, internal pressure,
and longitudinal force.
With regard to the deformation of the tread rubber
section, the adhesive contact area (the area of elastic
deformation) of the contact patch was calculated from
the relationship between the maximum deformation of
the tread rubber, as determined by the elastic modulus,
the static friction coefficient, and the contact surface
pressure, as well as the necessary deformation, as
determined by the tire slip angle, slip ratio, and degree
of belt deformation. The rest of the contact patch was
considered a sliding contact area (Fig. 17).
The forces on the tire are determined in the adhesive
contact area by the elastic modulus of the rubber and the
degree of deformation, and in the sliding contact area by
the slip friction coefficient and load. At this stage, a
model was formulated in which the elastic modulus was
defined as bulk temperature functions (bulk temperature
will be discussed below), the static friction coefficient
as surface temperature functions and contact surface
pressure functions, which is to be corrected based on its
distribution in the direction of the tread. The slip friction
coefficient was further defined as tire slip speed
functions in the model.
Using this method, the tire force in the adhesive and
sliding contact areas were calculated, and their synthesis
enabled calculation of the final tire force.
4.1.3. Thermal (temperature) model
The cross-sectional structure of the tire was
formularized as a one dimensional node model divided
into three layers, a tread surface layer, tread bulk rubber
layer, and carcass/tread belt rubber layer, with each layer
at a single temperature.
Sliding
contact area
In addition, the work performed by the tire was input

as heat chiefly in the tread surface layer for the sliding
contact area and in the tread bulk rubber layer for the
adhesive contact area. The rolling resistance of the tire
was treated as heat generated in the tread bulk rubber
layer.
Heat transfer between the three layers, convection
heat transfer to the air, and heat conduction to the road
surface at the contact area are all calculated to enable
calculation of the temperature of each layer. Heat
transfer by radiation was considered to have represented
2% or less of the total figure in comparison to heat
convection and conduction, and was therefore omitted
from the model (Fig. 18).
4.2. Identification of Parameters
4.2.1. Method employed
The difference between the calculation results for
force and moment from the model, obtained using time
series data from track tests for parameters including
wheel load, slip angle, slip ratio, and camber angle, and
the target data were treated as objective functions, and
parameters were identified using an optimization method
to minimize this difference. Parameters that changed with
the degradation of the tires, such as maximum friction
coefficient, were identified for each lap, while
parameters that did not change, such as the temperature
characteristic, were identified as a single value for all
the track test data. Measurement results from sensors and
the results of calculations using a combination of sensor
results and bench test results or theoretical values (in the
case of parameters including wheel load and slip angle)
were used as the data necessary for parameter
identification.
The following two methods of identification were
employed, depending on the target.
(1) Method using the results of 6-component wheel force
measurements
The results of 6-force measurements (longitudinal
force, side force, self-aligning torque) were used as
targets. Because the accuracy of identification was
increased and the level of variation reduced the greater
the diversity of combinations of parameters such as
wheel load and alignment, data for both the left and right
tires were employed in the parameter identification.
Tyre cross section
Thermal Input
Static
contact area
y
Tire workload
on contact area
Center line of tread rubber
x
Center line of tread belt
Fig. 17
Thermal transfer
to ground or air
Tread surface layer

Tread bulk rubber
Tread belt
Tire contact patch
Tread rubber
Thermal transfer
between layers
Tire rolling resistance

Carcass / Tread belt rubber
Side slip angle
Deformation model for tire contact patch
Fig. 18
169
Tire thermal model
(2) Method using vehicles dynamic states

Vehicles dynamic states (longitudinal acceleration,
lateral acceleration, and yaw moment) measured using
sensors were employed as targets. The vehicles dynamic
states were calculated using a variety of data including
calculation results from the tire model, vehicle
specifications, and figures for air resistance and tire
rolling resistance, and parameters were identified
simultaneously for the front and rear tires that enabled
these dynamic states to be matched with the target
dynamic states.
4.2.2. Optimization method
Broadly speaking, two types of optimization methods
are available. Gradient methods use the gradient of the
objective function (the error in relation to the target)
against the design variables (the optimization parameters)
to enable optimization. In stochastic methods, design
variables are varied at a specific probability to produce
a variety of figures that are potential optimum solutions,
and the optimum solution is selected from among this
group. The project discussed here used an evolution
strategy, which is a type of stochastic method.
4.3. Results of Model Verification
Because results are affected by the measurement
accuracy of the data used in the model calculations and
the target data for comparison, it can be challenging in
the verification of a model that uses actual vehicle data
to determine whether a specific issue originates in the
model or in the data. In addition, because the actual
values of the tire parameters are unknown, quantitative
verification of the results of parameter identification also
represents a challenge.
Artificial target data was therefore created by
constructing a vehicle model from body, suspension,
steering, and aerodynamics models and the tire model,
conducting a one-lap circuit simulation, and by adding
noise corresponding to actual driving conditions to the
obtained vehicle dynamics data. This artificial data was
then used to verify the model. This method enabled
issues of measurement accuracy to be set aside, and
because the tire parameters that should be obtained as a
result of parameter identification were already known,
it was possible to verify the tire model and its parameter
identification section.
Figure 19 shows a time series comparison of the
average error in vehicles dynamic states (longitudinal
force, lateral force, yaw moment) for one lap in the
target data and the model, using a specific identification

calculation. Because the evolution strategy is a type of
stochastic method, the identification calculations have
been conducted 10 times for the purposes of comparison.
Error for all the results is approximately 1.2%, and as
the time series graph shows, sufficient accuracy was
obtained for the analysis of vehicle dynamics. However,
despite the fact that the figure for error in vehicles
dynamic states was largely identical in both sets of
results, a phenomenon could be observed in which
synchronized variations in multiple parameters were
observed. The combination of different parameters to
produce an identical output from the model is termed a
modal characteristic, and is a phenomenon that is often
observed in multi-parameter models. Increasing the
diversity of the target data is an effective method of
controlling the modal characteristic in analysis results,
and this study provided insights into the degree of
diversity necessary.
Based on the verification of the accuracy of the
model described above, results for tire force and
temperature were compared with results from track tests.
Figure 20 shows a comparison of measurements
using a tire force meter and calculation results from the
tire model for the fifth lap of five laps around the
Barcelona circuit. The parameters for the tire model were
identified using data from the third lap. Side force is
calculated accurately across the entire range. Calculation
results for the peak value of longitudinal force during
braking are somewhat low, but results for turn-in
correspond well. The slip ratio, one of the input
parameters for the model, changes rapidly when the
vehicle is braking, and it is therefore not easy to be
confident of its accuracy. The effect of error due to input
data accuracy was therefore considered to be greater in
producing this result than the effect of the accuracy of
the model.
Figure 21 shows a comparison of measurements and
model calculation results for the temperature of the tire
surface and bulk layers. During the period shown, the
vehicle makes a pit stop (t=220 sec), and then reenters
the track (t=280 sec). The model was able to reproduce
the drop in the internal temperature of the tire due to
the cessation of internal heat generation and the increase
in the temperature of the tire surface due to the decline
in heat dissipation into the air after stopping, in addition
8000
6000
4000
Force [N]
Car lateral force
Car longitudinal force
Force [N]
Force [N]
Target
Identification
Fig. 19
0
-2000
Measurement of tire longitudinal force
Measurement of tire lateral force
Measurement of tire vertical load
Calculation of tire longitudinal force
Calculation of tire lateral force
-4000
-6000
Target
Identification
Time [s]
2000
-8000
400
Time [s]
410
420
430
440
450
460
470
Time [s]
Identification result (car dynamics error)
Fig. 20
170
Comparison of tire forces (front)
to the gradual increase in temperature until a state of

thermal equilibrium is reached after the vehicle starts
moving again. The model calculated the internal
temperature of the tire at a somewhat high level during
lap driving, but the level of accuracy can be considered
sufficient for evaluating warm-up performance.
4.4. Summary of Tire Model Development
The project discussed in this paper developed a tire
model that is able to perform coupled calculations of tire
force and tire temperature that affects tire force. The use
of a model configuration, in which parameter
identification was performed for elements affected by
driving conditions based on track data, enabled the
realization of a level of calculation accuracy and speed
satisfying a broad range of demands, from vehicle design
to vehicle setup.
120
80
60
40
Car speed
Measurement of tire surface temperature
Measurement of tire bulk temperature
Calculation of tire surface temperature
Calculation of tire bulk temperature
20
250
300
350
400
450
500
550
600
650
700
400
200
0
750
Car speed [km/h]
Temperature [C]
100
Time (s)
Fig. 21
Comparison of tire temperatures (rear)
5. Conclusion
As demonstrated by the technology introduced in this
paper, vehicle dynamics simulation technologies play a
significant role in developments relating to the dynamic
performance of Formula One vehicles. In addition to the
benefits of simulations in reducing costs and increasing
development efficiency, the very process of observing
phenomena, analyzing their mechanisms, and formulating
models to develop a simulation technology in itself leads
to dramatic enhancements in dynamic performance.
More recent developments, seeking further
performance increases and enhancements in development
efficiency, are progressing to the development and use
of a driving simulator, which represents an evolution of
vehicle dynamics simulation technology. The removal of
the various external factors associated with track tests on
a circuit enabled high quality test results to be obtained.
The driving simulator was applied not only in evaluating
the performance of development items, but also in the
formulation of vehicle development concepts and the
creation of indices for drive feeling in relation to
dynamic performance. It is expected that these
technologies will be actively employed in the
development of mass production vehicles.
Author
Hideaki SHIBUE
171
Kazuhiro TANEDA
Yasutaka KITAKI
Weight Reduction and Stiffness Enhancement

Technology in Formula One Chassis Development
Hajime WATANABE*
Terukazu EBISAWA*
Jun IWAI*
Kazuhiro TANEDA*
Naoki HIROMASA*
ABSTRACT
When developing Formula One chassis, reducing weight while still satisfying the functions and requirements of
each part was an issue for chassis design. The authors took into account the degree of contribution to circuit lap time,
and found a balance between functions and weight reduction during chassis development. The method used consisted
of reviewing materials and structures, and the weight of chassis parts was successfully reduced.
This paper introduces various examples of weight reduction.
1. Introduction
introduced to circuit tests. Finally, parts for which safety

has been confirmed in the running state in track tests are
introduced to actual races.
In this manner, much effort was given to maintaining
safety and quality, even though one characteristic of
Formula One chassis development is that time resources
are always limited. Normal annual development consists
mainly of parts development that aims to reduce weight
by a few grams while securing the same stiffness as the
previous model, or parts development that aims to
enhance stiffness by a few percent while maintaining the
same weight.
However, Honda aimed for weight reduction that far
exceeded the bounds of this normal development, by
introducing new materials and structures. In addition,
emphasis was placed on reducing the weight of unsprung
parts and parts located far from the center of gravity or
in high positions, taking into account the map shown in
Fig. 1 and other factors, from the standpoint of reducing
In chassis development, reducing weight and

enhancing stiffness while satisfying the functions and
requirements of each part is an issue that automobile
design engineers are constantly working at in their daily
development work, and Formula One chassis
development is no exception.
Most racing car parts are related with the functions
of accelerating, cornering, and stopping. Therefore,
development placed particular attention on safety, and
care was taken to develop parts that would not become
factors resulting in retirement from a race.
In particular, parts that are directly linked to safety
are given a Class A rank, and stress analysis is
performed in the design stage to check that the
prescribed safety factor is met before design release and
manufacturing. Only the parts of which strength and
durability have been confirmed by bench tests, can be
Rear suspension
Radiator
Rear wing
Airbox tray
Front suspension
Driver Monocoque
Nosebox
Front wing
Brake caliper
Wheel
Exhaust
Engine
Ballast
CoG
Fig. 1
CG_x versus CG_z plot for each component

172
Wheel
Brake caliper
the moment of inertia. Furthermore, the contribution of

chassis stiffness to performance was considered using
evaluation standards that differed from those
conventionally used for Formula One, and development
also attempted to enhance the stiffness of key chassis
locations including the engine.
2. Weight Reduction and Stiffness

Enhancement of CFRP Parts
CFRP composites have high specific strength and
specific stiffness, and are widely utilized in current
Formula One chassis. Approximately 80% of chassis
parts are made of this material, and most body parts in
particular are made of CFRP composites.
Efforts to reduce weight and enhance stiffness in the
body area focused almost exclusively to make the most
out of the characteristics of CFRP composites. This was
done by development of material or by applying new
construction to the parts using small-lot production
technology.
Weight reduction was investigated focusing on parts
located far from the center of gravity, that is to say the
nosebox, rear impact structure, and rollhoop, from the
standpoint of reducing the moment of inertia. These parts
are for crash safety, and must absorb collision energy by
crumpling themselves. That requirement is stipulated in
the regulations prescribed by FIA in such terms as the
amount of energy absorbed, initial velocity, and rate of
deceleration. We had to meet this collision requirement
and reduce the weight at the same time while keeping
the aerodynamically required shape and size, durability
with respect to force inputs during circuit running,
among other requirements. To realize this, a structure
designed to increase the amount of collision energy
absorption and the management of the manner how the
part crumples during a collision was required.
CFRP parts are created by manufacturing a master
model from a block of easily-machinable resin, and then
producing a reverse mold by laminating pre-impregnated
composites (prepreg), which are used to make molds,
over this master model. The product materials are then
laminated inside this mold, and formed in an autoclave.
The manufacturing process involves a lot of manual
work, so complex shapes and structures can be created.
However, being able to manufacture parts easily in
a short time means that more time can be allotted to
parts design, and more proposals can be investigated and
tested. This can make a decisive difference when
launching a new chassis in the beginning of each season,
or when implementing a large-scale update partway
through a season. Therefore, shortening the part
manufacturing time is also an issue when aiming to
reduce weight using new structures.
Typical examples of developing CFRP parts are
described below.
Fig. 2 (a) to secure shape stiffness with respect to front

collisions. However, this meant that the energy
absorption ratio per unit weight was small.
Therefore, the honeycomb was eliminated and a
pillared structure with pipe sections in each corner was
used as shown in Fig. 2 (b). This increased the energy
absorption ratio while maintaining the shape stiffness,
and achieved a weight reduction of 13% compared to the
conventional nosebox.
In terms of manufacturing technology as well,
eliminating the honeycomb enabled one-piece molding,
which reduced the process time by 30% compared to the
conventional nosebox. However, establishing predictive
design technology took time, so parts with the
conventional structure had to be made before
manufacturing parts with the new structure. Changes to
the external shape occurred repeatedly before all the
parts with the new structure were completed, and
therefore, these parts could not be applied to races.
2.2. Development of the Rear Impact Structure
The rear impact structure is a part that had a high
energy absorption ratio from the beginning. However,
when the external shape was stipulated by Formula One
regulations, an even higher energy absorption ratio was
required to prevent an increase in the weight.
CFRP absorbs energy by crumpling, so the more
fractured the structure after a collision, the higher the
energy absorption ratio. Focusing on this characteristic,
a thin-walled structure with a window frame shaped
cross-section [Fig. 3 (b)] was used instead of the
conventional thick-walled structure [Fig. 3 (a)], to
2.1. Development of the Nosebox

The nosebox formerly used a sandwich structure
(CFRP, aluminum honeycomb, CFRP) such as shown in
173
(a) Conventional structure
Nosebox
(b) Four pillared structure
Fig. 2
Nosebox
(a) Conventional
cross section
(b) Window frame shaped

cross section
Fig. 3
Rear impact structure
achieve further fracturing of the components in the event

of a collision.
This structure was created by laminating the window
frame shape materials over a silicon mandrel, placing an
outer mold containing the laminated outer shell materials
over, and performing one-piece internal pressure
molding.
The new structure achieved a 28% weight reduction,
and was applied to races while continuing to evolve the
design to more aerodynamic shapes every year.
2.3. Development of the RollHoop
The rollhoop must protect the drivers head in the
event of rollover, and secure a sufficient air intake area
for the engine. In addition, an aerodynamic external
shape with respect to the rear of the chassis is also
demanded.
A pentagonal shape as viewed from the front was
introduced to reduce thickness and weight while still
satisfying the above requirements. Within this structure,
the roof lines of the pentagonal shape were set against
to the load direction as stipulated by Formula One
regulations. A thinner rollhoop that meets the
requirements was successfully developed by increasing
the load resistance of these portions to reduce
deformation (Fig. 4).
This reduced the weight by 18% compared to the
conventional rollhoop, and we were able to apply this
new rollhoop to races. In addition, the shape of the
connection with the monocoque was also reviewed when
switching the monocoque, which achieved another 16%
weight reduction.
As engine development advanced and it became
possible to reduce the aperture area, the next step was
thought to be to increase the sectional area of the
rollhoop to further reduce weight. The new design
received loads with a planer surface, and weight was
reduced by making use of the shape stiffness of the
larger section, using a lightweight core (Rohacell)(Fig.
5). Prospects for lamination were obtained through
simple investigation using past results, and the maximum
load was predicted by fracture propagation analysis using
a nonlinear analysis tool. In this way, the new rollhoop
was designed. A structure that does not fail even when
subjected to the regulation load was achieved by
efficiently distributing the load over the entire structure.
This realized a weight reduction of approximately
20%, and the new rollhoop was applied to races.
2.4. Development of the Engine Cover

The weight of exterior cover parts such as the
engine cover (Fig. 6) must be reduced to the greatest
extent possible while realizing aerodynamic functions.
Therefore, as long as the part stiffness is kept by the
shape, a sandwich structure with the minimum layers
is used.
Weight was further reduced by making the CFRP
materials that comprise the inner and outer shells of this
sandwich structure as thin as possible. This was achieved
by using a new material developed jointly with the
material supplier, and by applying lightweight resins.
This new material was an open fiber material made by
spreading carbon fiber yarn into a flat, thin layer, making
the fibers into a plain fabric, and then soaking the fabric
in resin to create a prepreg. This realized a molded
thickness half or less than that of normal materials.
As a result, the engine cover weight was successfully
reduced by 60% in the qualifying specifications. To
achieve durability, applied areas are modified, and the
part with open fiber material was put in to the race
specifications.
This material was also used to laminate various other
external cover parts, and contributed greatly to reducing
the weight of non-structural chassis parts.
2.5. Development of the Monocoque
Clarifying how chassis stiffness contributes to vehicle
dynamics performance was a major issue. An attempt
was made to enhance vehicle dynamics performance by
increasing the stiffness of the joint between the power
train and the chassis, and by using the airbox mounted
on top of the engine as a stiffener, but an increase in
performance could not be confirmed.
Rohacell core
Fig. 5
Rollhoop structure
Distributed load
Fig. 4
Fig. 6
Pentagonal rollhoop
174
Engine cover
Therefore, it was hypothesized that the degree of

contribution of stiffness to vehicle dynamics performance
is influenced more by the local stiffness of the
suspension mounts than by the overall stiffness, and the
monocoque weight was reduced with the sole restriction
of equivalent local stiffness.
Figure 7 shows the weight reduction methods.
- Reinforcing materials were eliminated by allowing
a 20% drop in the overall torsional stiffness.
- Exclusively developed side panel materials were
applied.
The specifications were investigated using the local
stiffness evaluation method employed in production car
development, and the suspension mounts were reinforced
as shown in Fig. 8. This showed that the weight could
be reduced by 6.5% and the center of gravity lowered
by 2 mm compared to the conventional specifications.
Track tests were performed to compare the new
specifications with the conventional specifications.
The results confirmed that even when the overall
torsional stiffness is lowered, as long as the
suspension mount stiffness can be secured, there is
almost no effect on steering stability. Therefore, a
monocoque that reflects these specifications was
applied to races the following year.
Reinforcing plies eliminated,
Parts downsized
included maximizing tire performance by optimizing the

suspension geometry, controlling ground contact load by
tuning the spring and damper systems, optimizing the
power steering assist characteristics, optimizing the
steering variable gear ratio, developing brake cooling
technology that uses dynamos, reducing fuel sloshing,
and so on.
On the other hand, the Formula One car weight is
prescribed, so it was also important to reduce part
weights to help lower the center of gravity and increase
the freedom of longitudinal weight distribution using
ballasts. The authors constantly set goals to promote the
simultaneous achievement of both performance and
reduced weight.
Various examples of developed parts are described
below.
3.1. Development of Suspension Parts
3.1.1. Development of anti-roll bars
Figures 9 and 10 show schematic diagrams of the
suspension systems. The anti-roll bar is a spring that acts
only during rolling, and is a key part that controls the
rolling angle when cornering, and also determines the
longitudinal balance of rolling stiffness. In the
conventional structure, the rocker arm acts on the torsion
bar via links as shown in Fig. 9, and the mechanism is
such that in the reverse phase (roll), the right and left
Torsion bar spring
3rd damper (spring)

Rocker
Front push rod
Front top wishbone
(FTWB)
Anti-roll bar
Plies on sidepanel
eliminated
Fig. 7
Corner damper
Camber plate
Front track rod
Weight reduction method
Front lower wishbone

(FLWB)
Front upright
Suspension mounting
area reinforcement
Brake caliper
Disc bell
Axle
Brake disc
Wheel nut
Fig. 9
Front suspension system (2003)

Rear anti roll bar
3rd damper (spring)
Rocker
Fig. 8
Rear top wishbone

(RTWB)
Reinforcement method
Rear track rod
3. Weight Reduction and Stiffness

Enhancement of Functional Parts
Weight reduction and stiffness enhancement of
functional parts such as the suspension, steering, axles,
brakes, fuel system, radiators, and exhaust pipes is
described below.
Development of these parts pursued part performance
that enhances vehicle dynamics performance. This
175
Rear lower wishbone

(RLWB)
Rear push rod
Fig. 10
Drive shaft
Rear suspension system (2003)
link generate reaction force in the direction that works

to operate the torsion bar. In contrast to this, the
structure was changed to link the right and left flat
springs with a double end spherical joint as shown in
Fig. 11. In this structure, bending force is generated on
the flat springs, which deform in the primary bending
mode, only in the reverse phase, enabling the same work
as that of the conventional structure. In addition,
simplification of the structure achieved both weight
reduction and enhanced space efficiency.
3.1.2. Development of wheels
The performances demanded of wheels are
lightness in weight to minimize the unspring weight
and the loss of turning inertia energy, and high
stiffness with respect to the input from tire. However,
the ideal balance between weight reduction and high
stiffness was not clear, so development was first
performed by selecting means of stiffness
enhancement by modifying the shape and replacing
materials with a minimum impact on weight.
For each suspension part, the degree of contribution
to camber stiffness was investigated using bench tests
after that, and wheel deformation was found to have an
effect. Increasing the camber stiffness and enhancing the
camber relative to the ground, even at the expense of
weight, was judged to be advantageous for lap time, so
development was performed with the target of increasing
the camber stiffness of the standalone wheels by 80%.
A topology optimization method was used to investigate
Anti-roll blade
Fig. 11
(a) Before optimization
Fig. 12
Front anti-roll blade
the spokes, and the spokes were changed from the shape
shown in Fig. 12 (a) to that shown in Fig. 12 (b). This
achieved a 56% increase in the stiffness to weight ratio,
which reached the target with a weight increase of
approximately 10%. The new wheel design was
confirmed in track tests to enhance the rear lateral force
and rear stability, and was therefore introduced to races.
3.2. Development of Brake Parts
Brake development focused on the issue of how to
achieve a high level of balance between the initial
braking force, braking force stability, and drag reduction.
One method attempted was to reduce the caliper body
weight. Formula One regulations stipulate that the caliper
body material must have a Youngs modulus of 80 GPa
or less, so this essentially limits the material to
aluminum alloy. Among various aluminum alloys, A2099
was selected, which features good durability (high
fatigue strength characteristics), low density, high
stiffness, high specific strength, and heat resistance
strength. Hydraulic stiffness is important for the caliper
body to secure a good pedal feeling, so the center bridge
shape was reviewed in consideration of the stiffness
balance, and unnecessary material was eliminated from
around the cylinder, which succeeded in reducing the
weight by 6% (Fig. 13).
3.3. Development of the Water Radiator
Water radiators and oil radiators are mounted in
Formula One cars. Water radiators radiate the heat from
engine cooling water in the same manner as those in
production cars, and are keeping the temperature ranges
at high engine efficiency. The cooling water is
pressurized to raise the boiling point and to prevent
cavitation. Current Formula One regulations stipulate to
install a pressure relief valve that opens at 3.75 bars, and
this valve is supplied from FIA.
The pressure upper limit is stipulated, and the liquid
side heat radiating area has only a small effect, so
development of the radiators was performed first for the
water radiator side with the goal of weight reduction
(Fig. 14). Development was initially started for the
qualifying specifications so that priority could be given
to weight reduction, even at the expense of durability.
Application to the qualifying specifications started from
2001, and the design was repeatedly modified for race
use thereafter, and introduced to all races from 2006.
Even though the pressure is stipulated, the radiator is still
(b) After optimization
Fig. 13
Wheel spoke topology optimization
176
Brake caliper
subject to approximately four times the pressure in

production cars, so the outside of the stacked tubes is
subject to the greatest deformation, and the stress
increases in this area. In the initial type water radiator,
focus was placed on the heat transfer surface area of the
water side, which has a lower contribution to heat
radiation performance than the heat transfer surface area
of the air cooled side. The inner fins were eliminated,
and the outer frame was utilized to suppress bulging in
each tube due to the water pressure and the force from
the engine. The tube thickness was increased in locations
with local increases in stress, and measures were also
taken such as reducing the fin thickness, while taking
into account pressure resistance. The size was reduced
compared to the original base specifications, and
performance was enhanced by approximately 9%, which
successfully reduced the weight by 37%. However, this
structure had its limits when considering a chassis
package with enhanced aerodynamic performance, so
exclusive tube was developed thereafter. Louver fins
were also employed with the aim of increasing toughness
against chipping, and together with a greater degree of
shape freedom and enhanced durability, achieved further
weight reduction of approximately 7% and enhanced heat
radiation performance of approximately 4.4%. In
addition, boundary layer formation on the water side was
considered, and separate specifications that further
increase heat radiation performance by approximately 8%
was also ready to be applied. Copper, which has a high
thermal conductivity, was also investigated as the
material, but aluminum was used from the standpoint of
weight.
3.4. Development of Titanium Exhaust Pipes
The requirements for exhaust pipe materials are
strength, resistance to oxidation, and fatigue strength,
when subjected to high temperatures, and a low
coefficient of linear thermal expansion. Exhaust does not
have a uniform temperature distribution due to the
exhaust flow, and resonance also occurs at some engine
speeds. Therefore, durability was analyzed and the
exhaust pipes were designed in consideration of these
effects. That said, exhaust reaches a maximum
temperature of 1100 C, so the materials that can be used
are limited, and Inconel is generally used.
A titanium alloy with a low specific weight (density
4.4 g/cm3) approximately half that of Iconnel (density 8.5

g/cm3) was used to develop titanium exhaust pipes with
the aim of a weight reduction of approximately 30%
(Fig. 15). Titanium oxidizes at high temperatures, so
development of an anti-oxidation treatment was a key
element. The anti-oxidation surface treatment applied
was NCC treatment, which consists of first plating parts
with aluminum, and then burying the parts in aluminum
and alumina powder. This surface treatment generates a
uniform Ti-Al alloy layer on the surface of the Ti base
material, and oxygen reacts preferentially with aluminum
at high temperatures, which enabled the strength of the
base material to be maintained. The wear resistance of
joint insertion parts must also be increased, so nickelchromium treatment was applied as surface treatment
according to the location. In addition, the exhaust
temperature distribution differs according to the exhaust
flow and other factors, so also considering the
manufacturing efficiency, various Ti materials were used.
Durability of 600 km or more was confirmed by track
tests, but not all requirements were satisfied, so the
titanium exhaust pipes were not introduced to races.
3.5. Development of Carbon/Carbon (CC) Exhaust
Pipes
Like the titanium exhaust pipes described above,
which were developed mainly with the goal of weight
reduction by material substitution, carbon (CC, density
1.6 g/cm3) exhaust pipes were also developed to further
reduce weight. Normal carbon materials have carbon
fibers in resin matrix. However, CC is a carbon fiber
composite that uses carbon as a matrix. That is to say,
CC exhaust pipes are exhaust pipes made of charcoal.
Development started with a target durability of 25 laps,
aiming for use in qualifying rounds.
The first step was the collector and tail parts, but the
ultimate goal was a complete set of CC exhaust pipes
including the primary, and a weight reduction target of
approximately 39% was set. In addition, conventional
exhaust pipes are made by welding together Iconnel plate
materials. This made it difficult to set a thinner plate
thickness, and manufacturability and durability issues
meant that excess materials were unavoidable. In
contrast, CC exhaust pipes are manufactured using
molds, so the minimum thicknesses can be set regardless
of welding toughness, and the intermediate sections also
Cold roll forming
Precision casting
Fig. 14
Water radiator
Fig. 15
177
Titanium exhaust
Hot stamping
have a high degree of shape freedom. The final

specifications aimed for a compact layout that would be
able to enhance chassis aerodynamic characteristics.
In the first step with only the collector and tail parts
made of CC, gaps opened at high engine speeds between
the CC, which has an extremely low coefficient of linear
thermal expansion, and the Iconnel materials used for the
primary. This produced the issue that the air-fuel ratio
did not stabilize, so a gasket was also developed to fill
these gaps that appear at high temperatures. Pure carbon
has high heat resistance but low oxidation resistance, and
may burn up if measures are not taken, so development
of a surface coating to prevent oxidation was a key
element. Ultimately, resistance to surface oxidation for
3 hours at 900 C was successfully secured using a fourlayer coating, and durability of 30 laps was confirmed
in engine dyno tests, which exceeded the initial target.
Efforts were also made at the same time to shorten the
manufacturing time, but application only to the
qualifying specifications became impossible, so the
development was halted.
reduce the Formula One chassis weight, and achieved a

certain degree of effects. The development directions and
issues in the future was to continue ceaseless
development to reduce weight, focusing on higher
contributions to reducing circuit lap times, and using
ideas conceived during the course of development for
further enhancement of functions.
4. Conclusion
Formula One regulation stipulates a minimum weight
of 605 kg, including the driver. Therefore, the weight
reduction does not contribute to reducing the absolute
weight of the car, but to enhance vehicle dynamics
performance, which means increasing chassis
competitiveness. This is accomplished by replacing the
reduced weight with a ballast to lower the center of
gravity, concentrate the weight in the center of the
chassis, and expand the longitudinal weight distribution
adjustment range.
In addition, detailed regulations also stipulate the
materials allowed to use, so there is a limit to weight
reduction that can be achieved by simple material
replacement, which made us to work on approaches such
as structural changes based on an understanding of the
required functions and performance were needed.
(1) Reducing the weight of parts located above the cars
center of gravity is particularly effective from the
standpoint of lowering the center of gravity. This was
achieved with the rollhoop, engine cover, water radiator,
exhaust pipe, and other parts.
(2) Reducing the weight of parts that overhang from the front
and rear tires is effective from the standpoint of enhancing
turn-in performance (reducing the moment of inertia
around the Z axis (Izz)). This was achieved with the
nosebox, rear impact structure and other parts.
(3) Reducing the weight of the suspension parts and other
unsprung parts is effective from the standpoint of
enhancing vehicle dynamics performance, and targeted
parts such as the brake calipers, anti-roll bar, and wheels.
In addition, not only reducing the weight of these parts,
but also stiffness and function enhancements were
achieved.
(4) The monocoque weight was also reduced from the
standpoint of optimizing the stiffness.
In this manner, the development was performed to
Author
Hajime WATANABE
Jun IWAI
Terukazu EBISAWA
Kazuhiro TANEDA
178
Naoki HIROMASA
Measurement and Analysis Techniques of

Formula One Chassis Development
Hideaki SHIBUE*
Akihiro IDA*
Tomokazu SUZUKI*
Atsushi TSUBOUCHI*
Mamoru URAKI*
Yasutaka KITAKI*
ABSTRACT
Measuring and analyzing chassis data are the basics of Formula One vehicle dynamics development. The acquired
information is used in every stage of development, from design concept to race management. This paper explains
examples of data analysis techniques including competitor performance analysis, dynamic performance indices used
for chassis development and various techniques of measurement systems and sensors used in Formula One.
1. Introduction
The decisive difference of developing a Formula One
car against a passenger car is that the engineers cannot
drive and feel the cars themselves. Therefore, various
high quality onboard data and analysis tools are needed
to accurately understand the vehicle dynamics issues in
order to improve car performances. Some of them are
standardized to maximize tire performance during each
run. These data are obtained through real-time processing
using a telemetry system and being analyzed
automatically so that the engineers can determine car
setup for the next run whilst the car is still on the track.
Simulating tools had been developed as well to predict
and understand the phenomenon more in detail.
This paper introduces examples of the cutting edge
measurement systems and sensors, competitor
performance analysis technique including simulation
tools and the latest analysis techniques of the dynamic
performance indices.
the highest priority, some 60 different types of sensors

with close to 100 channels are mounted on the chassis.
During tests when greater emphasis is placed on data
acquisition, even more sensors are mounted with as
many as 170 channels. In addition, video cameras and
recorders are mounted to record moving images, and
driver operations, racing lines, infrared images of tire
temperatures, air flows, and other images are recorded
in sync with the measurement data according to the
application.
Figure 1 shows an overview of the measurement and
analysis system. The sensor signals are first processed
inside the onboard ECU and recorded in the data logger.
In addition, some performance evaluation indices are also
calculated inside the ECU in real time, and recorded
together with the sensor data. Some of this data is sent
by real-time telemetry to the pit, where it is processed
instantly and continuously on a server (vTagServer:
Sensor
Sensor
2. Measurement Systems
ECU
VTS
Data
logger
Car settings
VAP
Sensor
Sensor
2.1. Overview of Measurement Systems

The tasks of gathering data while running on a circuit
track, analyzing that data, and implementing
countermeasures form an important cycle in Formula
One vehicle dynamics development. This cycle should
be repeated accurately within a short time, so advanced
measurement systems are built into Formula One chassis
from the design stage. In this sense, a Formula One
chassis could itself be considered a measurement
apparatus. Even during races when performance is given
179
Sensor
Simulation
Lap chart
Camera
Radio
Video
logger
HRF
server
HRD/HGT
server
Data viewer
Simulation
Analysis
Design
Car system
Database
Software
Circuit network
Fig. 1
Circuit server
Measurement and analysis system
Measurement and Analysis Techniques of Formula One Chassis Development
VTS). Performance indicies are also calculated

simultaneously at this time, enabling engineers to obtain
analysis results and investigate the next setup changes
before the car returns to the pit. This real-time data is
distributed over a network, enabling simultaneous use not
only in the pit, but by many members in the team factory
and other locations with Internet connections.
When the car returns to the garage and the data
downloading starts, the automatic data analysis system
(VAP: Vehicle Analysis Package) transforms the data
immediately to physical values within minutes. This data
consists of approximately 2500 items that record the car
status at sampling rates of up to 10 kHz, and the data
volume can reach up to several GB per run. In addition
to running data, virtually all information related to
running is also managed centrally on the server. This
information includes the lap times, setup information,
driver comments and wireless audio for each car,
weather and track condition information, work
instructions and communication history between
members, trouble information, and simulation analysis
results.
The data is also transferred to the team factory (HRF)
and Automobile R&D Center Tochigi (HGT) servers,
enabling engineers in various fields to promote
development while sharing data on the same information
infrastructure, without the limitations of place or
environment.
Including both races and running tests, Formula One
running is performed for approximately 125 days per
year (2008 results). Running data for up to 200 laps per
day may be recorded, so measures to increase and
maintain the accuracy and reliability of this enormous
amount of measurement data are demanded. On the other
hand, the chassis needs to be as slim, lightweight and
have as low a center of gravity as possible, so the weight
and space that can be used are limited. In addition, the
engine that is rigidly connected to the chassis produces
vibration over 3000 G, and the exhaust heat over 1000
C. Therefore, efforts are made to develop original
sensors and new measurement methods to support the
needs of high accuracy, compact size, light weight, and
high durability. Specific examples are introduced below.
calculation box (MCB) installed on the chassis, are

mounted on the wheel disc surface. Matrix calculations
and rotation angle and temperature compensation
processing are performed inside the MCB on the data
received by the chassis side to obtain six-component tire
force data that is transmitted via CAN to the data logger
on the chassis.
Loads of up to 30 kN and a bending moment of up
to 4000 Nm can be measured under an operating
environment with a maximum speed of 360 km/h and
maximum temperature of 120 C, which covers tire force
measurement in the limit running state of a Formula One
car. In particular, heat from around the brakes, which
gets as hot as 700 C or more, affects the temperature
characteristics of electronic components including the
strain sensors, and may result in a drop in measurement
accuracy. Therefore, temperature compensation logic and
a corresponding calibration method were established to
address the issue of heat-induced changes in
characteristics.
2.3. Tire Bulk Temperature Sensor (TBTS) and Tire
Internal Surface Temperature Sensor (TIST)
Formula One tire performance is exercised by
precisely controlling the temperature, load, slip, wear,
and the like. Of these, the tire internal (structural portion)
temperature is one of the most important control
elements. Usually the tire surface temperatures are
Support arm
for signal output
Fig. 2
2.2. Strain Gauged Wheel (SGW)

A six tire forces measuring wheel system (SGW) has
been developed for Formula One excessive use. The
requirements were to give accuracy of 1%, compact and
easy to mount. This SGW uses a wireless system that
is lighter weight and has less aerodynamic effect than
the conventional wired system (Fig. 2). This reduced the
aerodynamic effect of mounting, shortened the mounting
preparation time, and enabled the acquisition of useful
tire data with limited track testing (Fig. 3).
The force generated on the tire is measured using a
strain sensor-type load cell configured by four pairs of
bridge circuits (4 gauge method) built into the wheel. In
addition, a measurement value processing box, a rotation
angle calculation box, a battery box, and a transmitter
box that wirelessly transmits the data to a matrix
180
Transmitter box
Processing box
Fig. 3
Conventional SGW
Angle calculation box
Battery box
Wireless SGW
measured during the run, and the internal bulk

temperature inside the garage using probes. A hole can
be made in part of the tread rubber and a thermometric
element embedded to measure the temperature of
passenger car and truck tires while running. However,
Formula One tires have thin tread rubber just several mm
thick, and the high loads and high-speed rotation make
it a challenge to apply a similar method. Therefore, two
types of sensors were developed: TBTS that embeds
elements inside the tread rubber during tire manufacture
(Fig. 4), and TIST that affixes elements to the tire
internal surface (Fig. 5). A compact multi-channel
telemetry system was also developed at the same time
to acquire the data from the rotating tires, and this
enabled measurement at racing speeds.
2.4. Multi-point Tire Temperature Sensor and Onboard
Thermal Camera
To obtain continuous performance from Formula One
tires, it is also important to maintain a proper pressure
distribution within the tire contact patch. The load on the
tire due to this pressure distribution can be estimated by
measuring the tire surface temperature distribution during
running, so infrared tire temperature sensors capable of
simultaneously measuring multiple points were
developed (Fig. 6). First, a 5-point sensor using multipoint thermopiles was developed to support grooved tires
(5 ribs), and the number of points was then expanded
to 8 and 16 points to increase the distribution
Molded
temperature
sensor
Transmitter
measurement resolution.
In parallel with this, a method that uses an onboard
infrared camera to measure the tire surface temperature
was also developed. A compact infrared camera was
installed on the roll hoop where the video camera for TV
broadcasts is normally mounted, and temperature images
of all four tires were recorded using a four-sided mirror
(Fig. 7). There is also the method of mounting multiple
cameras aimed at each tire, but the mirror method has the
merits of light weight, compact size, and high data quality
due to recording all four tires at the same calibration. The
recorded temperature images alone are useful information
for evaluating the tire temperature distribution, but even
more information can be extracted by analyzing the
images using originally developed software, and
converting into temperature distribution data.
2.5. Strain Gauged Suspension (SGS)
The strain gauged wheels that directly measures the
force generated on the tires during running are useful in
track tests that evaluate dynamic performance, but the
increase in weight and the aerodynamic effects of the
special wheel shape cannot be eliminated, so use is
limited to tests. That is to say, a condition for the
application of measuring systems to races is that there
be no effect on vehicle dynamics if at all possible.
Therefore, development of the SGS was promoted based
on the concept of driving six tire forces from the loads
on the suspension arms (Fig. 8).
A Formula One car suspension uses a double wishbone
suspension type, but each arm is basically viewed as a
Camera
Fig. 4
Tire bulk temperature sensor (TBTS)

FL
Temperature sensor
Transmitter
FR
RL
RR
Fig. 7
Fig. 5
Onboard thermal camera
Tire internal surface temperature sensor (TIST)

Sensing range
Sensor
Fig. 6
Multiple tire surface temperature sensor
Fig. 8
181
Strain gauged suspension (SGS)
single straight link, and the six tire forces can be

calculated from the combined force of each axle force
including the push rod. That is to say, the six tire forces
are obtained by performing finite element analysis and
bench tests beforehand to derive the component force
matrix that expresses the relationship between the six tire
forces and the suspension arm axle forces, and then
applying the inverse matrix of this component force matrix
to the axle force measured during running. Realization of
this system requires technologies that can handle A-shaped
arms that are not straight links, take into account the
bending stress due to the flexure joint, support changes
in the component force matrix due to cornering and the
suspension stroke, and take into account the effects of
temperature, vibration, and other factors.
contact pressure distribution in the state with

longitudinal and lateral forces acting on the tire. The
sensor sheet is resistant to compression but weak in the
shearing direction, so it was covered by a protective
plate (0.1 mm thick SUS material), and then a non-slip
sheet was affixed to the top of the protective plate (Fig.
11). Use of this system in combination with the
Dynamic Vehicle Simulator (DVS) or other bench tester
enables quantitative evaluation of the suspension
dynamic characteristics and vehicle setup from the
viewpoint of proper contact between the tire and the
road surface (Fig. 12).
Figure 13 shows an example of analyzing the
2.6. Wing Load Cells

The aerodynamic drag and lifts are usually calculated
by using pushrod loads, but a direct wing load
measurement system has been required to further
understand the phenomenon around the wing elements
itself. The requirements was to create a load cell that
does not affect the aerodynamics, having ability to
change wing elements without touching the load cell unit
and off course strong enough to run on the circuit. A
pillar type load cell (Fig. 9) was used for the front wings
and a rear impact structure integrated load cell type (Fig.
10) on the rears.
2.7. Measurement Technology Using Test Rigs
In addition to track tests, bench tests are also often
used in Formula One chassis development. Tests under
stable environments with few undetermined elements
enable the acquisition of detailed and precise information
that cannot be obtained from track tests. In particular,
the use of bench tests to obtain data that is a challenge
to measure during running, such as the tire contact
pressure distribution, helps to further deepen
understanding of vehicle dynamics mechanisms.
The Contact Patch evaluation (CPE) system was
developed as a bench test system that measures the tire
Protective plate + Nonslip sheet
Sensor sheet
Fig. 11
CPE sensor
DVS
CPE System
Fig. 12
CPE system on DVS

T11
T6
T9
T4
T9-10
T3-4
T9-10,end
T5
T10
T3
T8
EOS
T2
T1
Fig. 9
Straight
T7
Front wing load cell

RA106 Barcelona
BUT U2, T10
FL
FR
RL
RR
Load cell
Fig. 10
Fig. 13
Rear wing load cell
182
Circuit simulation with CPE
3. Competitor Performance Measurement

and Analysis Technique
Understanding the primary factors for differences in
lap times with competing cars during races is important
for determining development directions. The ideal is to
originally create the best in all technical areas, but in
reality it is necessary to analyze the weak points of ones
own car and formulate countermeasure policies to make
efficient use of limited time and resources. Unlike
production car development, it is unlikely that competing
cars can be obtained for direct comparison and analysis,
so original technology was developed to determine and
analyze the primary factors for differences in
performance without actually touching competing cars.
images and images filmed by digital cameras or other

means. The tire contact camber angle and the chassis roll
angle can be estimated from images that filmed
cornering (Fig. 16). Analysis of this information together
with the abovementioned thermography temperature data
enables an understanding of differences in the chassis
characteristics, suspension characteristics, and tire load
conditions.
Overlaying the running images of multiple cars
filmed during circuit running enables analysis of
differences in vehicle speeds, braking points, racing
lines, handling characteristics and other information (Fig.
17). In addition, minute time differences over the filmed
section can be calculated from these images. Time
differences over the entire course can also be calculated
from the onboard camera images of TV broadcasts.
3.1. Acoustic Analysis

Regulation changes in 2008 added a maximum engine
speed limit of 19000 rpm, but prior to that it was
important to know the engine speed of competing cars
as a guideline for estimating the maximum power output.
During TV broadcasts of races, sound is broadcast
simultaneously with the image from onboard cameras,
and the engine speed can be accurately known from the
frequency of the engine noise. At test tracks without TV
broadcasting, the sound is recorded and speed measured
simultaneously using a PC, microphone and speed gun,
and the engine speed is calculated from the engine sound
by compensating for the Doppler effect. This data can
then be further processed to estimate the vehicle speed
diagram, gear ratio, and other information (Fig. 14).
Temperature
constantly changing tire contact pressure distribution

during running, simulated using the circuit simulation
function of the DVS.
120
110
100
90
80
70
60
50
40
Average of tire rib temperatures
Out
Lap1
Fig. 15
3.2. Infrared Thermal Images

The load on the tires differs for each car due to
differences in the weight distribution, height of the center
of gravity, suspension geometry, setup, aerodynamic
characteristics, driving style, and other factors, and
results in differences in the tire surface temperature
distribution. Differences in the characteristics of each car
can be estimated by installing thermal cameras on the
course side, imaging the tire surface temperatures of each
car during running, and comparing the differences in the
temperature distributions (Fig. 15).
3.3. Image Analysis
Various analyses can be made based on TV broadcast
Lap2
2
1
Angle
Car speed
Lap3
Infrared thermal image
Roll
0
Engine revolutions
RL
RR
-1
Outside camber
-2
-3
Inside camber
-4
Gear ratio
Fig. 14
-5
Acoustic analysis
Car A
Fig. 16
183
Car B
Image analysis
In
3.4. High Speed Video Image Analysis

Various pieces of information can be obtained by
using high-speed cameras that have a faster shutter speed
and a higher frame rate than normal video cameras. The
slip ratio (traction control performance) of the driving
wheels during acceleration can be calculated from the
difference in the rotation angles of the front and rear
tires, and the tire surface wear conditions can also be
observed (Fig. 18).
3.5. Three-dimensional Modeling
Three-dimensional models are created using multiple
photos to understand the aerodynamic components, the
suspension geometry and other items (Fig. 19).
Comparing these models on the CAD system enables us
to help validate the competitor performances.
Car B
Car A
Fig. 17
Tire wear
Aerodynamic characteristics (down force and drag)

and the three-component force (wheel load, longitudinal
force, lateral force) acting on the tires during running are
important information for analyzing Formula One
dynamics. This information is used to determine the
policies for aerodynamic development and tire
management and the vehicle setup. However, it is a
challenge to directly measure these forces, so it is
required to perform high level estimation using a
combination of various sensor data, suspension
characteristic information obtained from rig tests and
needless to say about the vehicle dynamics
understanding.
During races and track tests, well over 100 different
pieces of chassis performance data are required for
analysis, which makes development of a system that
automatically calculates vast amounts of chassis
performance data vital to enable swift analysis.
Therefore, an automatic analyzing system called the
Vehicle Analysis Package (VAP) was developed to
support performance analysis. This system used the
onboard data and the suspension model information and
more to say it allowed real time analysis using the
telemetry system.
Image overlay processing
Wheel speed
Sensor FL
Video FL
Fig. 18
4. Measurement Data Analysis Systems
High speed video analysis
Sensor RL
Video RL
4.1. Overview of VAP System

The VAP system is introduced below using the
example of calculating the down force, which is a main
analysis item. Thus far limitations on down force
calculations meant that evaluation of aerodynamic
performance during running had been possible only
under the limited conditions of mode running at a
constant vehicle speed on a straight track. In contrast,
dozens of sensor values and complex models need be
used to enable calculations over all ranges, including
braking, driving and turning, while running at racing
speeds. Figure 20 shows an overview of the down force
calculation model. Here, the down force is derived from
the calculation results for the two main divisions of the
wheel load and the inertia force.
Fuel
Lateral acc.
Instantaneous
mass
Yaw rate
Inflation
pressure
Brake pressure
Engine brake
Lateral force
Tire loaded
radius
Long. force
Longitudinal acc.
Roll/Pitch
Speed
Swing arm angle

Steering wheel
angle (front only)
Camber
Drag
PW ratio
Damper length
Anti force
Pushrod load
Wheel stroke
Down force,
balance
Vertical force
Flexure force
Pitot pressure
Ride height
CoG position
Fuel
Pitch angular
inertia
Long. inertia
Vertical inertia
Vertical acc.
Fig. 19
Modeling in three dimensions
Fig. 20
184
Longitudinal acc.
Instantaneous
mass
Category
Calculation
Measurement
Overview of down force calculation in VAP
The wheel load comprises the load on the pushrod,

which bears the main load, the flexure force due to the
bend arm, and the jack-up force due to the suspension
geometry. Calculation of the jack-up force in particular
also takes into account the suspension stroke, tire
deflection, and other factors, so the wheel load is
calculated using a combination of over 20 different
sensor values and suspension and tire characteristics
models.
The calculation of inertial force is needed in order
to calculate the load transfer amount due to braking,
driving and turning, and to eliminate variation in the
wheel load, which is due to vertical centrifugal force
resulting from the vertical bending radius of the circuit
course, from the down force calculations. This means
that the acceleration calculation accuracy is important,
and requires accurate calculation of the pitch and roll
angles and the height of the center of gravity.
The calculation accuracy is verified using a wind
tunnel and bench testers such as the DVS (Fig. 21).
Figure 22 shows the results of accuracy verification
using the DVS. Here, circuit simulations were performed
using running data, and the VAP calculation results were
compared with the down force applied as a load from
the DVS. The results show a close match for both the
front and rear wheels.
4.2. Introduction of Data Analysis Systems that Support
Telemetry
Regulation changes implemented by the FIA from
2008 onward mandated the introduction of a common
ECU made by McLaren Electronic Systems (MES). One
of the functions of this ECU is a telemetry data analysis
system (vTagServer: VTS).
The main function of telemetry systems thus far had
been to transmit the information from onboard sensors
to the pit, and the main role for this had been to monitor
trouble occurring in the chassis while running. In
contrast, one of the functions strongly desired from the
Fig. 21
Correlation test with wind tunnel and DVS

Front
12000
y = 0.9973x + 0.2781
VAP rear (N)
VAP front (N)
4000
12000
8000
4000
0
4000
8000
DVS front (N)
Fig. 22
12000
standpoint of vehicle dynamics development was realtime analysis of dynamic performance while running. At
circuits it is necessary to analyze running data, determine
the next setup, and determine which running modes need
to be added, all within the limited time after the vehicle
returns to the pit until the next run. Therefore, real-time
analysis was strongly desired by engineers engaged in
vehicle dynamics development.
With the introduction of this new telemetry system,
the logic of the automatic data analysis system was
transplanted to the Simulink model, which achieved realtime data analysis by processing data on the VTS. In
addition, the VTS can also be used as a post processing
system, and development of an integrated system is
proceeding with the aim of realizing a seamless
environment from real-time analysis to more detailed
analysis (Fig. 23).
VTS (live processing)
VAP (post processing)
Car (running)
Car (running)
Telemetry
Garage
Garage
Car (back from track)
Car setup data
Car setup data

ATLAS data server
Outing D/B
vTagVAP
System monitor
ATLAS
ATLAS
ATLAS
Engineers
Fig. 23
Car data
vTAGServer
WBApps
Data offload
Outing D/B
ATLAS
VAP results
VAP
ATLAS
ATLAS
ATLAS
ATLAS
Engineers
Overview of live and post processing system
5. Creation of Vehicle Dynamics Indices

To evaluate the analysis data as an expression for
vehicle dynamics, and to use those results for vehicle
design and setup at circuits, the ability to extract physical
quantities that represent performance and to discuss these
quantitatively and simply is required. That is to say,
dynamic performance indices are important.
Without dynamic performance indices, it is a
challenge to quantitatively compare multiple setups in a
short time at circuits, where speed is required, and such
work cannot help but rely on driver comments and the
experience of the engineers. This also makes quantitative
investigation a challenge in the area of vehicle design,
such as when discussing vehicle dynamics targets or
tradeoffs between various design elements.
Rear
16000
y = 0.9979x + 0.8252
8000
4000
8000
12000
DVS rear (N)
Correlation test on DVS
16000
5.1. Development of Braking Stability Indices

Vehicles with good braking stability not only enable
shorter lap times by delaying braking, but also increase
the chances of overtaking using braking. For these
reasons, braking stability is one of the most important
dynamic performance in Formula One. However, while
it is discussed as the drivers sense of confidence in the
car while braking, there was no index that expressed
the braking stability in a quantitative manner. Therefore,
185
Body slip angle [deg]
Fig. 24
Correlation coefficient
0.6
0.4
5% significance level
0.2
Insignificant
0
10
15
20
25
30
35
Number of samples
Fig. 25
Significance test for Spearmans rank

correlation coefficient
Circuit 1
45
Close to apex point
Beginning of braking
35
Beginning of braking
Close to apex point
CoP [%]
40
30
50
Circuit 2
45
Baseline (B)
Setup 1 (S1)
Setup 2 (S2)
40
35
Baseline 1 (B1)
Setup 1 (S1)
Setup 3 (S3)
Baseline 2 (B2)
Setup 2 (S2)
30
100
Fig. 26
150
200
Car speed [km/h]
250
50
100
150
200
250
Car speed [km/h]
Setups with different CoP transition in braking
1.0
0.8
Stabilizing yaw moment margin (MM)
0.6
Lower envelope of vertical load on rear outside tire (Fz)

0.521
0.4
0.2
0.447
Insignificant region
0.0
-0.2
Weak
-0.4
Medium
-0.6
-0.8
Actual load
Strong
Circuit 1
Time
T8 (S2, S3)
T8 (B2, S2)
T8 (B1, B2)
T2 (S2, S3)
T2 (B2, S2)
T1 (S2, S3)
T1 (B2, S2)
T1 (B1, B2)
T11 (B, S1, S2)
T9 (B, S1, S2)
T10 (B, S1, S2)
Lower envelope of vertical

load on rear outside tire
T5-T6 (B, S1, S2)
-1.0
T3-T4 (B, S1, S2)
Current
car state
Significant
0.8
0.0
Stabilizing yaw moment margin

Tire vertical load
Stabilizing yaw moment [kNm]
Slope of stabilizing yaw moment
1.0
Forward
5.3. Verification of Indices

The premise was that cars with higher braking
stability enable stronger, more stable braking, with the
result that section times should be enhanced. Running
was performed using various proposed setups thought to
provide different stability, the correlations between each
physical quantity over the braking section and the section
time were analyzed, and the physical quantities with a
correlation coefficient of 0.7 or more, which indicates a
strong correlation, were extracted from the results. In
addition, the braking sections were divided into straight
braking sections and turn braking sections, and the
correlations were analyzed for each section.
The relationship between the two factors may be
nonlinear, so a non-parametric analysis method called
Spearmans rank correlation coefficient was used. In
addition, a no correlation null hypothesis to be rejected
was established, and hypothesis testing was performed
using a 5% standard to examine whether the calculated
correlation coefficients are significant. The correlation
coefficient value that could be obtained only at a 5%
probability when there was no correlation between the
5.4. Test Methods and Verification Results

Proposed setups with different transitions in the
center of down force (CoP) during braking were set.
Running tests were conducted on two different circuit
tracks to compare a baseline setup (the CoP shifts to the
front from the beginning of braking towards the corner
apex point) with other setups changed so that the CoP
shifts towards the rear, which was presumed to enhance
braking stability (Fig. 26).
Figure 27 shows the correlation coefficient analysis
results for straight braking sections. Of all the
CoP [%]
5.2. Candidates for Braking Stability Indices

Ten different physical quantities thought to
potentially express braking stability were hypothesized
as index candidates, and each candidate was verified to
determine whether it was suitable as an index. The
typical index physical quantity candidates are described
below (Fig. 24).
(1) Stabilizing yaw moment margin (MM)
The body slip angle was imaginarily increased from the
current car state, and the driver was hypothesized to feel
confidence according to the difference (margin) from the
maximum value of the calculated stabilizing yaw
moment.
(2) Stabilizing yaw moment variation rate (dM/ds)
The degree of sensitivity relative to the body slip angle
of the stabilizing yaw moment in the current car state
was hypothesized to affect confidence.
(3) Lower envelope of vertical load on rear outer wheel (Fz)
It was hypothesized that the tire gripping force is unable
to track high frequency fluctuations in the wheel load,
with the result that only force equivalent to the lower
envelope of the wheel load can be exercised.
two factors was calculated, and when a correlation

coefficient that exceeded this value was calculated, the
null hypothesis was rejected and the factors were judged
to have a significant correlation (Fig. 25).
braking stability indices were developed as an index of

dynamic performance.
Circuit 2
Turn
Candidates for braking stability indices
Fig. 27
186
Correlation analysis in straight braking
candidates, this figure shows the results of only MM and

Fz, which exhibited stably high correlations. Of the MM
data suitable for analysis, 80.0% of Circuit 1 samples
exhibited a significant correlation, and 20.0% showed a
strong correlation. For Circuit 2 these figures were
87.5% and 50.0%, respectively. This exceeded the Fz
results for Circuit 1 (40.0% significant correlation, 20.0%
strong correlation) and Circuit 2 (75% significant
correlation, 25% strong correlation), indicating that MM
is more suitable as an index in straight braking sections.
In turn braking sections, MM and dM/ds exhibited
higher correlations compared to other candidates (Fig.
28). However, these correlations were not as stably high
as that of MM in straight braking sections. Here, 33.3%
of MM samples for Circuit 1 showed a significant
correlation (33.3% strong correlation), and 66.6%
(0.00%) for Circuit 2. The respective results for dM/ds
were 66.6% (33.3%) for Circuit 1, and 25.0% (8.33%)
for Circuit 2.
conditions such as vibration heat and speed are so severe

that a very high task is required. These techniques
obtained in this project will be useful in passenger car
development as well.
Acknowledgements
The authors wish to express their deep thanks to
related parties at Tokyo Sokki Kenkyujo Co., Ltd., for
their cooperation in developing the Strain Gauged Wheel,
Kyowa Electronic Instruments Co., Ltd., for their
cooperation in developing the wing load cells, TUV SUD
Automotive GmbH for their cooperation in developing
tire temperature sensors, and Bridgestone Corporation for
their cooperation in developing vehicle dynamics indices.
5.5. Summary of Index Creation

It is necessary to create dynamic performance indices
in order to scientifically and quantitatively discuss
vehicle design and setup without being dependent on
engineer experience and driver senses. Indices were
created for braking stability, which is one of the many
items that comprises dynamic performance, and
knowledge was gained regarding physical quantities that
can serve as indices. Indices need also be created
successively for other vehicle dynamics items.
1.0
Stabilizing yaw moment margin (MM)
0.8
Slope of stabilizing yaw moment (dM/d)
0.6
0.521
0.4
0.447
0.2
Insignificant region
0.0
-0.2
Weak
-0.4
Medium
-0.6
-0.8
Strong
Author
Circuit 1
T9 (B2, S3)
T9 (S2, S3)
T9 (S2, B2)
T9 (B1, S3)
T9 (B1, B2)
T9 (B1, S1)
T8 (B2, S3)
T8 (S2, S3)
T2 (B2, S2)
T2 (B2, S1)
T1 (B2, S3)
T1 (S2, S3)
T9 (B, S1, S2)
T7 (B, S1, S2)
T1-T2 (B, S1, S2)
-1.0
Circuit 2
Turn
Fig. 28
Correlation analysis in turn-in braking

Hideaki SHIBUE
Tomokazu SUZUKI
Mamoru URAKI
Akihiro IDA
Atsushi TSUBOUCHI
Yasutaka KITAKI
6. Conclusion
On a qualifying session, the lap time differences
between the top teams are usually in within 0.1 s.
Assuming a lap time of 100 s, this could be called a
competition over differences in performance of
approximately 0.1%. Therefore, the demand to have the
cutting edge technology for both measurements and
analysis is very high. In Formula One, the environmental
187
Descriptions of Vehicle Control Technologies
Development of Traction Control Systems for

Formula One
Kazuharu KIDERA*
Yoichiro FUKAO*
Tatsuya ITO*
ABSTRACT
Efforts have been made toward the development of traction control that controls the amount of tire slip at a level
that exceeds the abilities of Formula One drivers and enhances turning acceleration performance, and the development
of overrun control that is intended to prevent rear tire lockup by using engine torque when the brakes are being applied.
Numerous control systems have been applied in racing based on modern control theories, including high-precision
engine speed control and wheel speed feedback control, which is extremely challenging in Formula One, where driving
at the limit is the norm.
Launch control, which is control applied in the start of a race, has also realized complete direct control of clutch
transfer torque. Development has proceeded on start assist systems that can enhance start performance with reliable
repeatability while continuing to satisfy regulations even after prohibition of automatic clutch control. The direct push
clutch and other new technologies have been actively adopted.
1. Introduction
Traction control was extremely important to Formula
One activities of the third era, which started with the
Australian Grand Prix in 2000.
The Federation Internationale de lAutomobile (FIA)
had basically prohibited traction control until the Spanish
Grand Prix in 2001. The FIA changed its stance at that
time, completely lifted the ban, and development of
traction control started.
Vehicle dynamic performance continued to evolve
even while restrictions were subsequently placed on tire
performance. In this context, the optimal control of
traction and the maximum employment of tire
capabilities occupied crucial positions in the evolution
of vehicle dynamic performance.
In 2008, the FIA unified control specifications by
making it obligatory to install a common electronic
control unit (ECU). Even after control systems that could
be used as driver aids were completely eliminated, there
was continuing high demand for control systems to make
it easier for drivers to manage torque during cornering.
Given these circumstances, development of traction
control continued largely throughout the period of thirdera activity.
This article describes the history and development
technology of various systems, including traction control
(TC) in the acceleration range, the development of
methods for setting traction that were implemented by
throttle pedal input after TC was prohibited, overrun

control (OC) in the deceleration range, launch control
(LC) that enables fully automatic launching with the
press of a single button in the launch range, and the race
start system (RS) presupposing manual clutch operation
that was developed after 2004 when LC was prohibited.
2. Development Goals
The purpose of traction control in Formula One
racing is to assist the drivers so that they can drive the
car as fast as possible with repeatability. It is necessary
with racing cars in general, however, not only to bring
out their performance limits, but also to sustain that
unstable state. With mass-production vehicles, it is a
requirement of similar control systems that, for safety
reasons, they keep the vehicle inside the limit range. This
difference in what the two types of system are intended
to achieve is also the difference in the performance they
are required to produce.
Formula One cars and mass-production vehicles also
differ greatly in the hardware that is subjected to control.
Formula One cars are, of course, light in weight, in the
600 kg range, and their drivetrains are built to have the
lowest possible rotational inertia. Meanwhile, they have
engine power, braking power, and tire grip force that are
several times greater, as absolute values, than in massproduction vehicles.
For the above two reasons, Formula One cars have

190
much more sophisticated traction control than massproduction vehicles. They demand highly precise, highly
responsive systems. Generally speaking, bringing out the
performance limits of a vehicle will mean making the
most of the performance of its tires, which are the only
points of contact with the ground. Control of Formula
One power plant systems is no exception in this. While
effectively transferring the limited engine power to the
tires, it must never exceed the tires limits. The control
methods used can be generally classified into two types:
one is the method for directly controlling wheel torque,
and the other is the method for controlling the amount
of tire slip. Another, indirect method for accomplishing
this is to control just the engine torque or the engine
speed. Drivetrain components have rotational inertia,
backlash, and twist. Strictly speaking, therefore, control
by wheel criteria and control by engine criteria are
different, and the various methods are employed
according to the application, the degree of ease or
otherwise of control, and the advantages or disadvantages
involved.
Figure 1 shows the history of the three systems
introduced below. Figure 2 shows a configuration
diagram of the control systems.
other than engine power control, but what it is ultimately

intended to control is the amount of tire slip.
The degree of challenge represented by control, given
the precision required in achieving the desired amount
of slip and response, depends on the delay in generating
torque by the engine, which serves as the actuator, and
the existence of torque transfer delay due to backlash and
twist in the component parts of the drivetrain, which is
situated between the engine and the tires. TC is further
required to assure robust performance with respect to
disturbances from rough road surfaces, curbs, and the
like.
2.2. Aims of Overrun Control (OC)
OC in Formula One cars is a system for control of
engine torque during braking in order to prevent rear tire
lockup and to optimize the amount of tire slip.
Active brake balance control and antilock braking
system (ABS) implemented by braking systems are
forbidden in Formula One racing. However, the
enormous tire grip force and the maximum deceleration
of as much as 5 G realized by carbon brakes result in a
large amount of front and rear load shift, so that the
optimum front and rear brake force balance changes from
moment to moment. OC does more than simply adjust
the engine braking force. When circumstances make it
necessary, even during braking, it produces torque all the
way to the drive side so that front and rear brake force
distribution are optimized while producing performance
that, although limited to the rear tires, is intended to be
similar to ABS.
2.1. Aims of Traction Control (TC)

The aims of TC in Formula One cars are to control
traction at a level surpassing the skill of the driver and
to limit the amount of tire slip appropriately so as to
enhance straight-line acceleration performance and
turning acceleration performance. The TC system is none
00
01
Engine specification
Regulation
02
03
04
05
06
07
V10 cylinder engine
V8 cylinder engine
TC/OC permission
TC/OC
LC
prohibition
08
Standard ECU
LC prohibition
LC permission
LC/RS
TC/OC
prohibition
Clutch auto control prohibition

Rev control prohibition
Interface
(Engine & Chassis
control data)
Throttle & Rev limit

Torque & Rev limit
Torque drive
Rev
Pattern lg TC
Ignition control
Rev
PID F/B TC
Rev
Model base TC
Wheel slip
F/B TC
TC
Torque capping
Th TC
Throttle control
Rev
Blip OC
OC
LC
Rev+
Clutch position control
Fig. 1
Torque reduction F/B Th TC

Rev
One side bank OC
Wheel slip
F/B OC
Rev+
Clutch torque F/B
Rev+
Clutch position control
RS
Pedal
torque
map
2stage throttle
Clutch dump
Clutch pressure control

DPC + Torque rev
History of control system development
191
Pedal
torque
map
The result is that drivers can feel safe about pressing

down hard on the brake pedal even under conditions with
bumps or other such disturbances. When turning in to a
corner, the amount of rear tire slip can be limited so as
to obtain an oversteering reduction effect.
OC has many of its challenging aspects of control in
common with TC. One major difference, however, is that
the primary causal source of tire slip is the braking force,
while the role played by engine torque is just a fraction
of that. Consequently, the margin of control available to
OC is necessarily limited.
2.3. Aims of Launch Control (LC) and the Race Start
System (RS)
LC is a system that automatically implements
Formula One race starting by integrated control of
engine torque, engine speed, and clutch engagement
amount. If drivers activate the system according to a
determined procedure, all that remains is for them to
push a single button. LC then appropriately controls the
amount of rear tire slip and realizes the maximum
acceleration from zero to 100 km/h. (LC was prohibited
by Formula One regulations in 2004.)
On the other hand, RS could be termed the drivers
manual race start assistance system that followed the
prohibition of LC. Development of standing start control
within the range permitted by regulations also took place
following the prohibition of LC, and this has been a
battle of wits among the teams and with the FIA since
2004. The aim of RS is to realize the best standing start
performance, with repeatability, by arranging it so that
the driver only has to follow a predetermined standing
start procedure (throttle pedal operation, clutch paddle
operation) to have the proper settings for engine and
clutch control be selected by engineers to match with the
coefficient of friction between the road surface and the
tires, as well as with the meteorological conditions.
The two challenges that LC and RS face in common
can be summarized as control of clutch transfer torque
with a high level of precision and response, and
cooperative control of the engine and clutch.
3. Development of Traction Control (TC)

In 2001, the use of TC was explicitly permitted in
the regulations, and it was used in racing from that point
to 2007. In 2008, TC again came under complete
prohibition when the FIA instituted the requirement for
a common ECU. This section describes the development
of TC together with the measures used to enhance
drivability following the prohibition of TC.
3.1. TC Development
TC development for third-era Formula One activities
can generally be classified under one or the other of the
following two headings:
Engine controller
Chassis controller
Engine speed
Rev limit
Throttle pedal
TC controller
(rev limiter)
Target throttle
Throttle pedal
Fuel controller IG controller
Target fuel
Throttle controller
Target IG
Wheel speed
Lateral
acceleration
Pattern IG
Throttle
Environment
TC/LC
controller
OC controller
Target throttle
00 - 01
Engine
Engine controller
Chassis controller
Engine speed
TC controller
(rev limiter)
Throttle
pedal
Rev TC torque
Driver
torque
Pedal
torque map
Rev limit
Target torque
Throttle pedal
TC controller
(rev or wheel slip)
Wheel speed
Gearbox controller
Gear position
Torque controller
Map search
Calculated torque
RS controller
IG cut/retard
torque divide
Throttle
Environment
Lateral
acceleration
OC controller
Fuel controller IG controller

Target fuel
Target IG
Target throttle
07 final system
Engine
Fig. 2
Configuration of control system
192
Trackability of target tire slip

Determination of target tire slip that yields maximum
acceleration performance
At the initial stage of development, the target tire slip
was not very trackable, and some instances were
observed of factors that disrupted vehicle behavior. Such
factors included insufficient toughness with regard to
disturbances such as rough road surfaces and curbs, and
inadequate control due to insufficient consideration of
the transfer properties involved from the engine target
torque to the tire traction force. The TC development
effort in the beginning and middle periods of the third
era sought to resolve these matters primarily by
enhancing the trackability of target tire slip, and
development work focused on a controller for this
purpose.
respect to disturbance.
Figure 4 shows the actual track data for Rev TC. It
can be seen how the target tire slip is tracked by
controlling the engine speed. Rev TC has the advantages
of not being susceptible to system delays, and of fast
response. On the other hand, it does not allow
independent control using rear left or rear right tire slip,
and has disadvantages coping with curbs and the like.
In parallel with development of the Rev TC,
development of wheel slip feedback TC (W/S F/B TC)
also took place. This is a method of control that uses
the tire slip as an F/B parameter. This allows traction
control to use only favorable side of rear wheels or both
during cornering, and it increases flexibility of the
settings. It has a robustness with regard to the input of
disturbances from the road surface that was established
in advance through simulations, and it also enabled
limitation of excessive torque reductions.
Figure 5 shows the actual track data for W/S F/B
TC. The wheel torque is subject to finely tuned control
with respect to changing load on the rear tires (rear inner
and outer tire Fz), and the tire slip is retained in the TC
slip target value.
In the third era, the above two control methods (Rev
F/B TC and W/S F/B TC) were employed differently
according to the circumstances. Rev TC was mainly
employed with low speed gear (in racing starts and when
building up speed coming out of a low speed corner) or
on wet road surfaces, when the traction is large against
reaction force from road surface and the F/B system
requires high response and high resolution. W/S F/B TC,
which allows more flexible control, was employed
mainly for driving in middle and high speed gears and
under conditions when disturbances such as curbs could
have an influence.
Important points of TC F/B systems include not only
the consideration of system delay and robustness with
respect to disturbances, but also the fact that, in the
interest of drivability, it limits excessive ignition control
fluctuations that ignore the will of the driver, and that,
in the interest of fuel economy, it limits the amount of
fuel injected. Throttle TC (Th TC) implements control
3.1.1. Enhancement of trackability of target tire slip

Methods for controlling tire slip include direct
methods of using actual slip as a feedback parameter for
target tire slip, and indirect methods of calculating a
target engine speed from the target tire slip and reduction
ratio, and then using the engine speed as a feedback (F/
B) parameter (Fig. 3).
The engine speed control TC (Rev TC) that had been
in use from the start of the third era corresponds to the
latter method. The controller of those days used the
deviation from the target engine speed as a basis for
specifying a predetermined ignition pattern [Pattern
Ignition (IG)]. It did not implement control using the
engine torque. In the subsequent course of creating
various different drivetrain control systems, engine
torque was taken as the common parameter, and
consolidating it in the torque interface led to enhanced
precision of the control system and expansion of the
degrees of freedom (introduction of torque drive system).
Rev TC subsequently had the convergence and torque
linearity enhanced by the adoption of PID F/B control.
The further application of modern control theory that
takes engine characteristics into account enhanced
convergence still more, while the optimization of the
torque control amount also heightened toughness with
Rev TC
Throttle pedal
Target
wheel speed
Lateral acceleration
Target
engine
torque
Target
engine
speed
Gear
ratio
Car speed
Rev
F/B
Controller
Engine
Chassis
Engine speed
Wheel slip F/B TC

Throttle pedal
Car speed
Target
wheel speed
+
Lateral acceleration
Target
engine
torque
Target
wheel torque
Speed
F/B
Controller
Gear
ratio
Wheel speed
Fig. 3
Feedback control system in TC
193
Engine
Chassis
for the above purposes. It continued to make

enhancements following from Torque capping Th TC,
which was close to the sequence control employed in
initial development, and ultimately was able to employ
control by a torque reduction F/B system that realized a
fuel economy increase of approximately 3% and ignition

control that provides a stable, small amount of fuel
injection even during the input of disturbances.
3.1.2. Calculation of target tire slip
Enhancement of feedback control systems came to
assure stable trackability of target slip, and TC
development shifted to ascertaining optimal target slip.
Starting in the summer of 2006, attempts were made to
use the car model and tire model in the ECU together
with data from the various types of sensors to estimate
the vehicle state quantities, and then to calculate target
slip through an integrated control that also included diff
control (electronic limited-slip differential gear control).
The method that was finally adopted, however, was to
store multiple maps of target slip based on driver throttle
pedal positions and vehicle lateral G forces.
The above methods adopted from the beginning of
the third era had the advantage of enabling settings that
responded to the feelings of drivers and reflected their
will. On the other hand, it is crucial to be able to
determine the settings that are appropriate for the
corners, road surface, and tire conditions of each
individual circuit, and that further support the skills and
predilections of the driver. The challenges with respect
to the time and labor involved and the methods
adaptability were never fully resolved.
Throttle pedal
Gear ratio
Target
engine speed
Engine speed
(F/B parameter)
Target wheel slip

Wheel slip
5.0
Error slip frequency
4.0
ERR SLIP TC0
3.0
2.0
1.0
0.0
3.2. Development Following TC Prohibition

The change of regulations in 2008 placed control of
overall vehicle systems, including the engine and
gearbox, in the ECU manufactured by McLaren
Electronic Systems, which would become the common
FIA ECU. TC and OC were prohibited. As a result, lack
of traction during cornering and poor drivability emerged
as issues. It therefore became necessary to enhance both
hardware and software aspects of the systems.
1.0
Target slip line
2.0
3.0
4.0
5.0
Overshoot area
Fig. 4
Rev TC track data
6000
Tire slip
10
5
Throttle torque demand
4
2000
TC wheel torque demand

3
6000
4000
2
0
Rear outer tire Fz

Gear position
Rear inner tire Fz
2000
29
30
Fig. 5
31
Engine friction torque

32
Wheel slip feedback TC track data
194
33
Time: 27.870 sec
2000
1
Tire Fz [N]
10
8000
4000
Gear position
TC wheel torque demand [Nm]
TC slip target [km/h]
10000
TC slip target
Development of powertrain systems proceeded with

a focus on torque accuracy. The reason for doing so was
that when TC was in effect, torque values other than
high magnitude torque errors were absorbed by the
feedback system so that weaknesses tended not to come
out in the open. The prohibition of TC, however,
heightened the requirement for torque accuracy,
including transient engine torque characteristics.
Furthermore, the method for defining the target
torque against the throttle pedal could not depend only
on driver feeling. It became necessary to establish setup
methods to substitute for TC.
Evaluation items were therefore examined together with

methods of test bench evaluation.
Since the project was frozen in mid-development,
some parts of activities such as evaluation of the stability
of a vehicle as a whole that relied on driver comments
still remained. In powertrain evaluation, however, and
particularly in evaluation of torque response
characteristics and accuracy, quantitative benchmarks and
test methods can be established. Benchmarks and
methods can be checked in advance before running on
the circuit, and front-loaded development can now be
implemented.
3.2.1. Torque accuracy

Important points for the powertrain from the
perspective of drivability include:
(1) Ability to predict the driving torque of throttle pedal
operation by the driver;
(2) Linear torque characteristics of pedal operation; and
(3) High torque response.
Maintaining the vehicle stability limit during a turn
when any one of these factors is missing becomes a
challenge. With regard to items (2) and (3), solutions are
likely to differ according to driver characteristics and
predilections.
Engine development sought to enhance the air intake
and the exhaust and fuel systems. This yielded
enhancement of combustion stability under low load
(under rich air/fuel ratio conditions), limited variation in
transient throttle opening and closing, and promoted
enhancement of torque response and torque repeatability.
Map measurements of various kinds made with a view
to combustion stability as well as data settings made with
consideration of usage conditions on the circuit also
contributed to increased torque accuracy and enhanced
drivability.
4. Development of Overrun Control (OC)
3.2.2. Definition of target torque

With feedback control for TC being prohibited, the
setting of target torque for the throttle pedal becomes an
item with significant effects on drivability. Drivers
operate the throttle pedal to control the amount of tire
slip when coming back up to speed out of a corner, and
one thing that is crucial at this point, in addition to
assuring the resolution and linearity of the driving torque
against throttle pedal operation, is implementing torque
reduction when excessive slipping occurs. During the
2008 season, the settings were optimized in light of the
traction utilization ranges on the low-speed corners at
every circuit in order to realize all of the factors noted
above. At that time, circuit simulations were also carried
out in advance in order to heighten the efficiency of the
process leading to optimization.
3.3. Establishment of Evaluation Items and Test Bench
Evaluation Methods
The ability to quantitatively analyze the influences
that power plant torque behavior and vehicle settings
have on TC controllability and drivability is a matter of
importance in performing every kind of development.
4.1. Engine Rev Control OC

The development of OC began at the suggestion of
a Honda team member in the autumn of 2001, during
the BAR-Honda period. It became a full-fledged
development item in 2002, and progressed to the point
of being adopted for use in racing.
This was initially a system to control the handover
of the lower engine speed limit (the opposite of the rev
limit), in accordance with the feed forward throttle
position demand and tire slip limit, from the chassis
control unit (Pi-Sigma MCU) to the engine control unit
(ATHENA ECU). From the autumn of 2002, as with TC,
the throttle position demand handed over between the
MCU and ECU was replaced by torque demand.
Subsequently, there were requests from the drivers for
engine braking and antilock performance, which are
mutually contradictory, and in 2003, throttle torque
demand was switched over to an on-demand system
(Blip OC) that responded to engine speed deviation.
Up to the start of 2006, however, the drivers
complained of a pushing feeling and other issues in the
OC caused by a phase delay in the control. Its use was
therefore limited. Changes went on being made in
methods of setting target values, and further tuning
continued, but the rear tire antilocking performance that
was the real purpose of OC remained incomplete.
4.2. Wheel Slip Feedback OC
Wheel slip feedback TC was adopted for racing at
the beginning of 2006, and after several races, a
completely identical control algorithm was also applied
to OC. This was not conventional engine speed control,
but instead allowed the use of rear wheel speed as a
source for direct feedback. Consequently, it allowed fast,
finely calibrated control of engine torque with respect to
disturbances from the road surface, fluctuations in brake
torque, and other such factors.
Figure 6 shows a view of control in OC. The OC
wheel torque demand will have risen and prevented
lockup of the rear tires long before the tire slip amount
reaches the limit slip amount (OC slip target), and when
the slip comes under control, the torque goes down (each
torque value is benchmarked to the wheel torque). The
throttle is opened according to the OC torque demand
with an added torque margin. This enhanced the antilock
195
performance and received favorable comment from the

drivers.
This new OC was based on wheel torque control, and
therefore was easily capable of adjusting the engine
braking torque simply by feed forward control. Since the
necessary amount of engine braking differs according to
the driving style, it is basically intended to provide an
amount suited to the drivers predilections. However,
considering that it tends not to be affected by changes
in the gear ratio, as well as its utility in eliminating
stepped changes in torque during gear shifts, it was
decided to implement control by a map of wheel torque
demand relative to car speed. The value given to
simplicity and ease of tuning in the Formula One context
was another reason for this choice. The data from the
41-second point to the vicinity of the 41.4-second point
in Fig. 6 shows this OC feed forward torque demand
in operation.
During braking, the engine would fire at a higher
frequency, so that the rise in exhaust temperature in the
exhaust pipes was initially an issue in development. This
was addressed by firing only one bank when torque was
adequate to the demand, and using the left and right
banks in alternation every time engine braking was used.
The use of an FIA-common ECU became obligatory
in the regulations starting in 2008, and TC and OC were
prohibited. Driver complaints, however, centered more
on their being unable to use OC than TC. This is because
controlling engine torque during braking of the kind
applied in OC is challenging even with the skills
possessed by Formula One drivers.
5. Development of Launch Control (LC) and

the Race Start System (RS)
As of the end of 2008, it was thought that the RS
control method using the FIA-common ECU would
continue to be used by ex-Honda Racing Formula One
Team even in 2009 and beyond. The discussion here,
10
5.1. Clutch Position Control LC

The development of LC started in 2001, the same
year that the ban on TC was lifted. Up to 2002, one issue
faced in LC was the controllability of the initial tire slip
from immediately after the standing start (launch) until
the clutch is fully engaged. This was because position
control of the pull clutch of that time resulted in large
changes in clamp load due to the amount of disk wear
and thermal expansion that would occur even if the
clutch were held at the same position in a partially
engaged state. The changing clamp load meant that the
transfer torque would become unstable. Originally, such
changing factors should have been absorbed by the tire
slip feedback control portion of the system, but the
processing cycle of systems at that time was too slow
to provide adequate compensation capacity.
5.2. Clutch Torque Control LC
In 2003, in order to resolve the above issues, a noncontact magnetostrictive torque sensor was installed on
the input shaft located between the clutch and the
gearbox, and a system for direct feedback control of
clutch transfer torque was developed. This clutch torque
control had practically no delay compared with TC
controlling engines that experience a torque generation
delay time of 10 msec or more. It therefore provided an
edge in tire slip control.
The realization of direct control of transfer torque
brought the capability to produce feed forward torque
matched to the coefficient of friction between the road
surface and the tire at the starting instant. Once the
vehicle started moving, it could switch over to slip
feedback control for the rear tires, and the amount of tire
slip could be controlled so as to maximize acceleration.
In order to accomplish this switchover as quickly as
possible, a dedicated arithmetic processing unit was
Gear position
Tire slip
6
OC slip target
4000
4
2000
Throttle torque demand

OC wheel torque demand
2
OC feed forward torque demand
Engine friction torque

0
-2000
40
Fig. 6
41
Wheel slip feedback OC track data
196
42
Time: 39.430 sec
Gear position
OC wheel torque demand [Nm]
OC slip target [km/h]
6000
10
therefore, will be limited to the LC of 2003 and earlier,

and the RS used after it up to 2007.
5.4. Clutch Dump RS

A super-short first gear was introduced and RS was
switched over to a clutch dump procedure, starting from
the Japanese Grand Prix in the closing stage of 2004,
Clutch
torque target
For
race start
Fig. 7
197
For manual
operation
Torque level adjustment

for different tire
For full
engagement
5.3. Partial Clutch Engagement RS with Pull Clutch

Position Control
In 2004, the FIA entered on a program of eliminating
driver aid systems, starting with the immediate
prohibition of LC. The regulations required clutch
control to move the clutch to a target position, or to exert
a target pressure on the clutch, determined by the
drivers manual operation. It was too great a challenge
to explicitly state the position with regard to TC in the
regulations, and so TC operation following full clutch
engagement was still permitted. BAR-Honda sought to
simplify the drivers clutch operations by devising a
special clutch paddle map, as shown in Fig. 7, and
combining it with a double paddle system in which
clutch paddles were attached on either side of the reverse
face of the steering wheel, as shown in Fig. 8, so that
they would be in standby at different positions and
would be released at different times.
The paddle on the disengagement side of the two
paddles would serve as the actual paddle demand value.
Therefore, the portions marked with heavier lines in Fig.

8 are the actual clutch position demands. According to
this procedure, the driver shifts from the fully disengaged
state to that of partial clutch engagement, and then to
full engagement, enabling a sequence of operations that
could be carried out rapidly and accurately. In a state
of partial clutch engagement, there will be a fixed target
torque in accordance with the horizontal parts of the
paddle map. This makes use of the fact that at the
extremely low speed in the initial phase of acceleration,
the maximum driving torque of a Formula One car is
close to being a certain fixed value (Fig. 10).
The ECU was equipped in advance with various
internal paddle maps on which the horizontal portions
represented several different torque target values. The
final selection of the map to be used for the actual race
start was determined several tens of seconds before the
start by a control engineer who was standing by in the
pit, and the driver was instructed by radio. It was done
this way because the start was practiced on the actual
grid at the beginning of the formation lap, and the
coefficient of friction between the road surface and the
tires was estimated from telemetry data.
Regardless of ingenious measures of this kind, the
system of that time, whereby a pull clutch was subjected
to position control, did not manage to realize transfer
torque for the launch as intended. This situation
persisted, caused by the state of disk wear, changing
temperatures, and the like, much as was the case with
the issues faced by LC up to 2002.
Clutch
paddle position
installed to detect extremely low wheel speeds. This was

done because slip control of tires when they are
beginning to roll was a major contributing factor to
victory from the start of a race.
With LC, the speed of the drivers reflexes also
makes a significant contribution. Actual measurements
were therefore made of the brake release system, button
operation, paddle operation, and other such systems for
each driver, and the pattern with the smallest delay was
adopted. In the case of an outstanding driver, there is a
reaction time of less than 0.2 second from seeing the
starting signal to taking action on the controls. If the
driver fails to start, then a Formula One car with
acceleration performance taking it from zero to 100 km/
h in under three seconds does not, in practical terms,
allow for recovery even with an outstanding control
system.
Another factor that determines the results of a
Formula One race start is tire temperature management.
The traction performance of a tire becomes optimal in
the vicinity of 80-90C, as no doubt every team was
aware, and very few teams were able to manage this in
an actual race start at that time. BAR-Honda finally
managed to apply a tire surface temperature sensor in
racing from the time of the British Grand Prix in 2003.
This showed for the first time how abnormally low their
tire temperatures were on the starting grid, and the team
was having to play catch-up on this and related
operational matters. When road surface conditions were
good and tire temperatures were optimal, LC had a very
real punch that could provide acceleration from zero to
100 km/h in under 2.6 seconds, and BAR-Honda was
probably ahead of the others in this regard. There were
other teams, however, that took the lead with a rearward
weight distribution package, which was advantageous in
starts, or with their knowhow on using tires, and it was
regrettable that the BAR-Honda teams starting
performance did not raise the teams overall strength into
the top rank.
Clutch
paddle position
Clutch paddle map
First paddle
release
Second paddle
release
Disengagement
Engagement
Start
Fig. 8
Full engagement
Double paddle procedure
Time
200
Clutch actuator position

6
Input shaft torque
Engine rev limit

Engine speed
Gear position
Rear wheel speed
10000
200
Input shaft speed

Start
Front wheel speed

282
284
Time: 279.550 sec
Clutch dump RS track data (2005)
stage method of pressing the throttle pedal, and

continued the use of dump RS (Fig. 9) that realized
controllability on a par with the real rev limit from the
standby state to the standing start while seeking to
simplify driver operations.
5.5. Partial Clutch Engagement RS with DPC Pressure
Control
In 2006, the 3-L V10 engine used until then was
prohibited, and the use of 2.4-L V8 engines was
required. This meant that engine torque was diminished
by 20% or more, so that the clutch dump system could
effectively no longer be used. Figure 10 shows the
relationship between torque in a V8 engine and the
required torque calculated from the coefficient of friction
between the road surface and the tires. The on a dry
road surface ranges from 1.6 to about 1.9, so that in
order to satisfy the relationship of Engine Torque > RS
Required Torque, an engine speed of from 8000 to 9000
rpm or more is necessary. With the V10 engine, the
relationship between required torque and engine torque
was such that = 1.0, as shown in Fig. 10, allowing
the dump system at 5000 rpm or slower. However, the
V8 engine was required to provide torque at higher
Throttle %
300
250
200.0
180.0
5
10
160.0
15
200
140.0
150
120.0
100
100.0
Tire
50
80.0
= 1.7
60.0
0
Tire
-50
= 1.0
-100
19000
18000
17000
16000
198
15000
Fig. 10
14000
Engine speed [rpm]
13000
12000
11000
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
in order to resolve the weakness of pull clutch position

control. This method involved holding the engine in
ready state at the rather high speed of 11000 rpm or
more, fully engaging the clutch when the clutch paddle
was released immediately after racing starts, and thus
forcing the wheel spin to be reduced by applying the
engine rev limitation for extremely low speed, 5000 rpm
or less. Figure 9 shows actual track data for clutch
dump RS as it was in 2005.
Although the wheel spin prevents the initial
acceleration from rising as high as when a partial clutch
engagement start succeeds and there is no wheel spin,
this method realized highly repeatable starts because the
amount of wheel spin generated was the same every
time. The only indeterminate factors in starting were the
timing for release of the clutch paddle and the tire
temperatures. There were primarily two challenging
aspects of control. One was that the engine speed had
to be kept at a low speed, even lower than its idling
speed, even though for just one second or less, and
satisfactory ignition conditions and hydraulic pressure
had to be assured during that time. The other aspect was
that the regulations specified that the target clutch
position (or the hydraulic pressure) determined by the
clutch paddle and the actual clutch position (or the
hydraulic pressure) are allowed no more than a 50 msec
delay. If full engagement occurs instantaneously
according to the paddle demand, the shock torque
generated in the drivetrain would have easily exceeded
the maximum allowable torque for gearboxes of that
time. That was why the clutch control was tuned with
great care with the order of several milliseconds so that
the maximum transfer torque could be reduced as much
as possible within the allowed 50 msec time period.
There were no regulations relating to engine rev limit
control in 2004, and drivers were permitted to keep their
throttles wide open even before the start. In 2005,
however, use of the rev limit while in the standby state
was prohibited. BAR-Honda proposed an engine torque
trimming map exclusively for use in starting and a two-
Engine torque [Nm]
Fig. 9
283
Car speed [km/h]
Wheel speed [km/h]
400
10
Gear position
Clutch actuator position [mm]
20000
Engine speed [rpm]
20
25
30
35
45
55
40.0
65
20.0
75
0.0
93
Mu=1.7
Mu=1.0
Car Speed
V8 engine torque and required torque for RS
200
Clutch actuator pressure
0
Input shaft torque
Rear wheel speed
Engine rev limit
Engine speed
Gear position
10000
200
Input shaft speed

Start
Front wheel speed
0
88
Fig. 11
89
90
Wheel speed [km/h]
200
Gear position
400
Clutch actuator pressure [bar]
Engine speed [rpm]
20000
91
Time: 87.623 sec
Clutch partial engagement RS with DPC track data (2006)
engine speeds, also making it necessary to alter the RS

strategy on the assumption that the double paddle system
with partial clutch engagement would be back in use.
Hondas unique, newly developed direct push clutch
(DPC) (1) was adopted as a measure to enhance clutch
torque controllability. This clutch resolved the issue of
thermal expansion and wear state of the disk causing the
clamp load to change, which was a weakness of the pull
clutch. DPC used a system with a hydraulic actuator that
pushed directly against the disk for the clutch
connection. This succeeded in obtaining clamp load with
generally linear characteristics with respect to the
hydraulic pressure, and the transfer torque was stable.
The clamp load was already known, so where race
operations was estimating the tire as described earlier,
now it also estimated the clutch disk , further enhancing
the precision of paddle map selection in racing starts.
Torque/rev control was also introduced to combine
torque control with the conventional rev control method
of engine control. In this method, control of the absolute
value of torque using the engine formed the primary
means for control, and it would switch over to rev
control only when the engine speed reached a specified
upper or lower limit. The purposes were to stabilize
clutch transfer torque by optimizing engine torque
output, and to create conditions facilitating recovery from
the wheel spin by holding torque at a minimum even
when wheel spin occurred.
Figure 11 shows actual track data for partial clutch
engagement RS by means of DPC pressure control. The
double paddle method keeps DPC pressure (clutch
actuator pressure) demand at a certain level, thus holding
the transfer torque (input shaft torque) at a constant value
as close as possible to an ideal value in the state of
partial clutch engagement. This realizes an ideal standing
start without causing wheel spin.
The development of this system up to 2007 generated
knowhow that was applied wherever possible in the area
of tuning in the FIA-common ECU that was made a
requirement in 2008.
6. Formula One Control System

Development Environment
When third-era Formula One activities got started, the
focus was initially on testing in the actual vehicle, and
the software development for control systems was also
done by hand coding, working from specifications based
on flowcharts. However, the project for development of
Honda made chassis control systems, begun in 2002 to
enhance development efficiency, was the start of a
gradual shift to development based on the use of
MATLAB Simulink and Stateflow, and from that point
on, technology for autocoding was actively incorporated
into the development effort. Development of software for
control on the chassis side was gradually converted to
autocoding with the LC in 2003, and in 2006,
development of the chassis-side control unit was
switched from Pi-Sigma MCU, which had been used up
to that time, to the ATHENA system, which was made
by Honda and used for engine control. With this,
development on the chassis side was converted 100% to
autocoding. Autocoding technology was incorporated on
the engine side, as well, but ATHENA had already been
employed continuously since 2000, and for that and other
reasons, the use of autocoding did not progress beyond
partial adoption for use on relatively new systems.
There were also advantages to the use of Simulink
and Stateflow. Offline simulation was performed across
a wide range on both the chassis and engine sides, from
the debugging level to the development of new control
algorithms, and the Hardware In the Loop Simulation
(HILS) environment, which included a real ECU, was
also actively employed. In fact, the LC and RS
introduced from 2003, as well as the wheel slip feedback
TC and OC introduced in 2006, were developed by
simulation using a chassis model that included a
drivetrain.
The development of third-era Formula One control
systems sought to successfully develop systems with
limited opportunities for testing in actual vehicles and
199
to deploy them as quickly as possible in racing

machines. Simulation technology can be said to have
played a central role in this effort.
7. Conclusion
Throughout the third-era Formula One activities,
there was an impetus to move beyond traction control
that dealt simply with acceleration, and to incorporate
traction control for limit ranges in a variety of different
circumstances including deceleration and standing starts.
This yielded results and many new insights in connection
with (1) the adoption of torque drive system that
converts throttle pedal input to engine torque demand,
(2) the adoption of a torque interface with chassis control
systems, (3) wheel slip control that directly controls
wheel slip, (4) coordination with clutch control, and so
on.
All this brought a keen awareness that not only the
powertrain field, naturally enough, but also the chassis
field and coordination with the driver were particularly
important for traction control. It also became apparent
that alternative testing approaches, to include virtual
methods, were of importance in developing control
systems for use under limit conditions that are not
readily reproducible in the actual machine. It is to be
hoped that the experience gained on these two points will
also be put to active use in developing mass-production
models.
Reference
(1) Kishi, T., Nagatoshi, Y., Nakamura, H., Fukao, Y.:
Development of Direct Push Clutch Control during
Honda Formula One Third Era, Honda R&D Technical
Review 2009, F1 Special (The Third Era Activities),
p. 207-210
Author
Kazuharu KIDERA
200
Yoichiro FUKAO
Tatsuya ITO
Development of Gearbox Control during Honda

Formula One Third Era
Takayuki KISHI*
Hiromasa NAKAMURA*
Yoshiki NAGATOSHI*
Yoichiro FUKAO*
ABSTRACT
One main development topic for the Honda Formula One third era gearbox was the seamless gearbox (Quick Shift),
which reduced the interruption of acceleration to zero.
Gearshift control provided a way to control dog damage and excessive torque, which had been issues affecting
Quick Shift feasibility, and this enabled the team to realize first deployment of the Quick Shift in racing, ahead of
the other teams. The search for even smoother gear shifting continued after that, and led to the establishment of a
mechanism that would reduce to a minimum the torque fluctuation resulting from correlation of the torsional vibration
of the driveshaft and the inertial torque generated during gear shifts.
1. Introduction
The structure of the gearbox in the Honda Formula
One third era was the seven-gear sequential gearshift of
the constant mesh type widely used on motorcycles. This
was electronically controlled using a hydraulic pressure
actuator, and gearshift operations were carried out using
a paddle attached to the steering wheel. Given this
structure, shifting gears requires the engine torque to be
first reduced to nearly zero. Reducing this gearshift time
to the ultimate extent would contribute to increased
racing competitiveness, and development of such control
was therefore begun. The development effort sought, on
the upshift side, reduction of acceleration loss,
stabilization of wheel torque, and limitation of torque
oscillation when shifting gears. On the downshift side,
it sought reduction of minus torque when shifting gears,
and reduction of idle running time. In this way,
measures were taken to achieve a balance of hardware
evolution combined with performance enhancement and
establishment of reliability.
The Quick Shift mechanism, development of which
started in 2004, was an innovative mechanism that
enabled upshifting while still maintaining engine torque
at the maximum level. It brought about a major evolution
in performance when shifting gears. However, it was
necessary to develop new controls to take advantage of
this mechanism. Even before the introduction of the
Quick Shift, the demand for better gearshift performance
had been addressed by enhancing control equipment and
changing the controlled subjects. With the seamless
gearshift, however, there was a likelihood that significant
damage will be made to the dog or the gears depending

on the gearshift timing. Control that would eliminate this
likelihood and assure a prescribed level of reliability was
an absolute condition for the feasibility of this new
mechanism. This paper will introduce the main controls
that established the reliability of the Quick Shift and the
control methods that realized smooth upshifting and
downshifting.
2.1. Establishment of Quick Shift Reliability
The Quick Shift is a mechanism that enables
upshifting without lowering the engine torque. However,
when one dog collided with another edge to edge, the
shock was significant, and sometimes the dog edge
would wear so that it could no longer provide driving
power. If the next gear provided driving power when
there was dog delta speed, an impact torque was also
generated that resulted in the gear exceeding its torque
limit so that the gear was damaged. The following two
points were therefore defined as control objectives:
(1) Zero collision of one dog with another edge to edge
(2) Zero occurrence of gear over-torque
2.2. Realization of Smooth Upshifting and Downshifting
The inertial torque that occurred in gear shifting
when shifting up or down could result in excessive wheel
slip, rear tire lock-up, deterioration in drivability due to
torque oscillation after gearshifts, and other such issues.
The following objectives were defined to address these
issues:

201
(1) Upshifting
Maintain wheel driving torque
Limit oscillations in engine speed and in torque after
gearshifts
(2) Downshifting
Limit minus torque after gearshifts
Limit oscillations in engine speed and in torque after
gearshifts
3. System Overview
Figure 1 shows the gearbox system configuration.
The gear position is determined by the angle of the
gearshift barrel, and the gear shifting is executed by
input on the gearshift switch attached to the steering
wheel. While shifting gears, the system controls the gear
position and simultaneously requests the clutch transfer
torque, engine speed, and engine torque.
The clutch operates according to analog input from
the clutch paddle attached to the steering wheel. In
addition to engaging and disengaging, the amount of
clutch disengagement is also controlled according to
gearshift and anti-stall requests.
The gearshift and clutch actuators are driven by 200bar hydraulic pressure controlled by a Moog valve. (The
control cycle is 1 ms.)
4. Establishment of Quick Shift Reliability

4.1. Avoiding Edge to Edge Collision of One Dog with
Another
An overview of the control should include the
development of Trigger Shift control. The Trigger Shift
control uses rotation sensors for the drive shaft
(layshaft), which is connected directly to the engine, and
for the driven shaft (main shaft) to constantly calculate
the angle of the gear and of the dog ring, memorize the
phase difference between the dog ring and the gear when
the gear is engaged and driven, and calculate the timing
for driving the gearshift barrel so that it can occur in the

interval between dogs at the next gearshift. Figure 2
shows a conceptual image of the target position.
The Trigger Shift learns the phase differences in the
below order from (1) to (4), and drives the shift barrel
accordingly.
(1) The rotation angle of the dog ring is calculated from the
main shaft rotation.
(2) The rotation angles of gears 1 to 7 are calculated from
the layshaft rotation.
(3) The angle difference between (1) and (2) (phase
difference (1)-(2)) for each gear when engaged with the
dog ring is calculated and memorized.
(4) The angle up to the point where the target phase
difference is reached is predicted from the barrel
movement time and the dog delta speed, and the barrel
is driven accordingly to shift gears.
The above describes the basic logic involved.
Figure 3 shows a block diagram of the Trigger Shift
control.
The following two data settings are necessary to
implement this control:
(1) Response time until the barrel reaches the in-gear position
(2) Target position between dogs (upshift and downshift)
These values are determined by the structure, and
they do not require fine tuning while the car is actually
in operation. The control is efficient also in helping to
ensure reliability by reducing the likelihood of
operational errors during racing, and the like.
Figure 4 shows example Trigger Shift race results
in graph form from the ninth competition in 2006, the
Canada Grand Prix. The graph summarizes the in-gear
positions between dogs when shifting up from 3rd to 4th
gear. The triangle symbols ( ) in the figure represent
the positions targeted at by the Trigger Shift system. The
open circle symbols ( ) in the figure represent the actual
in-gear positions. The figure indicates that the system
generally functioned with in-gear positions as targeted
Dog ring
Dog delta
speed
Shift
request
Gear
dog
Enough in
gear depth
to drive
Paddle
request
Target position
+
Pressure
Rotary actuator
Actual position
Current
Trigger shift
Engine
Target pressure
+
F/B
controller
Current
Actual pressure
Demand
arbitration
Fig. 2
Layshaft
pulse
Clutch
Main shaft
Dog ring
Main shaft
pulse
Gear angle calc

Dog ring
angle calc
Layshaft
Barrel delay
Pressure
Pressure voltage
15 deg
Start of in-gear
position
Upshift target pos

Up shift
Downshift target pos
Phase
calculation
Engine speed
Engine torque
Target
phase
prediction
+
+
Target
phase
Offset
=
=
1 dog
(72 deg)
Engine controller
Fig. 1
Memorize
drive side
engaging
phase
Gear dog
Anti-stall
Speed, torque
request
Image of trigger shift
Pot sensor Fork

Pressure
Layshaft, Main shaft speed
Clutch
adaptation
Paddle
map
Barrel
Sensor voltage
Ratchet
lock/unlock
Shift
sequencer
Target phase
Moog valve
Current
F/B controller
System configuration
Fig. 3
202
Trigger shift control system
Barrel
drive
enable
(trigger
signal)
for, and did so repeatedly with good reproducibility.

The Trigger Shift performance limits will also be
described here. If the barrel response is taken to be
constant, then the dogs engage with a shallower
intermeshing when the dog delta speed is greater, so that
the dogs are more likely to collide edge to edge with
each other. Furthermore, even when the intermeshing
depth is the same, there will be greater damage to the
dog edge when the dog delta speed is greater. Figure 5
shows the results of testing on wear limits by
intermeshing gear depth and dog delta speed. The
maximum value for dog delta speed between dogs is
determined from the upshift speeds and the gear ratios
in Formula One racing to be approximately 2000 rpm.
The intermeshing gear depth necessary to reduce the
occurrence of wear in this case, as shown in Fig. 5, is
1.5 mm. By compiling fork speeds in all the 2006 races,
it was found that the four sigma point of the distribution
was 1.54 mm/ms, and this was taken as the worst case.
Given this, the in-gear position capable of avoiding dogto-dog collision could be expressed in terms of time as
1.5/1.54 1 ms, which is the necessary safety margin
when the dog delta speed is 2000 rpm.
In actuality, as shown in Fig. 6, it is necessary to
take into account variations in the target position due to
barrel response times and dog delta speed changes.
Taking these factors into consideration, together with the
fact that the dogs are placed every 72 degrees, the limit
dog delta speed at which the Trigger Shift can assure
performance is expressed by formula (1):
In this formula, the value will be low for 2000 rpm.

The recommended target position will be 32 degrees at
the middle of the variation.
4.2. Avoiding Gear Over-Torque
Adjustment of the amount of clutch disengagement
when shifting up so that the clutch transfer torque will
be between the engine torque and the gear limit torque
allows the impact torque to be absorbed by the clutch.
However, the torque characteristics of the clutch lacked
stability while the car was in operation due to changes
in the coefficient of friction and the clamp load (the
load pushing against the disk) resulting from disk wear.
Therefore, stable absorption of the impact torque could
not be realized with a fixed amount of clutch
disengagement, so an auxiliary control of the clutch
disengagement amount called clutch adaptation was
applied. This control is focused on the amount of clutch
slippage while shifting up, and it applies cumulative
adjustments to the amount of clutch disengagement so
that it will match the target slip amount. The clutch
slippage amount while shifting up is calculated, and if
it has not slipped more than the target amount, then the
clutch will be adjusted in the direction of disengagement
for the next shift up. If it has slipped more than the
target amount, then it will be adjusted in the direction
of tightening the clutch. Figure 7 shows a block diagram
of the clutch adaptation control.
Barrel response offset +0.66/-0.61 ms
(Dog pitch) - (Dog delta speed variation) / (Control

cycle error) + (Barrel speed error) + (Safety margin)
(1)
Delta speed changing +15/-7 deg

Job timing 1 ms
OS - Apps timing -1 ms
60
Phase [deg]
Target
Actual
-2.61 ms -7 deg
40
+1.66 ms +15 deg
20
0
Safety margin 1 ms
-20 0
100
200
300
Number of gearshifts
Fig. 4
Target
Trigger shift result (2006 Canada GP)
72 deg
Delta speed [rpm]
2500
y = 1160x2 - 360x
Fig. 6
Control tolerance of trigger shift
2000
Clutch torque demand
1500
1000
Dog
Safe
Dog
Wear
Fig. 5
Target integration slip

Engine speed
Adjust
IPS rpm
Actual integration slip
500
0.5
Engine speed
1.0
1.5
In-gear depth [mm]
Compensated clutch
torque demand
2.0
IPS rpm
Relation of dog delta speed and in-gear depth
203
Fig. 7
++
+
Error
+
1/Z
Configuration of clutch adaptation control
Gear
graph shows the rate at which the engine speed is

dropping, while the oscillation level is plotted on the
vertical axis. It is apparent that the oscillation increases
whether the slope of the engine speed is steep (part A
in the figure) or too shallow (part B in the figure), and
that it reaches its minimum value at -90 rpm/ms.
Figure 13 shows the summarized results of the
correlation of the oscillation level and the delta engine
speed with the next gear when the clutch is engaged. The
graph has the difference between the engine speed when
the clutch slippage ends and the engine speed at the next
gear plotted on the horizontal axis, and the oscillation
level plotted on the vertical axis. It is apparent from
these results that engine speed oscillation when shifting
up depends on both the engine speed when the clutch is
engaged and the engine speed at the next gear. It is also
apparent that reducing the differences between these
speeds can limit oscillation, as well.
4th
3rd
Engine
speed
Control shock torque

by adjusting amount
of clutch slip
Layshaft
speed
Layshaft torque
Rear speed
Fig. 8
Example of upshift behavior

3-4 up peak torque
800
Peak torque [Nm]
Without adaptation
600
Engine
speed
400
With adaptation
Post shift oscillation
200
0
10
20
30
40
50
60
Number of Gearshift
Rear
speed
Clutch adaptation result
Figure 8 shows the waveforms for actual shifts up

from 3rd to 4th gear that absorb impact torque by
making the clutch slip. Figure 9 shows a comparison
of impact torque in the course of driving approximately
10 laps. It is apparent from this figure that when clutch
adaptation is applied, the impact torque is held in the
vicinity of 400 Nm with stability as intended.
5. Realizing Smooth Up and Down Shifting
Wheel slip
Front
speed
Fig. 10
Upshift oscillation issue
Engine speed
5.1. Upshift
When shifting up with Quick Shift, as shown in Fig.
10, the inertial torque when shifting gears results in the
wheel driving torque that exceeds the tire limits and
induces excess wheel slipping as well as oscillation in
the engine speed. The torque oscillation following a gear
shift can lead to impaired drivability. Therefore, the
following two points take on importance for upshift
control:
(1) Limiting engine speed oscillation
(2) Maintaining optimal wheel driving torque
In order to quantitatively evaluate the engine speed
oscillation, therefore, the oscillation level was
represented by the deviation from a logical value for
engine speed in the next gear integrated and depicted as
an area, as shown in Fig. 11.
Figure 12 shows the summarized results of the
correlation of this oscillation level with the slope when
the engine speed is dropping. The horizontal axis of the
204
Target gear
engine speed
Fig. 11
Definition of oscillation quantity

100
Minimum at -90 rpm/ms
80
(A)
(B)
60
40
Oscillation
Fig. 9
20
0
-200
-150
-100
-50
Engine speed drop rate [rpm/ms]
Fig. 12
Oscillation and engine speed drop rate
Analysis based on the above results showed that the

oscillation level was smaller in the vicinity of the drop
in engine speed at the rate of around -90 rpm/ms, which
largely coincided with the torsional oscillation cycle in
the drivetrain approximately 20 ms after the engine
speed began to drop. Therefore, as shown in Fig. 14, it
was found that oscillation could be limited by
synchronizing it with the torsional oscillation of the
drivetrain so that the clutch slippage is stopped at the
point when the speed of the layshaft (which is on the
other side of the clutch from the engine) is the same as
the speed of the engine at the next gear, thus canceling
out the torsional oscillation. It is necessary, therefore, to
have the period of time until the engine speed reaches
the next gear speed match with the torsional oscillation
cycle, and this necessitates appropriate control of the
engine torque reduction and the clutch slippage amount.
Control of wheel driving torque will next be
described. The torque when shifting gears is subject to
the influence of the inertial torque generated when the
engine speed is reduced to the speed at the next gear. It
is expressed by the following formula.
(2)
TLay = TENG IENG d /dt
(TLay: Layshaft torque. TENG: Engine torque. IENG: Engine

inertia. : Engine angular velocity)
In the above formula (2), the term IENGd/dt normally
had a negative value when shifting up, so that it would
be added on top of T ENG. This increased the wheel
driving torque, which would induce excessive wheel
slipping. It was necessary, therefore, to control TLay so
as to keep the wheel torque constant, and the target TTarget
would then be expressed by formula (3):
The logic that realizes these aims can be described

as follows. The target time for the engine speed to drop
to the next gear speed is replaced by the downward slope
of the target engine speed. The inertial torque IENGd/dt
is defined by this slope of the target engine speed. In
order to realize T Target , therefore, T Lay + I ENGd /dt is
supplied as the engine torque reduction demand during
the gear shift.
TTarget is the clutch transfer torque TClutch, and where
clutch adaptation had conventionally modified the clutch
slippage amount, here instead it modified the slope of
the engine speed by applying a correction quantity
defined as a coefficient equivalent to the clutch
coefficient of friction , thereby enhancing the accuracy
of TClutch. Figure 15 shows a conceptual image of these
elements of control.
Figure 16 shows the superposed actual upshift
waveforms for the conventional upshift and the upshift
with the new control applied. A smooth upshift was
realized by the optimization of the slope of the engine
speed and the engine torque reduction.
5.2. Downshift
Figure 17 shows the waveforms for a conventional
downshift. The issues when downshifting are to limit the
locking of the rear tires and to stabilize vehicle behavior
when braking. The mechanism by which rear tire lock
occurs during downshifting is described below.
(1) The dog delta speed immediately before being driven by
the dogs is large.
(2) The dogs engagement with each other leads to rapid
convergence of the dog delta speed.
(3) Inertial torque is generated, this applies torsion to the
Speed rate optimization
Optimization of speed error
at clutch engagement
Engine speed
TTarget = TENG Previous Ratio / Next Ratio
(3)
Layshaft speed
Target gear engine speed
100
Oscillation
80
Torque reduction to keep

wheel torque constant
Engine torque
demand
60
Engine torque
40
Wheel torque
Fig. 15
20
Target shock torque

Inertia torque
Layshaft torque
Image of smooth upshift control
20000
Engine speed (new control)
500
1000
1500
2000
18000
Delta engine speed at clutch engage [rpm]
Engine speed
(conventional)
16000
Fig. 13
Oscillation and speed error at clutch engage
14000
4000
12000
Engine speed
Less oscillation
Driving torque
(conventional)
Constant driving torque
3000
2000
1000
Driving torque (new control)
Target gear speed
Layshaft Speed
0
36.3
Fig. 14
Upshift example
Fig. 16
205
36.4
36.5
Upshift comparison
36.

Gear
(2) While being driven by the dogs, make the gear speed
lower than the dog ring speed.
In (1), the clutch is disengaged and the engine is
blipped to raise its speed. Then, the clutch is engaged,
with timing matched to when dog driving occurs. The
clutch torque raises the gear speed, and the dog delta
speeds converge. In (2), the engine speed demand is set
lower than the next gear speed to control the gear speed
so that it will be lower than the dog ring speed.
Figure 18 shows a conceptual diagram of this
control.
In the diagram, the barrel-driven timing (A) and the
clutch disengagement initiation timing (A) are
determined by the engine blip response. As noted above,
it is important that convergence of the dog delta speed
between dogs should occur gently, which is why the
engine torque and speed instruction values are defined
so as not to decrease the gear speed in (B). Also, as
shown in (C), the clutch transfer torque is controlled
appropriately so that the gear speed increases gradually,
thus relaxing the impact torque when in gear.
Figure 19 shows the results of comparison between
these controls and conventional control. Figure 19
presents the waveforms for the conventional downshift
(a) and the downshift (b) with the control described
above. Comparison of the torque behavior and the
oscillation in the engine speed after downshifting
confirmed that these are reduced when the new control
is applied.
Engine speed
Layshaft speed
Layshaft torque
(4)
Dog ring speed
(3)
(1)
(5)
(2)
Front speed
Rear speed
Gear speed
Fig. 17
Example of bad downshift

Downshift phase
Gear
(A)
Engine speed
(B)
Dog speed
(C)
Gear speed
Blip torque
0 Nm
(A)
Clutch
Fig. 18
Downshift control concept
6. Conclusion
Gear
The Quick Shift that was applied in racing from 2005

had enhancements made to the Trigger Shift in 2006, and
it achieved a record of zero retirement for gearbox
reasons in races during the 2007 and 2008 seasons.
The new control was applied from 2007, and the
enhancement of gearshifting yielded a zero level of
related driver complaints.
Engine speed
200
200
100
100
Layshaft torque
-100
-100
34
35
194
(a) Standard downshift
Fig. 19
Author
195
(b) New downshift strategy
Downshift comparison
driveshaft in the direction of deceleration, and oscillation

occurs.
(4) Engine speed is increased and the minus torque exceeds
the rear tire drive torque limit.
(5) The rear tires lock.
The above item numbers correspond to the numbering
in Fig. 17.
The major reason for deterioration of downshifting
is the dog delta speed when being driven by the dogs
while in gear. A smoother downshift was realized by the
following kind of control of this dog delta speed.
(1) Make the dogs drive while the gear speed is increasing.
(Cancel out the inertial torque while being driven by the
dogs.)
Takayuki KISHI
Yoichiro FUKAO
206
Yoshiki NAGATOSHI
Hiromasa NAKAMURA
Development of Direct Push Clutch Control during

Honda Formula One Third Era
Takayuki KISHI*
Hiromasa NAKAMURA*
Yoshiki NAGATOSHI*
Yoichiro FUKAO*
ABSTRACT
One main development topic for the Formula One third era clutch was the direct push clutch (DPC), which was
developed in order to increase the accuracy of transfer torque.
Conventional pull clutch position control could not readily stabilize the amount of torque transfer during standing
starts, and therefore pressure control of the DPC was used in order to stabilize the clamp load. The non-linear
characteristics of the hardware posed some issues with the hydraulic pressure response. However, these issues were
addressed by control, and as a result response that had no issues in practical terms was able to be assured and the
method was successfully deployed in racing.
1. Introduction
In the standing start range, the main focus of
development has been to drastically curb wheel spin at
the initial stage by enhancing the accuracy of clutch
transfer torque estimation and raising the performance of
the partial clutch engagement start. For clutches with the
conventional structure, both the clamp load and the
coefficient of friction were indeterminate factors,
making it a considerable challenge to estimate torque.
The DPC with its control of clutch transfer torque by
means of hydraulic pressure was developed in 2005, but
the hardware and the control were integrated as one, so
that assuring the requisite pressure control performance
became an important subject for development as had
been experienced in the case of the gear box
development. This paper introduces the DPC pressure
control system.
In the development of DPC, the adoption of a device
that uses hydraulic pressure to determine the clamp load
increased the accuracy of estimation of transfer torque.
However, the relationship between the hydraulic pressure
and the length the piston travels in the action from the
fully disengaged state to the bite point where pressure
against the disk begins is different from that from the

bite point to full engagement. As a result, pressure
controllability became an issue. Controllability was also
diminished by the hydraulic pressure vibration resulting
from sticking and slipping of the piston that occur when
the hydraulic pressure is raised from the fully disengaged
state. In standing starts, speedy and accurate delivery of
the amount of torque that results in the tires exercising
the maximum gripping force is sought. The clutch
response and convergence that will transfer the necessary
torque are important, and the following objectives were
defined regarding the controllability of the hydraulic
pressure:
(1) Hydraulic pressure response of 30 ms or less up to the
target value
(2) Hydraulic pressure overshoot of 1 bar or less
3. Overview of the DPC

Figure 1 shows the DPC configuration. The force
applied to the piston is conveyed to the pusher in a lever
structure with the fulcrum ring as the fulcrum and the
pressure plate as the point of action. Thus the clamp load
is transferred to the clutch disk.
Consequently, the DPC differs from the conventional
pull clutch in that the clamp load can be directly
controlled using hydraulic pressure.

207
DPC characteristic
Piston
200
Pressure plate
Pusher
Crankshaft
Layshaft
Lever
Fulcrum ring
Full engage
120
Bite point
80
(A)
40
Worn
condition
(A)
0
4
Fig. 1
Full
disengage
New
condition
160
Configuration of direct push clutch
Clutch travel [mm]
Full disengage
4. Establishing Pressure Controllability of

the DPC
Lever
4.1. Issues in Pressure Controllability of the DPC

Figure 2 shows the DPC stroke and hydraulic
pressure characteristics as well as the clutch state, which
are factors diminishing its pressure controllability.
In Fig. 2, the area (A) extends from the state of full
clutch disengagement to the bite point where the clutch
begins to engage. The amount of actuator movement is
large relative to the amount of change in the hydraulic
pressure.
When the wear of the clutch disk progresses, the
distance that the pusher moves from the fully disengaged
position until it hits the pressure plate grows longer. The
area (A) in the figure expands to become (A).
An area like this became a dead zone when under
pressure control, interfering with the response, and not
only did it diminish the effectiveness of gearshift,
standing start, anti-stall, and other such types of control
that require rapid clutch control, but it was unable to
satisfy the limit on response time (50 ms) before
reaching the target value as defined under FIA
regulations. Therefore, this was very likely to result in
a violation of the regulations.
In order to address these issues and achieve the target
value explained earlier, the following types of measures
were taken with regard to DPC pressure control.
4.2. Substance of DPC Pressure Control
4.2.1. Variable P-term gain, variable low-pass filter
In order to achieve the enhanced hydraulic pressure
response described earlier, the variable P-term gain and
variable low-pass filter (LPF) shown in the conceptual
diagram in Fig. 3 were employed. This gain and LPF
cutoff frequency are hydraulic pressure variables, and the
aim was to raise the P-term and LPF cutoff frequency
in transient states of the low-response area in the dead
zone in order to achieve better hydraulic pressure
response.
4.2.2. Simple state predictive P-term control
With variable P-term gain and LPF, the switching
hydraulic pressure for the gain and cutoff frequency
208
Pressure plate
Disk
Full disengage - Bite point
Bite point - Full engage
Fig. 2
DPC pressure-stroke characteristic

Pressure target
Threshold of
replacing gain
Actual
pressure
Error
Press FB P-gain
Press FB LPF
Use strong P-gain

and LPF to
enhance response
Controller
output
Fig. 3
Image of variable P control
depends on the hydraulic pressure in the dead zone. As

explained earlier, the hydraulic pressure in the dead zone
changes according to the wear of the clutch, and
robustness could not be assured by variable control
alone.
Therefore, the hydraulic pressure that follows after
the time of non-response by hydraulic pressure (waste
time) was estimated from the slope. Predictive control
was implemented as shown below in order to use this
predicted value as feedback to raise the P-term gain
while reducing overshooting. Figure 4 shows the
conceptual diagram.
(1) The time of non-response by hydraulic pressure is defined
as the specified time. The hydraulic pressure following
the specified time is estimated from the slope of the
current hydraulic pressure.
(2) The amount of operation is calculated according to the
difference between the predicted value and the target
value.
(3) If the predicted value is greater than the target and the
current value is smaller than the target, then the difference
is considered to be zero and the amount of operation zero.
(4) If the predicted value is greater than the target and the
current value is also greater than the target, then the
amount of operation is calculated according to whichever
has the least difference relative to the target.
4.2.3. Step limit moving average

Limits were imposed on the size of the steps in actual
clutch pressure, taking the effective response frequency
of the actual hardware into account. This limited the
delay and resulted in the occurrence of significant
damping to occur. This damped the vibration generated
by sticking and slipping of the piston. Figure 5 shows
the conceptual diagram.
4.2.4. Additive I-term control
Overshooting and undershooting were determined by
the deviation from the target value and the slope of the
hydraulic pressure, and the I-term was converted into a
step form to increase the convergence. Figure 6 shows
the conceptual diagram.
The filter value discussed earlier is fed back to the
integral control. When this filter value deviates from the
target value by more than a certain figure and the slope
remains small for a set period of time, it is determined
to be overshooting. The I-term is then reduced regardless
of the feedback so that convergence with the target value
will occur more quickly.
4.3. Results
Figure 7 shows the pressure controller that was
Pressure
Filtered pressure
Error threshold (high)
(3)
Predicted
pressure
Error
term
Error threshold (low)

Target pressure
(2)
(4)
Time
Integral term
Actual
pressure
Target
pressure
Increase integral term

to enhance conversion
(1)
Dead time
Fig. 4
Fig. 6
Predictive P-term control
Additive I-Term control
Filter
Target pressure
Actual
pressure
Step
limiter
SMA
Pressure (raw)
Step
limiter
2nd order
IIR LPF
Predictor
2nd order
IIR
SMA
Pressure (filtered)
Filtered pressure
Step filtered
Target pressure
Controller
Pressure target
Step filtered pressure
Feedforward
map
Pressure (filtered)
Predictor
Variable
P-term
Transient
state
Controller output
Constant
P-term
Steady
state
Filtered pressure
Actual pressure
Fig. 5
+
I-term
Additive
I-term
Fig. 7
Step limit moving average
209
Configuration of DPC pressure controller
ultimately realized.
Figure 8 shows a comparison between standard PID
control and the control described above. The comparison
was carried out during step response assuming a standing
start, when accuracy and speed of hydraulic pressure
Stroke
27ms
response are required. Comparison was also carried out

with 3.0 mm of clutch wear, which corresponds to the
amount of wear from a single race, since response can
be assumed to diminish when clutch wear results in an
expanded dead zone.
The hydraulic pressure response time was
successfully reduced, without overshooting, by 44%
when there was no wear and by 71% when there was
wear. Even when wear had occurred, the diminishment
of standing start performance was kept to a minimum.
5. Conclusion
Target pressure
Actual pressure
(a) Standard PID (no wear)
12ms
The establishment of pressure control of the DPC

enabled an increase in the accuracy of clamp load
(transfer torque) control.
The DPC was employed in racing from 2006, and the
implementation of a partial clutch engagement start
throughout the entirety of the standing start enabled
realization of slip starts without any torque feedback
control. This yielded the advantage of a maximum 0.46second reduction in 0-100 km/h acceleration (an 11-m
advantage in terms of distance).
(b) New control (no wear)
Stroke
105ms
Author
Target pressure
Actual pressure
(c) Standard PID (3.0 mm wear)
30ms
Takayuki KISHI
(d) New control (3.0 mm wear)
Fig. 8
Comparison of pressure control
Yoichiro FUKAO
210
Yoshiki NAGATOSHI
Hiromasa NAKAMURA
Descriptions of Electronic Equipment Technologies
Development of Electronic Control System for

Formula One
Kenichiro ISHII*
Masaki NEGORO*
Toshiyuki NISHIDA*
Masataka YOSHIDA*
Kohei TOSHIMA*
Yutaka MARUYAMA*
ABSTRACT
The application of electronic control systems has been rapidly increasing in Formula One cars as well as in other
vehicles. System performance is a crucial element in conducting precision control and measurement of cars.
Starting with the 2006 season, the Honda works has been working to apply original Honda systems not only in
the engine control system, as before, but in all vehicle electronic control systems.
In order to pursue higher performance at the same time as enhanced in-vehicle mountability in Formula One
electronic control systems, it is necessary to carry out optimization of these systems together with thoroughgoing
miniaturization. On-board systems have an electronic control unit (ECU) with integrated functionality linked by a highspeed network with units located in every part of the vehicle. High-speed telemetry is used to coordinate these with
the garage system in order to optimize systems.
The enhancement of unit performance by means of higher speed and greater precision contributes to heightened
controllability, and particularly to enhanced precision of driving force control and gearbox control. High-speed
communication also contributes to greater measurement performance in the pit.
1. Introduction
Formula One cars in recent years have incorporated
more than just the engine control and gearbox control
found previously. From mid-2001 to the 2007 seasons,
traction control and engine brake control have come to
be allowed in the regulations. Clutch control was not
prohibited up until 2007.
Traction control and gearbox control, in particular,
require precisely coordinated control of the engine and
chassis in real time. This has required advanced
computational capability and measurement performance.
Furthermore, Formula One cars demand aerodynamic
performance, so that there are limited on-board
installation locations for electronic control units in
Formula One cars. This means that systems require
greater compactness and lighter weight.
The functions of electronic control systems for
Formula One are generally for two purposes, for driving
and for analysis. In order to achieve compactness and
light weight, the latter functions are assigned as much
as possible to pit systems. In this way the former
functions can be optimized, and this was the basic
conceptual approach to the construction of these systems.
This article will provide an overview of Honda
Formula One electronic control systems, introduce the

issues involved in development of the systems, and
describe the contents of the development.
2. System History
Figure 1 shows the history of electronic control
systems for the third era of Honda Formula One.
Development of third-era Formula One systems began in
1998 with a view to applying integrated engine and
chassis systems to racing.
Under the BAR-Honda system from 2000 to 2005,
engine control systems were supplied to the team. From
2006, when Honda works joined in the competition, the
Honda full system was provided as a chassis integrated

212
Year
Team
2000
2001
2002
2003
2004
British American Racing - Honda
2005
2006
2007
2008
Honda Racing F1 Team
System
Chassis Control
Data Acquisition
Telemetry
BAR System
Engine Control
(PGM-FI)
Data Acquisition
Telemetry
Power Supply
Honda System
Athena
Fig. 1
FIA
system
Data
acquisition
History of Honda Formula One system
together the various units, sharing sensor data and fail

data to operate the systems. The ECU has the
functionality for integrated control of the engine, chassis,
and measurement. It acquires all information from the
chassis and conducts the operation of every device as
well as the processing of measurements. Each unit that
has been distributed can manage the measurement and
control functions for the different applications, and is
connected to the network as necessary. Of these, the
Front Data Acquisition (FDA) unit acquires data from
numerous sensors at the front of the chassis. Since it
processes signals used for real time control computations,
it is connected with the ECU by high-speed
communications at 10 Mbps.
In the pit, the client PC and the ECU are connected
management system. The system was named Athena

after the Greek goddess of wisdom and victory.
In 2008, it was made mandatory to apply
International Automobile Federation (FIA) standard
systems. The purpose was to standardize engine and
chassis control and to reduce costs. As a result, Hondas
in-house developed systems that were used in racing
were limited to the devices for engine use, sensors, and
data logging systems.
3. Configuration of Electronic Control

Systems (Athena)
3.1. Optimization of System Configuration (Functional
Integration and Distribution)
A cars aerodynamic performance, which determines
its body design, makes a very important contribution to
lap times. In order not to place constraints on the
freedom of the design, it is necessary to have system
configurations that are compact and have outstanding
mountability.
The main locations for mounting electronic unit were,
as shown in Fig. 2, beneath the radiator ducts on either
side of the chassis. The only other locations were near
the throttle pedal toward the front of the car and inside
the cockpit. The systems were therefore mounted in a
distributed configuration and were linked by a highspeed network in order to achieve a balance of
mountability and functionality.
Figure 3 shows the system configuration. The
network in the chassis is centered on the ECU and ties
Electronic power unit assembly
Electronic control unit assembly
Right layout
Left layout
Fig. 2
Electronic unit layout (front view)
Sensors
Sensor terminal (ST)
Engine
Gear Box
Differential gear
AC-DC (PBOX)
Front data acquisition
(FDA)
CAN
1 Mbps
Arcnet
10 Mbps
Sensors
CAN
1 Mbps
Ignition (CDI)
Electronic control unit
(ECU)
CAN
1 Mbps
Sensors
Steering wheel
Laser
2008system -
GPS
Camera
Logger
Tire temp/press
Video camera
Wing driver
SIO
2 Mbps
Telemetry
(HTX)
CAN
1 Mbps
CAN
1 Mbps
Team data logger

(TDL)
Ethernet
100 Mbps
Program
setting data
Telemetry data receiver
C
Telemetry server
Fig. 3
Server
Athena system
213
Model base development
3.2. Explanation of Each Component

(1) ECU
The ECU has three roles, in engine control, chassis
control, and measurement control. In engine, it controls
the ignition timing, volume of fuel injected, and throttle
opening, and realizes the requested torque. In chassis
control, it uses vehicle behavior information to perform
wheel drive torque transfer control and seamless shift
control computation for gear shifting. In measurement
control, it performs measurement computation of two
kinds, in telemetry and data logging.
(2) Capacitive Discharge Ignition (CDI)
The CDI is the unit that supplies the energy for
ignition to the engine ignition coil. The CDI method
provides greater real time effectiveness than the full
transistor method and greater precision in drive torque
control (traction control and engine brake control) and
engine speed control. This unit also simultaneously has
misfiring detection functionality by means of ion current
detection.
(3) Power Box (PBOX)
The PBOX is the unit that operates as the voltage
regulator and power distributor (sequencer function). The
voltage regulator part uses a highly efficient converter
method and supplies two power outputs (14 V and 7 V)
so that the power components where this power is
supplied can be made more compact.
(4) Batteries
The system uses NiMH batteries for compactness and
lighter weight. Battery temperature monitoring and
charge control are handled by the PBOX.
(5) Telemetry transmitter and receiver (HTX, HRX)
The HTX is the data transmitter unit by which data
processed by the ECU is sent by radio to the pit. The
HRX is the data receiver unit that receives radio data.
(6) Front Data Acquisition (FDA)
The FDA is the unit that acquires data from sensors
at the front of the chassis. It is equipped with input/
output (I/O) capable of multi-channel analog input. This
functions to process acquired analog signals by filtering
computations and then transmits them to the ECU by
high speed communications (10 Mbps).
(7) Sensor Terminal (ST)
The ST is a compact sensor measurement unit used
for measurement of the environment in the chassis and
in the engine. Up to eight ST units can be connected to
the ECU and CAN communication lines.
(8) FIA/FOM units
The FIA/FOM units include a data logger, camera,
and GPS units. Installation of these units is required by
the Fdration Internationale de lAutomobile (FIA), the
umbrella organization for automotive matters, and by the
Formula One Management (FOM), the organization that
conducts racing. Car data from the ECU is sent to the
units by CAN communications, and the racing circuit

data for each team is managed and used by the FIA and
FOM.
(9) Team Data Logger (TDL)
The TDL unit has a logging function and reads data
from all types of sensors. The use of FIA standard
systems has been a requirement since 2008. This unit is
therefore developed so that assistance tool used up to
2007 can continue to be used. The TDL is connected
with FIA standard systems by CAN communications.
(10) Garage system
The garage system is system infrastructure composed
of the server installed in the pit and the client PCs used
by each engineer. It was designed exclusively for use in
Formula One racing. As for the communication between
the car and PC, Ethernet was used.
3.3. High-Speed, High-Precision Control (Engine,
Chassis, and Measurement Control)
The ECU was required to have high computational
capability in order to provide satisfactory control
performance and measurement performance. In the third
era of Formula One, the main items for control
development have been drive torque control and
seamless shift control. The efficiency of this control
development was enhanced by the introduction of model
base development using floating point operations. Figure
4 shows the controls that were applied, together with the
changes in computational capability. The regulations
made driver assistance control (slip control, typified by
traction control) permissible in mid-2001. From then up
to 2007, when it was outlawed, this demanded the
greatest computational capability. This also involved an
increase in the volume of data for control analysis, and
the CPU and data transfer speeds were increased in order
to be able to support the demand for high-speed
sampling and logging on multiple channels for
measurement function.
Seamless gearbox control
Control
Torque control
Traction control
(MIPS)
3000
Total calculation instruction
by Ethernet communications (100 Mbps) so that logging

data and other high-volume data can be received in a
short time. All vehicle checking applications that handle
high volumes of data are also placed on the client PC
side in order to simplify on-board systems.
214
Engine control
Measurement control
Chassis control
Measurement control
2000
1000
2001
2003
Fig. 4
Processing speed increase
2005
2006
2007
M/Y
3.4. On-Board Layout

The main electronic control units were placed beneath
the radiator duct. This location was at the center of the
chassis as a whole, as well as at a low position, and it
was the largest area where units could be mounted
without affecting the vehicle dynamics. As shown in Fig.
5, the ECU and CDI are placed on the left side of the
chassis between the radiator duct and the under floor.
They are affixed to carbon fiber brackets on antivibration mounts. As shown in Fig. 6, the PBOX,
battery, and HTX are placed on the right side between
the radiator duct and the under floor, like the units on
the left side. In order to cool the units, a duct shape
CDI
ECU
Fig. 5
Left side electronic units
PBOX
Battery
HTX
Fig. 6
Right side electronic units
All outlets open-NACA duct
All outlets open-Deep NACA duct
High temp
NACA duct
Fig. 7
Cooling duct simulation
capable of introducing the necessary cooling airstream

was simulated (Fig. 7), and on that basis a duct shape
that would have minimum effect on the aerodynamics
was created. This meant that the heat radiation structure
of the electronic control units was crucial.
4. ECU Architecture and Functions

4.1. ECU Hardware Architecture
In order to realize the performance requirements for
the electronic control systems described at the start of
this article, the hardware performance of the ECU in
particular was important. An architecture with multiple
CPUs optimally suited to the various functions carried
out by the ECU was created. The memory architecture
that coordinates those CPUs and their communications
links are described below.
The basic architecture of the ECU, as shown in Fig.
8, includes three CPUs: the Application CPU (ACPU),
the Device CPU (DCPU), and the Gateway CPU
(GCPU). The ECU was configured in two blocks, one
for the engine and the other for the chassis, and their I/
O were configured for their respective input and output
devices.
A balance between real time operation of the devices
and high-speed computational processing of applications
was sought by distributing functions so that processing
would not be concentrated in a single CPU. In the part
for device operation, operations of the injector and the
ignition device take place at 150 s intervals when the
engine is running at high speed. Consequently, these
were made into functions handling only device I/O
operations, and were separated from the processing of
applications that require high throughput computation.
The selection of a dedicated automotive CPU (DCPU)
with better I/O functionality therefore satisfied the
functionality requirement.
A high-speed processing CPU (ACPU) was selected
to handle the control applications used with the engine
and the chassis. What had until then been controlled
using three automotive microcomputers was combined
into one. This reduced the need for data access between
CPUs for applications and reduced unnecessary
processing.
The CPU used for measurement control carries out
high-volume data communications with PCs and
communications with telemetry units. A CPU (GCPU)
that is high-speed and is equipped with many varied
communication devices was therefore selected, allowing
data operations that make use of high-speed serial and
Ethernet communications.
The multiple CPUs with distributed functions have to
look up each others data. For example, data that is
necessary for telemetry and logging is sent to the GCPU
and the device instruction values that are output by
control computation are sent to the DCPU. The
calibration data and commands that are sent from client
PCs in the garage system are shared by every CPU. In
order to implement these actions, DPRAM was placed
between all the CPUs and high-speed synchronized
215
TDL developed for the 2009 model, and it achieved a

30% increase in data download speed together with an
increase in capacity (to a maximum of 8 GB).
access was enabled by means of trigger signals when

they make connections.
Local memory is assigned to each CPU according to
the functions that are required. Memory functions include
flash memory for program saving, SDRAM for program
execution, NVRAM for data backup, and Compact Flash
(CF) memory for use in high-volume logging. CF
memory is installed with two cards in parallel in order
to expand the data bus width from 16 bits to 32 bits.
The purpose was to achieve higher-speed data processing
and access processing during logging.
The engine and chassis blocks are connected by
ARCNET communications at 10 Mbps. This is to
transfer the parameters required for coordinated control
above 1 kHz process cycle rate.
Up to 2008, the GCPU performed logging data
retrieval processing and data download processing in
alternation. How to conduct these processes
simultaneously, therefore, became the technical issue
when a balance between increased logging volume and
reduced download time was sought.
The high-speed memory controller shown in Fig. 9
was therefore developed. The logging data retrieval
process and download process could be conducted in
parallel by using the PCI bus and the CPU local bus.
The increased speed and greater volume in reading and
writing logging data was realized by means of a circuit
architecture with multiple SD memories connected in
parallel. This memory controller was installed in the
Ethernet
100 Mbps
2 ch
To PC
To
Telemetry
CPU (GCPU)
PCI
( 32 bit / 66 MHz )
20 Mbyte/s
SD
20 Mbyte/s
SD
Memory
controller
(FPGA)
Fig. 9
64
64
SDRAM
Work area
64 MB
LOGGER and
COMMUNICATION
UART
2 Mbps
2 ch
4.2. ECU Software Architecture

The ECU software can be generally divided into
these main parts: the engine control part that handles
control of engine ignition and throttle, the part that
handles control of the gearbox and vehicle behavior, and
the data measurement part that handles the recording of
circuit data and its transmission to the garage system.
A real time operating system was implemented on
each CPU. This met the demand for the advanced real
time performance needed to realize high-speed control
computations, sensing, and communications. As shown
in Fig. 10, the software is divided into three distinct and
independent layers, namely, the OS layer, the middle
layer that performs communications and I/O control, and
the application layer that performs control computations.
Flash ROM
Program
8 MB
32
Flash Disk
LOGGER
2 GB
CPU
SD memory controller
32
16
NVRAM
Back Up
4 MB
ARCNET
10 Mbps
1 ch
8
CAN
1 Mbps
2 ch
300 MHz
16
DPRAM
Work area
512 KB
32
SDRAM
Work area
64 MB
64
DPRAM
Work area
512 KB
16
Analog OUT
12
32
Flash ROM
Program
8 MB
NVRAM
Back Up
4 MB
To
Chassis block
To CDI, PBOX
To ST
64
CPU (DCPU)
DAC
12 bit
4 ch
MOOG driver
(2 ch)
CPU (ACPU)
32
Control
APPLICATION
16
DPRAM
Work area
512 KB
16
Digital IN
Single end
0-5 V(6 ch)
80 MHz
450 MHz
ARCNET
10 Mbps
1ch
DEVICE I/O
1 Mbyte flash ROM

48 Kbyte RAM
Analog IN
Single end
0-5 V(6 ch)
Analog OUT
ADC
16 bit
6 ch
16
12
DAC
12 bit
4 ch
CAN
1 Mbps
2 ch
MOOG driver
(2 ch)
Digital OUT
IINJECTOR
driver(10 ch)
Analog IN
To FIA
Single end
0-5 V(4)
PVRS solenoid
driver(5 ch)
DIFFERENTIAL
0-5 V(2)
Lo side
driver(10ch)
Fig. 8
Hardware architecture
216
DIFFERENTIAL
0-5 V(2 ch)
LAMBDA
Sensors(2 ch)
Thermo
Couple(2ch)
Internal
Monitor(26ch)
In order to enable independent control development by

application users in model base development, a
development environment capable of automatically
generating execute files was set up and control
development efficiency was increased (Fig. 10).
4.3. Circuit Data Measurement Function (Logging)
The logging function is used to make settings in the
chassis and power train, to monitor conditions, and to
conduct simulations. It therefore needs to provide
accurate data and handle large volumes of data. This
required specifications that include 2000 channels and
sampling rates from 10 kHz to 1 Hz.
For the efficiency of circuit tests, it is important for
the engineers to be able to quickly determine conditions
in the vehicle after it has been driven, and make it ready
for the next driving session. It was necessary, therefore,
to shorter the time from data download to data display,
and to provide techniques for easy analysis.
Logging data is accumulated each time the car is
driven. This is used not only on the circuit, but also for
analysis and simulation at the various development
centers. It was necessary, therefore, to associate the
circuit data and the test items.
4.3.1. Measurement data format
In order to facilitate data management in every
driving situation, data measurement using the logging
system was set up to split off the data from each driving
session and write it to the logging memory (Fig. 11).
A single driving data is composed of the Run Header
and the Run Data.
The Run Header area records the Lap Data, which
Application CPU
Device CPU
Gateway CPU
Application
layer
Driving force
Control
Device control
Measurement
communication
Middleware
layer
Arcnet
OS layer
RTOS
Data
exchange
Run header 1
Run data 1
RTOS
Run header
Lap data
Straight end data
Fail data
Warning data
Index data
Ch1
Ch6
Ch7
Ch3
Ch6
Ch2
Ch7
Ch5
Ch8
Ch7
Ch3
Ch8
Ch1
Ch5
Ch4
Ch4
Ch6
Ch4
Ch6
Ch2
Ch5
Ch8
Ch7
Ch6
Ch1
Ch1
Ch4
Ch6
Ch8
Ch7
Ch5
Ch3
Ch2
Ch8
Ch2
Ch4
Ch1
Ch3
Ch6
Ch8
Ch5
Ch1
Ch4
Ch6
Ch3
Run data 2
Run header 3
Ethernet
Arcnet
Software architecture
Run data
Run header 2
Data
exchange
RTOS
Fig. 10
Measured data
CAN
Block data
Channel
Channel ID
ID
Cannel
Cannel number
number
Next
Next channel
channel number
number
Sampling
Sampling interval
interval
Number
Number of
of data
data
Data 1
Data 2
Data 3
Data 4
Data 5
Data 6
Data 7
1 kByte
Run data 3
Virtual connection between same channel data
Fig. 11
Logging data format
is updated on every lap; the Straight End Data, which

gives the highest speed point at the ends of
straightaways; and the Fail/Warning Data, which gives
information on malfunctions. These were arranged to
enable engineers to readily identify characteristic points
in each lap. Index data for logging-related data is also
recorded so that circuit data, calibration data, and the like
can be tied together and accurate data management can
be conducted after driving sessions.
The Run Data area contains records for every channel
subject to logging.
4.3.2. Data recording techniques
As explained earlier, it is necessary for data logging
to record data from measurements conducted at a high
sampling rate. Data logging with a high sampling rate
involves increased amounts of data, which leads to
increased download times, and this ultimately diminishes
the efficiency of circuit tests. For this reason, a trigger
logging function was developed. Under specific
circumstances, such as when a gear change is performed,
when a malfunction occurs, or when some other
condition necessitating acquisition of detailed data at a
certain rate occurs, past data is recorded at a high
sampling rate for a maximum of one second back from
a trigger point. At other times, data is recorded at a low
rate so that the volume of logging data can be reduced.
The trigger conditions are defined on a client PC in
the C language and written into the logging configuration
file using reverse Polish notation. After the ECU has
decoded the trigger conditions, it implements trigger
processing when those conditions are met. The system
was set up to allow the configuration to be changed
according to test circumstances.
When initially developed, sampled data was simply
put in time series and the data on each channel was
recorded in logging memory using a discontinuous
method. Since the aim of this was to reduce the
processing burden on the ECU. However, when CPU
performance was enhanced, the volume of logging data
increased and the issue of the time required to display
a graph occurred.
The use of a high-speed CPU (GCPU) enabled the
data recording format to be changed so that the data in
each channel are in continuous sequence and this
upgraded the drawing speed. The data recording format
was changed so that the measurement data for each
sampling from the same channel is placed in a single
block with a fixed length of 1 kB, and measurement data
is recorded as an aggregation of blocks. (Figure 11
shows data from the same channel in blocks of the same
color in the middle column.)
With this change in format, the block of the same
channel is read by unit and taken united, so an increase
in graph display speed was achieved as a result.
This logged data can not only be displayed in
analysis tools, it can also be input to Hardware In the
Loop Simulation (HILS) and used to reproduce driving
conditions on a PC. It has also been used to feedback
measures to solve circuit-running issues.
217
4.4. Garage System Functions

The purpose of the garage system was to manage
circuit data, provide race strategy support, and assist in
conducting vehicle checks. The functions that are
important in running the car were built into the ECU.
Other additional functions for support purposes were to
be taken on by the garage system. This was the
conceptual approach in development, and it was tied-in
with miniaturization of the ECU (Fig. 12).
In the pit, the car is connected to the garage network
where such actions as logging data capture by client PCs
and transfer of calibration data take place. The logging
data is stored on the garage system server so that the
data can be looked up by multiple client PCs. It is also
given numerous functions for support of racing strategy
and test runs, and the system was set up to allow
operational commands to be issued to an ECU that is
connected to the network. Typical support functions and
their association with the ECU are described below.
(1) Data setting support function
At the actual race location, engineers change the
engine and chassis settings on the basis of circuit data.
Changes to settings are made frequently, and such
changes need to be made up to the point immediately
before a race begins. Therefore, the system was set up
so that the data transmitted from client PCs to the ECU
would be the only data changed by engineers on client
PCs, enabling instant setting changes.
(2) Auto Warm-up
Performance enhancements in the engine and gearbox
have brought increased complexity in engine starting and
warm-up as well as the engine checking mode. Auto
warm-up is a support function that automates these
activities. This check mode includes numerous check
modes for determining the engine status, and the engine
is controlled using control instruction values sent from
a PC. The ECU is not prepared beforehand with a profile
of these check modes. Instead, the system specification
calls for the profile to be stored in the form of files on
client PCs in the garage system. This allows the creation
of a variety of check modes adapted to engine
specifications and environment of use without changing
the ECU software.
(3) Configuration of Steering Wheel Functions
In addition to driving operation, the steering wheel
has the capability for the driver to check or make
changes to settings and car information while driving, in

accordance with racing strategy and changes in the
circuit environment. Figure 13 shows how the steering
wheel is equipped with multiple switches, buttons, and
an LCD display (Fig. 14). The functions given to all of
these controls and the content shown on the display can
be freely assigned and set up from applications on a PC
as requested by the individual driver. By sending
information to the ECU through the garage system, the
steering and ECU functions can be linked together.
Frequently used and important functions are assigned to
push switches and rotary switches so that those
operations are carried out with one action. Other
operations are placed in command hierarchies and
accessed using a Mode Function switch and the +/-
switch. This realized a balance between operability and
multifunctional switchability.
Display
Gear position
Mode confirmation
Message
Shift timing Light
Reverse
Alarm lamp
Engine brake control

Traction control
Diff
+/-
+/Speed limit
Shift paddle
(down)
Shift paddle (up)
Driver radio
Neutral / Oil check

Drink pump
Traction control
(tire warm up)
Clutch paddle
(for right hand)
Clutch paddle
(for left hand)
A/F
Fig. 13
Launch/Overtake
Toggle switch Mode function
Steering wheel functions
Measurement data
analysis
Fig. 14
Lap chart
Measured data
Client PC
Other statistic tools
Server
Feedback
5. Hardware Development
Race strategy
support
Setting data tuning
Measured data
Auto warm-up
Client PC
Fig. 12
Steering wheel display
Logged and telemetry

data chart
Garage system
Configuration change
The electronic control units were given a threedimensional structure using multiple circuit boards in
order to increase the package density. Circuit boards
densely populated with electronic parts have large
numbers of wiring connections, and it was necessary to
218
consider techniques for making connections between

boards. Higher-performance CPUs also operate at higher
speeds and generate considerable heat. Techniques for
heat dissipation in miniaturized packages were an issue.
5.1. High-Density Packages
Electronic component packaging techniques that had
no track record of use in automotive electronic
components at that time were employed in 2003 to
realize high-density packages. These were the Fine-pitch
Ball Grid Array (FBGA) and the Build-up printed circuit
board (PCB), and their use realized miniaturization.
Figure 15 shows a circuit board populated with CPUs
using the Build-up PCB technology. For the 2009 model
TDL, a prototype unit was fabricated through the
application of Device Embedded PCB technology. This
was confirmed to achieve 20% greater packaging density
than Build-up PCB products.
CPU operation at higher speeds was accompanied by
faster bus clock rates and CPU electrical power supply
at a lower voltage (1.8 V). This made the units more
susceptible to being affected by disturbance noise and
cross talk. Consequently, there was an even greater
demand than before for stabilization of pattern shields
and ground electric potential during pattern design, and
circuit board wiring design techniques based on highspeed transmission path simulation were adopted.
5.2. Inter-Board Connection Methods
In the case of unit structure from multiple circuit
boards, the method used to connect boards with signal
wires becomes an issue.
Larger units involve a great number of inter-board
signal wire connections. Therefore, not only the
connectors can cause dead space, but it is apparent that
even more space is required when wire routing is taken
into consideration.
In order to resolve this issue, it was decided that the
2008 model TDL would reduce the inter-board
connectors used up to that time by instead using RigidFlexible circuit boards to optimize inter-board wiring. At
the same time, pattern shields could be built in at signal
wire connections for communications and small signals,

and the connections could be made electrically stable as
well (Fig. 16).
5.3. Heat Dissipation Structure
Generally speaking, the automotive CPUs used in an
ECU often do not require heat dissipation measures. The
high-speed CPUs used in Formula One ECUs, however,
have power consumption that rises as high as about 4
W, and they exceed the guaranteed operating temperature
of 105C even at room temperature (25C).
In order to achieve a balance of high processing
performance and miniaturization in the Formula One
ECU, the greatest issue was how to deal with heat. This
was a question of whether the unit could be made
suitable for use in the actual car environment.
Heat dissipation for high-speed CPUs and other such
heat generating devices is generally accomplished by
forced air-cooling using heat sinks. In Formula One
ECUs, for which miniaturization is sought, maximum use
was made of the case and the circuit boards to enable
guaranteed operation even in the actual car environment
where the ambient temperature rises above 60C.
The heat dissipation measures will be described here,
using the ECU as an example. Figure 17 shows the
main electronic parts that generate heat, the heat
dissipation structure, and the heat dissipation pathways.
It is important to dissipate the heat that is generated and
to increase the heat capacity.
Regarding the former, a pattern was formed with a
circuit board affixed to an aluminum plate made from
the same material as the ECU case in order to enhance
heat conduction from the device and the circuit board,
thus using the case structure for heat dissipation.
Regarding the latter, metal core PCBs were adopted for
the circuit boards that have CPUs mounted. These have
high heat conductivity in the planar direction, and the
heat capacity of the circuit boards was increased as a
result.
Fig. 16
Rigid-Flexible PCB of TDL
Al base PCB
Cu core PCB
Thermal flow
A heating element (CPU, Driver)
Fig. 15
High density PCB surface
Fig. 17
219
PCB
Copper core PCB
Frame
High heat conditioning structure of ECU
By means of these various techniques, the unit could

be made suitable for the actual mounted environment in
the car.
Device package technology

Printed circuit board technology
QFP, SOP package

Through hole via
Performance
Refer to Fig.4
Volume
50%
Internal via hole
Rigid-flexible
Metal-cored build-up
FBGA package
1608 chip
Build up
00
01
-02
03
-04
Fig. 18
05
0402 chip
Buried chip
2009 TDL
06
07
09
Year
ECU package
6. Telemetry
6.1. History of Telemetry Development
Telemetry is the system that takes data from the
various kinds of sensor data and the like while the
vehicle is running and uses radio to transmit it to the
pit. Telemetry was first developed for use in races by
Honda in the mid-1980s, during the second era of
participation in Formula One competition. Telemetry was
used during the second era by acquiring important data
on engine speed, engine oil and water pressure, engine
oil and water temperature, fuel consumption, and the like
while the car was passing in front of the pit, and using
that data to manage conditions in the car. In the third
era, by contrast, that system evolved to acquire and
analyze all data concerning the engine and the chassis
in real time. Figure 19 shows the history of change in
communication speed and communication protocols used
in telemetry.
In 2000 and 2001, the telemetry system developed
during the second era was carried over. Its carrier wave
frequency was in the 400 MHz band, the modulation
method was Frequency Shift Keying (FSK), and the
6.2. Issues Associated with Increased Speed

Telemetry systems were checked for data acquisition
performance in a simulated desktop environment. After
that, circuit tests would be implemented; however, there
were many points where testing on the actual circuit and
desktop simulation indicated different characteristics.
This was particularly the case when Quadrature Phase
Shift Keying (QPSK) was adopted as the modulation
method to increase the speed. In desktop simulation, the
packet acquisition rate (the percentage of data packets
220
2 Mbps
Data transmission speed [kbps]
100%
5.4. History of Miniaturization

Figure 18 shows the historical changes in the ECUs
package size and the technologies employed. There has
been demand for higher processing performance since
the 2004 specifications. This was achieved employing the
technologies described above, so that higher performance
was realized while maintaining the size of the 2004
model.
Growing demand for higher performance and greater
miniaturization of ECUs in mass production vehicles, as
well, will no doubt lead to importance being placed not
only on measures for heat dissipation in electronic
components, but also on technology to inhibit the
generation of heat in devices.
communication speed was 19.2 kbps.

A change in the regulations in 2002 enabled the
transmission of data from the pit to the car, initiating the
development of interactive telemetry systems. The
communication speed in one direction was itself
increased to 38.4 kbps. In 2003, however, the regulations
changed again and interactive telemetry was outlawed.
In 2004, the carrier wave frequency was moved to
the 1.7 GHz band, and the communication speed was
raised as high as 460 kbps.
The communication speed was further raised
significantly again in 2006. Of all the data collected
while running, the items required for real time analysis
amount to approximately 1000 channels. To transmit
these by telemetry would require a transmission speed
of approximately 1 Mbps. The demand for high-precision
measurement described in section 4.3. required a
maximum data rate of 1 kHz. In order to meet these
demands, the chosen modulation method was Phase Shift
Keying (PSK), which has high-power efficiency and
frequency-use efficiency as well as low error bit rates
at low receiving levels. In addition, the communication
speed was increased to 2 Mbps, and a trigger function
that increased the rate of acquisition for certain data at
specified times was also implemented.
In 2006, as a result of the evolution in technology
noted above, practically all the data from running cars
could be acquired by the pit in real time. This capability
became an indispensable part of race operations, as the
data from the start of the formation lap could be used,
for example, to change the clutch control mode as the
race was starting.
2000
1000 channel data

100% data coverage
Trigger function
1500
Standard transmission
speed of 3G wireless phone
FSK
QPSK
1000
Two-way telemetry
460 kbps
500
19.2 kbps
38.4 kbps
0
2000 2001 2002 2003 2004 2005 2006 2007
year
Fig. 19
Progression in telemetry transmission speed
received normally) was 90% or higher, but in circuit

tests, the rate declined significantly. The data was being
retransmitted, however, so the coverage (the percentage
of necessary data acquired) maintained a level of about
100% (Fig. 20).
Figure 20 shows a conspicuous decline in the packet
acquisition rate on the circuit, even in sectors that are
relatively close to the pit. When data acquisition
performance declines even at short distances, in other
words, when reception levels are adequate, the cause is
conceivably the influence of fading, defined as mutual
interference on the receiving side from multi-pass waves,
which are reflections from the circuit road surface, the
stands, and the like. The extent of influence from fading
100%
Packet rate / Coverage
90%
80%
70%
depends on communication speed. Figure 21 shows how

the length of a single frame is reduced when the
communication speed increases, so that the amount of
interference is greater proportionate to the frame length.
As a result, the influence of fading becomes significant
and communication quality declines.
A variety of measures were tried to reduce the above
type of decline in the packet acquisition rate, including:
(1) transmission power was increased to mitigate the effects
of noise;
(2) transmission speed was lowered to mitigate the effects
of multi-pass waves;
(3) a directional antenna was used on the receiving side in
order to avoid multi-pass waves; and
(4) forward error correction (FEC) was adopted to lower the
error bit rate.
Ultimately, however, these measures did not achieve
any major reversal in the decline.
60%
50%
40%
Packet rate
30%
20%
Coverage
10%
0%
9 10 11 12 13 14 15 16 17 18 19 20 21 22
Sector
PIT
Fig. 20
Packet rate and coverage for circuit one

round (2007 Catalonia)
Slow data transmission speed
Direct wave
Frame 1
Delay wave
d: Delay time
Frame 2
Frame 1
Frame 3
Frame 2
Frame 3
Usable frame
d
Interference
(unusable)
Received signal
Fast data transmission speed

Direct wave
Frame Frame Frame

1
2
3
Frame Frame Frame

1
2
3
Delay wave
Bits/frame is
same as above
6.3. Consideration of a Radio Wave Propagation Model

for the Circuit
As the development of telemetry depends on desktop
simulation, an accurate grasp of the radio wave
propagation environment on the circuit is of major
importance. Consequently, telemetry development from
2007 on included efforts to raise packet acquisition rates
on actual circuits by conducting investigations into radio
wave propagation environments that resembled the actual
environment.
The environment used for Formula One racing differs
substantially from that of ordinary mobile
communications (mobile phones and the like). Movement
speeds can exceed 300 km/h at the maximum, there is
no clear line of sight from the pit to the farthest point,
and adaptive transmission cannot be used because the
communication is one-way from the car to the pit.
Figure 22 shows the radio wave model (fading
model) that was used in simulations up to 2007. In this
model, there are two routes (paths) followed by radio
waves, and the maximum delay time between paths was
0.2 s. There is generally said to be a reflection delay
time of 0.01 s on land where the line of sight is clear,
and a reflection delay time of 0.1 s (a maximum of 0.2
s) where there are buildings. This model is thought to
diverge from the actual situation on the following two
points:
(1) The path delay time is short.
The delay time of 0.2 s represents no more than 60
m when converted into an optical path length.
Considering that in the actual environment there would
be reflections from distant stands, surrounding
Many rates of
interference
Max delay time

: 0.2 s
2
Received signal
Fig. 21
Difference in influence of fading by data

transmission speed
Fig. 22
221
Fading model used for development to 2007
mountains, and the like, it can be inferred that the path

delay time would be longer.
(2) There are too few paths.
Since the telemetry transmission antenna is nondirectional and transmits radio waves in every direction
from the car, it can be expected that waves reflected
from every structure on the course will be arriving at the
receiving antenna at the pit, and the number of paths can
therefore be conjectured to be greater than two.
Measurements made on site are thought to be the
optimal way of ascertaining the radio wave propagation
environment. As explained earlier, however,
measurement on the circuit presents issues. Therefore, a
variety of values were assigned as fading model
parameters and the actual measured packet acquisition
rate and the error pattern in data received with errors on
circuit tests were compared with simulated tests. The
following results were obtained:
the maximum number of paths is six; and
the delay times of multi-pass waves of dominant
strength were about zero to 1 s, and weaker multipass waves had a maximum delay of 5 s.
It was found that the above fading model is similar
to the actual environment (Fig. 23).
The above fading model is known as the Rayleigh
fading model. Since the influence of overlapping multipass waves in large numbers can cause consecutive
errors to appear in the data, little effect is generally felt
from increases in transmission power or forward error
correction. Furthermore, the existence of a large number
of paths means that even the use of directional antennas
Max delay time
: 5.0 s
1
2
3
4
5
6
Fig. 23
Fading model used for development from 2008
will not fully avert the effects of multi-pass waves, and

since the delay times are so long, decreasing the
transmission speed of 2 Mbps to just one-half or onequarter cannot be expected to resolve the issue of
communication quality. When the model in Fig. 23 is
viewed in this light, it becomes quite understandable that
the measures described in section 6.2. to address the
reduced packet acquisition rate did not yield any major
effects to better the situation.
In development from 2008 on, the model in Fig. 23
was employed to conduct evaluations of new
communication protocols to realize higher-speed
communications while still maintaining an adequate
packet acquisition rate even under this model.
6.4. Technology for Next-Generation Telemetry
The aim from 2009 had been to do away with
downloading by developing a system that uses real time
telemetry to transmit all the data that has been logged
by the ECU and the logger up to that time. Investigation
was in progress with the goal of achieving further
increases in speed. In order to seek still greater speed
under present circumstances, it is important to adopt a
multi-carrier system capable of reducing the transmission
rate separately by carrier as a measure to increase fading
resistance. Conceivable specific techniques to realize the
aim of increased speed include the use of Orthogonal
Frequency Division Multiplexing (OFDM) for
enhancement of frequency-use efficiency and fading
resistance.
A prototype for desktop use of OFDM was created
and an evaluation of its characteristics was carried out
using the fading model shown in Fig. 23. Figure 24
shows a comparison of the characteristics of the
telemetry transmission equipment from 2007 and the
OFDM prototype. The vertical axis in Fig. 24 represents
the Bit Error Rate (BER: the number of bit errors
divided by the number of bits transmitted) while the
horizontal axis represents the number of fading paths. It
can be confirmed that the adoption of OFDM has
realized higher speed and higher reliability.
7. Component Development
BER (Bit Error Rate)
1.E+00
1.E-01
1.E-02
System in 2007 (Data rate 2Mbps)

1 carrier data rate 2Mbps
1.E-03
OFDM Prototype (Data rate 4.8Mbps)

1 carrier data rate 0.1Mbps
Good
1.E-04
Fig. 24
4
5
Number of paths
Comparison of conventional system and

OFDM
As the Formula One engine and chassis have 100 or

more sensors and actuators mounted on them,
miniaturization of these devices was required with a
view to mountability. Particularly, sensors mounted on
the engine (for measuring pressure, temperature, timing,
and throttle position), along with ignition coils and the
alternator, were thoroughly miniaturized and made highly
efficient. In addition, the greatest issue was the guarantee
of their sustained functionality under engine vibration.
This article will provide details on the development
of the following:
(1) throttle position sensor; and
(2) engine wire harness
7.1. Throttle Position Sensor
In the engine of an ordinary automobile, the intake
222
flow is measured by the intake manifold pressure or by

using an air flow meter, and that measurement is used
to control the fuel injection volume. In the highly
responsive engines used for racing and the like, however,
the throttle opening is taken as the basis for setting the
fuel injection volume. The throttle position sensor,
therefore, is one of the key sensors in a Formula One
engine, and requires high accuracy (within 1% at Full
Scale) and reliability.
When Honda returned to Formula One racing in
2000, the brush type of contact sensor (potentiometer)
was initially being used. There were various issues with
it, however, including wear of the brush contact surface
from vibration and foreign matter (such as oil or dust).
As vibration from the engine reaches approximately 500
G, there was an urgent necessity to isolate the sensor
from contact for the purposes of accuracy and reliability.
From late 2003, therefore, this sensor was changed to a
non-contact type using a magneto-resistance (MR)
element to detect the magnetic vector. As this is the
basic sensor for fuel injection control, it is necessary to
provide high accuracy across the entire temperature
range. An IC using two MR elements was adopted, and
accuracy was assured by a technique that cancels out
their respective temperature characteristics.
A split type sensor was adopted that has the sensing
element mounted on the throttle body, and a magnet
(polarized to generate the desired magnetic field)
attached to the shaft of the throttle butterfly, which is a
rotating body (Fig. 25).
This sensor was mounted toward the back end of the
engine, and heat damage occurred frequently.
The main cause of heat damage was exhaust heat
from the exhaust manifold. Due to the aerodynamic
requirements of the chassis, the engine cowl was being
squeezed to a smaller size every year. This had an
impact on the flow of air inside the engine cowl, in
addition to which the layout of the exhaust manifold was
also changed accordingly so that it was closer to the
engine where the sensor was mounted. The result was
an increasingly harsh thermal environment. Even when
the MR element and other IC parts are guaranteed at
high temperatures, such guarantees ordinarily cover up
to 130 C. In order to guarantee up to 150 C for racing,
high-temperature solder and other such materials were
used, while the range of guaranteed accuracy was

restricted and adjustment made by means of software.
Through these and other such measures, sensor accuracy
was able to satisfy the condition of staying within 1%
FS.
7.2. Wire Harness
The Formula One wire harness was designed with an
emphasis on reduced weight, durability, and on-site
maintainability. Up until engine homologation was
prescribed in 2007, there were no engine weight
regulations. Consequently, measures were specialized in
weight reduction, and the wire harness was fixed in place
using simple stays that were mainly placed around the
airbox and cylinder head. When engine weight
regulations were instituted, there was a shift in
orientation to maintainability and reliability rather than
weight reduction. A junction box was placed above the
cylinder head and the circuits for all the sensors were
brought together in that box (Fig. 26).
In terms of reliability, it was a struggle against breaks
in the wires. The top speed of a Formula One engine is
double or more that of a mass production vehicle, and
normal engine speed extends across the entire range, so
that areas around the engine are subject to constant highfrequency vibration. In the case of a mass production
vehicle, the wire harness is bunched together with tape
or other such material and then encased in outer
sheathing material (plastic tube) to counter wire breakage
and wear at points of contact due to vibration. In
Formula One cars, since the use of finer wires and
reduction in weight are also important objectives, heatshrink tubing made of woven polyester fiber is used to
protect against external contact as well as to realize the
use of finer wiring and lighter weight. Wire breakage
resulting from tension load on the harness generated by
vibration and acceleration has further been addressed by
the use of silver-plated high-strength copper wire for
racing use. The harness wires are then twisted together
and bundled inside shrink tubes, enhancing harness
Junction Box (J-Box)
Wire harness
Hall IC
Throttle butterfly fitting
Throttle body fitting
Magnet
Sensor case
Fig. 25
Throttle sensor
Fig. 26
223
Junction box on engine
strength against breaks (Fig. 27). This twisted structure

not only enhances strength against wire breakage, it also
heightens the flexibility of the harness itself and
enhances engine mountability.
Even when these techniques are used, however, there
are still places on the engine where vibrations at the
level of several hundred Gs are generated. There have
been cases, therefore, where just the copper wire has
broken inside its cladding or inside the shrink tubing
even when no external damage is visible. This is thought
to be generated by sympathetic vibration of the copper
wire or of the harness itself due to the high-frequency
vibration, and the vibration is conjectured to result in
displacement and tensile loads that cannot be absorbed.
The method was therefore adopted of fixing the entire
harness in place between the attachment points, including
the oscillating parts, by encasing it in rigid retaining
parts made of carbon material. This has succeeded in
preventing wire breakage.
Vibration has been an unavoidable obstacle in the
course of Formula One development. In the initial phases
of development, it was important to grasp the levels and
patterns of vibration in advance and to skillfully combine
the measures taken to counter them. Another point is that
the short Formula One development period allows
limited opportunities to conduct the durability checks
under engine operating conditions that are necessary to
guarantee reliability. Consequently, it is necessary to
ascertain the symptoms of faults and the effectiveness
of countermeasures early on, and methods of analysis
using X-ray and other such non-destructive inspection
devices in durability testing have been introduced to the
harness development process.
It has been necessary to manage the design effort so
as not to interfere with weight reduction measures by
pursuing vibration-resistant design that goes beyond what
is required in actual vehicles.
regarding new technology has been obtained. The

electronic control systems developed for Formula One
are for unique vehicles and are further specialized for
racing functionality. However, the conceptual approach
to system construction and the basic electrical system
technologies that were employed are also likely to be
held in common with the advancing automotive
technology of the future.
It has been decided that the development of electrical
and electronic technology for Formula One racing
underway since the adoption of FIA standard systems
will be superceded by the development of kinetic energy
recovery systems (KERS), which is employed from
2009. It is to be hoped that Formula One will continue
to stand as the highest achievement in automobile racing,
as well as the highest achievement in automotive
technology in times to come.
Author
Fig. 27
Kenichiro ISHII
Toshiyuki NISHIDA
Kohei TOSHIMA
Masaki NEGORO
Masataka YOSHIDA
Yutaka MARUYAMA
Harness
8. Conclusion
In retrospect, Hondas Formula One system
development proceeded without break from the time of
the test runs at Pembrey Circuit in the United Kingdom
in 1999, one year after their development started, up to
the Brazil Grand Prix in 2007, and a wealth of findings
224
Masataka YOSHIDA*
Masato KITA*
Hirofumi ATARASHI*
ABSTRACT
In 2009, the regulations for the Formula One World Championship were amended to allow the use of kinetic energy
recovery systems (KERS). The new regulations stipulated that the KERS drive shaft be limited to the rear wheels,
that output should be no more than 60 kW, and that the amount of energy used per lap be no more than 400 kJ.
Honda had been conducting R&D in this area since the summer of 2007, and had developed a high speed, high output,
direct oil-cooled motor, a water-cooled power control unit (PCU) which integrated a motor drive inverter unit and
voltage control system, as well as a high power density lithium ion battery, all based on being small and lightweight
enough for Formula One characteristics.
This system was first used to drive on straight roads in April 2008, and in May of that year Honda beat out other
teams to conduct the worlds first driving tests in an actual vehicle at the Silverstone Circuit, where the technologys
superiority and high level of safety were proven.
1. Introduction
Honda has been developing electric vehicles, fuel cell
electric vehicles and hybrid vehicles to find alternatives
to fossil fuels, reduce emissions and mitigate the impact
of automobiles on global warming.
As for the Formula One World Championship, the
regulation amendments of 2009 allowed usage of KERS,
which recovers and utilizes braking energy as drive
power assist. To work with the new rules, Honda chose
for its energy recovery method an electrical hybrid
system and proceeded with development for use in
Formula One based on the electric drive technology it
had been developing.
Because the maximum output and the amount of
assist energy that can be used per lap are stipulated by
regulations, this development focused on making
equipment as small and lightweight as possible with high
output and high torque technology without changing the
high level of dynamic performance unique to racing cars,
and therefore Honda developed a motor, PCU and
lithium ion battery capable of installation on a racing car.
Honda additionally achieved high responsiveness to meet
the requirement for output characteristics during racing.
Development began in earnest in the summer of
2007, and in just nine months, actual driving tests were
conducted using the prototype vehicle RA1082 (a vehicle
built to check functionality), and subsequently KERS
was run at full power on a racing course for the first
time in the Formula One environment. Based on the
basic functions that had been confirmed with the
RA1082, the technology then went into the RA1089

(race prototype) and then the RA109K (racing vehicle)
by the end of 2008 (Table 1).
This paper recounts the development of the Formula
One hybrid system.
2. Development Concept
Under the 2009 regulations, KERS could only be
connected to the rear wheel, with maximum energy use
per lap of 400 kJ and maximum output of 60 kW. The
necessary performance targets set as development themes
for this project were as listed below, not only to make
the equipment compact and lightweight in order to be
installed on a Formula One vehicle, but also to create a
vehicle capable of winning races.
(1) System weight: no more than 30 kg
(i.e., no more than 60% of the vehicle ballast weight of
the 2006 vehicle)
(2) Assist performance: at least 5 continuous seconds at
output of 60 kW
(i.e., the output and assist time enabling the vehicle to
overtake others)
When setting targets for performance, weight, center
of gravity and the like for the various functional
components, these parameters were investigated from
many angles, including race strategy and the use of
vehicle dynamics simulation, but some major conceptrelated issues were encountered when doing these
investigations.
One was whether the race strategy should emphasize

225
Battery
PCU
Motor
Table 1
Specifications of KERS
RA1082
2008-APR
60 kW
45 Nm
800 kJ
Transmission
100 x 202
21000 rpm
7.7 kg
Transmission oil
Si-IGBT/SiC-diode
Boost copper
680 V
11.2 kg
Water
Li-ion
114
22.4 kg
Air
Vehicle code
Shake-down
KERS power
Torque
KERS energy
Location
Dimension
Max rpm
Weight
Cooling
Power module
VCU type
Operating voltage
Weight
Cooling
Type
Cell number
Weight
Cooling
RA1089
2008-NOV
60 kW
45 Nm
400 kJ
Front of engine
100 x 190
21000 rpm
6.9 kg
Engine oil
Si-IGBT/SiC-diode
Switched capacitor
560 V
8.0 kg
Water
Li-ion
108
21.2 kg
Air
RA109K
2009-JAN
60 kW
45 Nm
400 kJ
Front of engine
100 x 190
21000 rpm
6.9 kg
Engine oil
Si-IGBT/SiC-diode
Switched capacitor
580 V
7.4 kg
Water
Li-ion
106
20.8 kg
Air
Battery
Front
Front
Layout
Motor
Motor
Motor
PCU
Battery
Front
overtaking or lap time. We compared single assist, which

uses the 400 kJ all at once, or multi assist, which splits
the maximum 400 kJ to use it on multiple instances;
simulations showed that compared to single assist, a
multi-assist system that settled the energy budget at each
corner could reduce lap times by about 0.1 seconds.
Figure 1 shows the time gain achieved on each circuit
owing to the difference in type of assist. A multi-assist
system allows smaller energy storage, so installation of
a super capacitor with low energy capacity but great
power density was considered. This would have the
synergistic effect of having a lesser impact on the
chassis. However, one cannot win a race unless one gets
out ahead of other cars. For example, if assist is begun
at 180 km/h, where the tire grip exceeds the drive torque
(i.e., the tire is not skidding), assuming that the output
of 60 kW will be used for 6.666 seconds (an energy
equivalent of 400 kJ), vehicle speed can go 15 km/h
faster than without assist, which in distance terms is a
PCU
Battery
PCU
difference of 20 m. Even supposing that it were not

possible to use assist for 6.666 seconds because of the
course layout or other reasons, this would be a distance
gain of 5 m (one car length) in 2.78 seconds and 10 m
(two car lengths) in 4.22 seconds over another vehicle,
which allows overtaking, so single assist, in this case,
is more effective in terms of race strategy (Fig. 2). The
opinions were thus divided on how to use KERS in
actual races, so there was even some wavering on target
requirements, but after several discussions with the
Honda Racing Formula One Team (HRF1), the concept
of emphasizing overtaking was ultimately decided upon.
The second issue was to maximize KERSs recovery
of energy from braking while also maintaining
drivability. Brake cooperative control, and the like is
prohibited under the regulations, and therefore the
amount of energy recovered varies not only according
to course layout but also according to driving style,
including such factors as how long and at what pressure
0.35
Single assist at longest straightaway

Multi assist
Lap time gain (sec)
0.3
0.25
0.2
0.15
0.1
0.05
0
Mel Sep Bah Bcn Mco Mtl
Ind Mgc Sil
Nur Bud Ist Mza Sha Int Suz Hok Spa Fuj Imo
Circuit
Fig. 1
Lap time reduction
226
the brake is operated. Additionally, simulations indicated

that, depending on the circuit, it may not be possible to
recover the target amount of energy. To address these
issues and minimize the differences in driving style, we
minimized loss from zero torque control that occurs
during shift change and when offsetting motor friction,
and furthermore achieved optimal allotment of engine
braking and regenerative braking and made settings to
let each functional component work with maximum
efficiency in the range of usage. In order to get more
regenerative energy, Honda also decided to propose to
the Federation Internationale de lAutomobile (FIA) that
we revoke a number of regulations relating to
regenerative conditions, for example, disallowing
regeneration if brake pressure is not below a set value.
same time it lessens the capacity of the gas bag (fuel

tank) located in this area; the gas bag could now hold
83 liters, approximately 40% less than the earlier 140
liters. This was considered a racing strategy issue. As
for the motor, originally it should have been directly
installed with the engine close to the center of gravity,
but because of the engine homologation (development
freeze) in effect since 2006, no engine alterations were
allowed and connections could not be made to the drive
shaft, so the motor was placed inside the transmission
case. Giving priority to minimizing the bulge of the case
to avoid interfering with the aerodynamics to the rear,
the motor gear was connected to the five speed gears of
the lay shaft (the driven side shaft). The motor diameter
of 100 mm was decided upon based on the installation
fastening of the transmission and engine.
3. Development of KERS
3.1.2. Development of RA1089 (race prototype)
We considered the issues of the RA1082 and began
designing the RA1089 with the aim of minimizing the
impact of KERS being on-board while maintaining a
high level of vehicle dynamic performance. This vehicle
also had an important position as the winter test vehicle
(i.e., pre-season test vehicle) in anticipation of the
coming 2009 season. The unprecedented installation
layout, offering both vehicle performance and KERS
performance, ran into troubles, and even once into
chassis production, there were numerous design changes
that put a burden on the production site. Moreover,
engine changes to install KERS that were originally
unapproved were later allowed after a request from
Honda to FIA, so the installation of functional
components was reviewed again. About that time, the
Car speed (km/h)
3.1. The Story of On-board Package Development

3.1.1. Development of RA1082 (functionality check
vehicle)
The RA1082 was built to demonstrate the advantages
of KERS in track tests and find any issues with the
conformity of functional components to Formula One
conditions. Since the PCU and battery are heavy
components, using the RA106 (the 2006 race vehicle) as
the base, we altered the area behind the drivers seat
inside the monocoque (a CFRP body forming a cockpit)
and placed these components there so that we would
have little impact on the vehicles center of gravity.
Although placing high voltage components like the PCU
and battery in this area has advantages in terms of the
vehicles center of gravity and electrical safety, at the
*Monza straight acceleration
350
300
200
+ 15 km/h
Assist start
180 km/h
150
Distance gain (m)
Non-KERS car
400 kJ assist
finish at 60 kW
250
30
Normal
40 kW
50 kW
60 kW
25
20
15
10
+ 20 m
5
0
Time (sec)
7
6.666 sec
Non-KERS car
10
Distance gain (m)
Scale up
10
9
8
7
6
5
4
3
2
1
0 0
2 car lengths
1 car length
3
2.78 sec
4
4.22 sec
Time (sec)
Fig. 2
Assist time / car speed and distance gain
227
6 Non-KERS car
time to decide the chassis layout for RA1089 was

already coming soon, but Honda felt that vehicle
dynamic performance was the number one priority and
so, in a short period of time, worked with the HRF1
engineers to modify the engine and reinvestigate the
cooling specifications of functional components and the
installation position as determined by all items, including
safety in the event of collision. As a result, from the
many ideas proposed, it was decided to install the motor
on the engine front side, the PCU in the left side pot,
and the battery in the monocoque front keel. Table 2
shows a comparative investigation based on motor
installation position. This decision not only made it
newly necessary to have a motor gear train that
connected the engine and motor drive shaft, but there
was also an urgent need to ensure toughness against
collision for the PCU and battery located outside the
monocoque and provide for electrical safety with
unprecedented techniques. It was decided to give these
the highest priority and reinvestigate the specifications
of functional components from the beginning. This put
much pressure on the period of stand-alone testing for
the PCU and battery, and meant that system launch
would have to be taken place in a short period of time.
The next issue was the relative position between the
gas bag and motor. If the engines motor drive is
transmitted directly to the crank shaft, it is necessary to
put a recess around the gas bag, which is located in front
of the crank shaft, of a volume equivalent to the motor.
The motor was offset to the left side where it would not
impact aerodynamics and the gear train was placed
between the back end of the monocoque (the bulkhead)
and the engine so the loss of gas bag volume was
minimized as much as possible.
Because the functional components underwent
completely new development, the challenging work,
performed both in Japan and in the UK, dragged on for
a long time, so the completion of the RA1089 was
delayed. As such, the system check was not in time for
the original target of the seasons first winter test in
November, and the target was changed to the second
winter test.
Table 2
Comparison of motor position
Motor position
Mass
Front-weight
distribution
C.O.G
Engine front - LHS
8.69 kg
-0.60%
+1.0 mm
Gearbox
10.30 kg
-1.13%
+1.4 mm
the Japanese and UK bases proceeded with development

24 hours a day until the end of November. In particular,
the battery module occupies the greatest volume and
weight of all functional components and affects the
vehicles weight distribution and aerodynamics package.
The decision on the placement of the battery disrupted
the original timeframe for deciding on specifications, but
ultimately it was decided to use a dispersed placement,
with the battery module located in the keel and also in
the nose, a position which until then had not been
approved. However, the FIA collision regulations are
very strict as far as the nose is concerned, so HRF1
conducted simulations and bench collision tests with the
collision regulations to prove the layouts safety and
finally earned the FIAs authorization (Fig. 3).
3.2. Overview of KERS
This section discusses the RA109K vehicle intended
for races.
Figure 4 shows an overview of the Formula One
hybrid system. The motor drive shaft is connected to the
engines cam gear train through the five-gear train for
the motor. The motor and PCU were on an anti-vibration
mount on the left side of the monocoque, with the motor
cooled by engine oil and the PCU by special cooling
water. The battery was fixed to the front of the
monocoque, and the temperature controlled by the draft
air from the front. Besides these modifications, the PCU
was connected to the FIA standard ECU (S-ECU) by
3.1.3. Development of the RA109K (racing specs

vehicle)
The RA109K was designed as the racing vehicle.
KERS components were located in about the same
position as in the RA1089, but in the RA109K, the
periphery of the engine cowl, one of the aerodynamic
components, was made smaller than in the RA1089, and
it was necessary to review from the beginning the forms
of the PCU and battery located inside the cowl.
Aerodynamic performance is affected by wind tunnel
testing time, and therefore aerodynamic components
were first produced with a target of being ready for the
opening race in March with subsequent updates
following in turn. However, the specifications of KERS
components cannot easily be changed because we take
so long to produce and check for reliability. To ensure
development of specifications meeting both these needs,
228
Battery module
(nose)
Battery module
(keel)
Fig. 3
Front collision test for FIA regulations and

battery layout for RA109K
CAN communication for mutual commands and

monitoring. Assist and regeneration were controlled in
linkage with the drivers operation of the vehicle.
In Formula One vehicles, the vibration environment
is more severe than in mass-produced vehicles, so the
PCU and battery were built so that even at a vibration
of 20 G, there would be no issues in terms of
functionality. The motor is required to be able to
withstand even greater vibration, because it is exposed
to engine vibration, but since this surpasses the capacity
of the vibration generator, no stand-alone evaluation was
performed; instead, reliability was judged by checking
on an engine bench and in track tests. Moreover, in light
of collision safety, a structure was chosen that met the
FIAs side impact requirements.
3.3. Motor Gear Train
A five-gear train was used as a means of connecting
the crank shaft with the motor, which had been moved
to the left side of the monocoque. The gears engaged
with the engines cam gear train to connect to the crank
shaft (Fig. 5). These KERS-specific parts were made
lighter by the use of magnesium covers, titanium bolts
and ceramic ball bearings, the use of which is prohibited
in the engine itself under the regulations.
There was already only a few mm of clearance
between the engine and monocoque, and to put the gear
train here required creating some space by putting a
recess in the bulkhead. Such a recess diminishes the
rigidity of the monocoque and the capacity of the gas
bag, so the layout characteristics were enhanced by
integrating the gear train housing with the engine front
cover. The gear train was in a parallel configuration of
five gears with a height of less than 30 mm, but this was
the our first experience with a gear train of this form
and a technique transmitting torque in both directions,

so during development there were frequent issues, such
as broken shafts because of resonance in the gears. After
subsequent changes of specifications in fine areas and
repeated durability testing, we were able to ensure
durability in time for track testing.
3.4. Motor
For a motor to be adopted on a Formula One vehicle,
it should of course be compact and lightweight and offer
high output, and in addition high efficiency is a crucial
factor for ensuring enough energy for overtaking; it is
no exaggeration to say that this factor determines victory
or defeat. A brushless DC motor was therefore used to
achieve both of these factors at a high level. From the
point of view of installing the motor in the vehicle, as
stated before, the diameter was designed to be within
100 mm and the full length within 200 mm, but
calculated from required output, the power density is at
least 8 kW/kg, which is a far more severe requirement
Motor
Engine
Front
Fig. 5
Energy storage
Li-ion 106 cells
Five-gear box
Five-gear box for motor
Water cooling
Water cooler
J/Box
Water pump
S-ECU
M/G ECU
PCU
Engine
oil cooler
Engine
Oil tank
Oil pump
Five-gear train
Oil cooling
Motor
Front
Fig. 4
Hybrid system
229
60 kW 45 N-m 21000 rpm
than with mass produced motors. Furthermore, we ran

KERS simulations with the driving data from past race,
and set the requirements that the motor could assist and
regenerate up to 60 kW at 13000 rpm and above so that
the system could ensure the minimum amount of
regeneration on any circuit. To do this, we reviewed the
entire motor design, including development of new
electromagnetic steel plates and magnetic materials,
review of winding techniques and using new cooling
techniques and rotor structures, with the result being the
achievement of a high power density of 8.7 kW/kg while
maintaining high efficiency (Fig. 6).
The rotor used a newly developed high coercive force
magnet (iHc = 1.1 MA/m at 160C) and at the same
time, in order to achieve high revolution speed of more
than double that of earlier motors, it used filament
winding with organic high strength fiber to prevent
magnet scattering, thus protecting the magnet
circumference. We additionally enhanced torque by 8%
by setting a magnetic field angle of in, out to
concentrate the magnetic field orientation of the magnets
to the polar center, and divided and stuck the magnets
along the length of the shaft to mitigate temperature
increases resulting from eddy current loss. For the
electromagnetic steel plate making up the stator, on a
base of iron and cobalt (49 Fe-49 Co-2 V), we succeeded
in reducing eddy current loss by making a 100 m thin
panel, in reducing hysteresis loss with a past-rolling

heating treatment, and in enhancing the volume fraction
with an oxidized insulation membrane. As a result,
saturation magnetic flux density was enhanced by 30%
and torque by 15% while iron loss was reduced by 60%
as compared to a conventional 200 m silicon steel panel
(Fe-Si). Lap winding was used for the stator winding for
high torque and low loss; connecting parts (turnaround
parts) at both ends of the stator were press-molded, using
injury-resistant copper wire, achieving an unprecedented
low connecting (turnaround) height (F i g . 8 ).
Furthermore, because smaller motors made it impossible
to keep the coil within the tolerable temperature during
driving of the motor in conventional water-cooled jackets
with stator housing, a stator structure was used such that
the coil ends were directly cooled by engine oil (Fig.
9), so no special radiator was needed. Additionally, the
agitation of cooling oil with the rotor increases friction,
so a cylindrical cover was used to prevent cooling oil
from sticking to the rotor, and the oil chamber of the
stator side was completely sealed. As a result, the
motors stand-alone efficiency averaged 95% at an
average motor speed of 20000 rpm during assist and
93% at an average of 16500 rpm during regeneration,
thereby achieving both high output and high efficiency
(Fig. 10).
Coil end
Three-phase terminal
Stator
Fig. 8
Fig. 6
Core assembly
Formula One KERS motor

Oil out
in
out
Prototype
Fig. 7
Oil out
Oil seal sleeve
Developed
Direction of magnetization
Fig. 9
230
Oil in
Section of motor with cooling oil flow
3.5. PCU
The PCU consists of an inverter unit (PDU) to drive
the motor and a voltage control unit (VCU) that allows
the voltage to be raised or lowered freely. The PCU is
used to supply the optimal amount of voltage and current
from the battery to the motor and to recharge the battery,
and it makes a great contribution to increasing motor
efficiency while making it more compact. An intelligent
power module (IPM), installed in the PDU, used a
special design to enhance compactness and reduce
electrical loss, while an SiC diode was used on-board for
the first time to reduce flywheel diode noise.
Development of the VCU started out with a boost
chopper form, but the PCU also faced severe on-board
installation requirements and had to be made smaller and
lighter, so system operation frequency was increased and
the form was changed to switched capacitor, which
offers the potential for a smaller, lighter reactor (voltage
step-up coil) (Fig. 11). An effort was made to make the
reactor smaller and lighter, not only by increasing the
system operation frequency as previously stated, but by
forming a 3D core by using dust core materials.
60
30 90%
93% 95%
20
97%
96%
98%
However, there is no history of using a switched

capacitor system on-board, and the technology was still
at the level of proving the principles behind it, so it was
difficult to establish the technology, and we were pressed
to take measures until just before moving to the circuit
for winter testing. The result of their effort was that the
PCU was 3.8 kg lighter than that of the RA1082.
The PCU of the RA1082 was located inside the
monocoque, but starting with the RA1089 it was
hurriedly put outside the monocoque, so that it faced
new collision requirements. To meet both the need for
lighter weight and collision requirements, we considered
changing the aluminum case used up to that point to a
CFRP case. In addition, the form of the PCU case was
reviewed every time an aerodynamic cowl component
was changed because the inside of the cowl, which
prioritized aerodynamic performance, had very little
installation space.
Under the initial specifications for the RA109K PCU,
the unit was cooled by a special cooling water circuit,
but to ensure high heat tolerance and low loss, we
developed a power module using an Si-C MOSFET
(metal oxide semiconductor field effect transistor), and
a new heat spreader (silver and diamond compound),
thereby creating a smaller, lighter PCU that shared the
engines cooling water; the plan was to release this at
the British Grand Prix, midway through the season.
Capacitor
10
Case
0
-10
VCU
Water jacket
-20
-30
96%
-40
87% 90%93%
-50
-60
9000
94%
95%
13000
17000
Speed [rpm]
Fig. 10
In brake
Output power [kW]
40
In boost
50
PCU
Control board
21000
Reactor
Fig. 12
Motor efficiency
View of PCU for race specification
PCU
VCU
PDU
P
U
V
W
N
Fig. 11
Race specification PCU block diagram
231
Side plate
3.6. Energy Storage

In order to meet the targets for the Formula One
KERS, targets for energy storage (ES) were set as below,
taking system efficiency into account.
(1) The system should have at least 70 kW at the ES output
end and at least 500 kJ of actual charge/discharge energy.
(2) To ensure collision safety, boxes should have at least as
much strength as required in the FIA side impact test,
with the internal construction having enough strength to
withstand at least 100 G-force.
The battery accounts for a high percentage of the
weight among KERS functional components, and to get
an output of 60 kW, a conventional lithium ion battery
weighs at least 30 kg, which reduces the competitiveness
of any race vehicle in which it is installed. For that
reason, a battery emphasizing output was newly
developed based on a lithium ion battery undergoing
R&D at that time for ordinary market vehicles. The
enhancement of power density continued until just before
the system for providing vehicles for the 2009 race
7000
Power density (W/kg)
6000
RA109K specifications
5000
RA1082/RA1089 specifications
4000
Original specifications
3000
Li-ion (Hi power type)

Li-ion (for EV)
2000
Ni-MH
Li-ion (for commercial use)
1000
50
Fig. 13
100
150
200
Energy density (kJ/kg)
250
Battery specifications
Sensor terminal unit

Contactor
Battery module
(nose)
Battery module
(keel)
Main switch
Fuse
Contactor
Sensor terminal unit
Fig. 14
Battery pack with control unit
season was established. Ultimately, the power density

advanced from approximately 5000 W/kg under the
RA1082 specifications to approximately 6000 W/kg for
the RA109K. This is about five times the power density
of a typical Ni-MH battery and about twice that of a
typical lithium ion battery (Fig. 13).
On the other hand, the battery of the RA109K was
split for installation into two areas, the nose and the keel,
to meet the collision requirements. To minimize the
impact on vehicle weight, the cooling system depended
on draft air from the front of the vehicle; for
waterproofing, cell connections and high voltage
terminals were molded when building the single module.
Furthermore, sensor terminal units, which constantly
monitor cell temperature and voltage, were placed in
each battery box (Fig. 14).
3.7. High Voltage Safety
Hondas basic stance on high voltage safety is that
high voltage components should be placed in an area that
is unlikely to be crushed even in the event of collision.
In races, however, for reasons of strategy and vehicle
packaging, KERS high voltage components were placed
outside the monocoque and could therefore potentially
be crushed. To ensure high voltage safety under these
conditions, we decided that it would be necessary to
actively announce our proposal for safety measures as
based on mass production experience when implementing
KERS, and Honda put together its own safety stance
prior to the FIAs move to implement unified safety
standards. The high voltage power source in KERS
consisted of the following three units.
High voltage battery
PCU capacitor (when recharging)
Motor (counter electromotive force is generated
when the motor is turned by external force)
To ensure high voltage safety, it is important that
KERS unit and cables be securely insulated.
Furthermore, if it is possible that the insulation can be
destroyed by accident or contact with another vehicle,
it is necessary that the supply of high voltage from the
above three units definitely be cut off. Consideration also
needed to be given to removal of the insulating structure
(the covers, case, and the like) when performing
maintenance. It is also important to provide a monitoring
function (ground fault detection). Based on the above
perspectives, safety was ensured with the specific
structure below.
3.7.1. Safety during collision
(a) Destruction of the insulating structure (case and covers)
by loads anticipated during collision was prevented.
Specifically, enough strength was ensured that the
insulating structure would not be destroyed under
conditions equivalent to those of the FIAs side impact
test.
(b) In the following cases, the electric charge in the PCU
capacitor was discharged at the same time the main
contactor was cut off.
When the FIA driving data inspection record unit that
232
is connected to the S-ECU has detected acceleration

beyond a set value
When the acceleration sensor on the PCU control board
has detected acceleration beyond a set value
Specifically, the electric charge was shifted between
the VCUs two capacitors by VCU switching, and
electric energy was changed to heat and released by
switching loss and conduction loss.
(c) A circuit that discharges electric charge from within the
PCU capacitor if the control power source supply is cut
off was located on the control board.
(d) A structure was used such that the high voltage connector
keep plate would be destroyed if more load than
anticipated were applied. The load required to destroy
the plate was set to be smaller than the load to break
cables, so that in the event of collision, connectors would
come off before high voltage wiring breaks (Fig. 15).
3.7.2. Safety during maintenance
(a) The electric charge in the PCU capacitor was discharged
at the same time the main contactor was cut off, when
the ignition was off.
(b) A structure was used such that an interlock mechanism
was used for the DC cable connector and high voltage
would not be applied to terminals when high voltage
connectors were removed. To remove or put on the high
voltage connector cover, one had to remove the control
harness, and if the control power source were cut off,
Connector
PCU
Keep plate
Fig. 15
Structure of high voltage connectors
the PCU would discharge and not apply high voltage to

connectors. A similar structure was also used on the
battery side to ensure safety.
3.7.3. Ground fault detection system
A detection circuit was mounted on the PCU control
board, and during KERS operation (i.e., when the main
connector is connected), the high voltage area ground
fault detection system was constantly put into operation.
Safety was ensured by making a warning lamp turn on
and simultaneously cutting off the main connector when
a ground fault is detected.
3.8. Control
KERS control system consisted of an S-ECU control
and PCU control (Fig. 16).
The S-ECU control calculated electric power
commands based on driver operations and monitored
motor output and per-lap assist energy. Assist occurred
when the driver pushed an assist button with the
accelerator pedal fully depressed, while energy
regeneration only occurred when braking. Motor output
was measured using values from torque sensors, and
other values, and if it was determined that a violation
had occurred, an output restriction penalty lasting a few
seconds was imposed. Assist energy was calculated from
the total motor output, and assist was stopped when 400
kJ was reached.
To use KERS to its full capacity, one has to use the
400 kJ efficiently and recover enough energy to make
that possible. Trial calculations indicated that to get 400
kJ of assist, one needs about 500 kJ of regenerative
energy, taking into account such factors as loss
depending on system efficiency and the energy
consumed by zero torque control. Because carbon disk
brakes have good brake force and can slow down
vehicles quickly, it is conceivable that if regenerative
torque is added, brake force will be unstable and stability
will decline, and we predicted that depending on the
circuit, road surface conditions and other factors, it may
not be possible to perform regeneration sufficiently and
thus there would not be enough energy. We therefore
Control board
PC for KERS
support system
Management
& Battery MPU
Battery
sensor terminal
FIA standard
ECU
Voltage
control unit
Battery
Power
drive unit
Motor
VCU control MPU
Motor control MPU

Power control unit
Fig. 16
Control system configuration
233
optimized each control component, including cooperative

control, and also optimized control of each functional
component by incorporating the various simulation
results into specifications. Through track tests, we also
endeavored to optimize the front-to-back brake balance
and distribution of regenerative braking and engine
braking. These tasks all require a high level of balance,
and track test settings were predicted to be difficult
because we are subject to S-ECU function limitations
and annual mileage limitations, but drivers were not
demanding any enhancements and potential issues were
considered unlikely for initial specifications.
It was decided that the operation of starting assist,
which is entrusted to the driver, would take effect at the
lowest vehicle speed, at which excessive driving force
would not result in wheel spin. This was to
simultaneously prevent both an unnecessary increase in
the driving force in low gear due to motor torque and a
decline in assist efficiency resulting from an increase in
driving resistance. Acceleration by assist is also affected
by gear ratios and aerodynamics settings, so the amount
of time and energy needed to overtake on each circuit
was calculated and the effectiveness clarified, and results
were used effectively in determining ES capacity and
designing how each functional component would be
cooled.
In the PCU control, each functional component was
controlled based on electric power commands from the
S-ECU control. This consisted of PCU management
control, motor control, VCU control and battery control,
and the necessary functions were created for racing use
based on existing specifications.
The system operations such as system stop and output
stop or power save were determined using PCU
management control, based on mediation between
electric power commands and system status.
Motor control and VCU control controlled motor
output. Because of the motors connection to the gearbox
in the RA1082, it was necessary to have some control
so that torque would be lost instantly (torque loss
control) when downshifting. When downshifting, on the
other hand, energy regeneration was performed while
decelerating, so torque loss control, during which energy
could not be regenerated, should be applied in as short
time as possible. If responsiveness were to be within 50
msec as stipulated in the regulations, the time it would
take to recover from torque loss control would be
wasted, since energy could not be regenerated then, so
we targeted responsiveness of within 15 msec to reduce
wasted time. It was decided that electric power
commands would change step-wise, at each change of
gear, from -60 to 0 kW or from 0 to -60 kW, and we
had to ensure controllability during great fluctuations of
revolution speed and load. For motor control, current
feedback control was performed using vector operations,
but the behavior of field component current in response
to step-wise electric power commands could not be fully
controlled and excessive current sometimes occurred. For
VCU control, output voltage feedback control was
performed, but this arrangement could not keep up with
sharp load fluctuations and excessive voltage frequently

occurred. The countermeasures against these issues
continued until just before track tests, and we were able
to achieve both stability and responsiveness to the target
of being within 15 msec by enhancing their techniques
for calculating target current values with motor control
and by applying voltage with feed-forward items in both
control parts.
In addition, we unstintingly added the technological
elements deemed necessary to the pursuit of speed, such
as performing zero torque control during ordinary driving
(without assist) with motor control to prevent
interference with engine acceleration. We furthermore
developed one-pulse control, which set the former pulse
width modulation (PWM) duty to 100% and reduced the
number of times switching occurred in order to minimize
PDU loss, and also developed VCU variable-voltage
control to maximize motor efficiency, with the aim of
releasing these developments during the 2009 winter
tests. These systems were also subjected to cooperative
control and systems were optimized simultaneously.
For battery control, we made it possible to use
lithium ion batteries safely and efficiently by monitoring
battery voltage and temperature as found at the sensor
terminal and accurately determining charge status as
based on calculation of the amount of electric power
charged and discharged.
4. Bench Tests and Track Tests

4.1. Bench Tests
Two Formula One engine test benches were used, to
check the system and its durability. A battery simulator
(BTS) and inverter stand (Fig. 17) were implemented
along with these benches to make them KERScompatible. A completely new BTS was implemented to
deal with a level of responsiveness unprecedented in
mass produced vehicles, making it possible to test the
overall system with all parts assembled (the engine,
motor, PCU, battery, and other parts) in coordination
with development of the RA1089 without experiencing
troubles under transient loads equivalent to actual
driving.
In this way, we substituted the bench for all testing
from confirming operation of each functional component
of KERS system to confirming the functionality of the
234
Battery simulator
Inverter stand
Fig. 17
Battery simulator and inverter stand

Table 3
overall system on an actual vehicle, including functional

assurance testing, and sought to bring KERS as a whole
to early fruition without the track tests that are so
severely restricted under the regulations.
Figure 18 shows the basic form of the test bench.
KERS was mounted with the engine as part of the jigs.
Using bench testing equipment, miscellaneous evaluation
tests were performed as shown in the events of Table
3, under steady state or transient load conditions. For the
transient load tests on the bench, the system underwent
mode operation, reproducing actual driving conditions
based on circuit data from actual driving experience, and
transient performance evaluation, particularly energy
management, measurement of the temperature of each
part, and the like, was performed, making it possible to
acquire useful data and obtain feedback in order to speed
up development. Figure 19 gives a typical example of
mode operation.
The data in the figure was acquired with the
following combination. This experience demonstrated
that it is possible to conduct function and durability
bench tests before actual circuit driving.
Driver: Jenson Button
Circuit: Monza
Car: RA1082 + KERS
Testing items on engine dyno bench
Function test
Durability test
-System operation / setting

-System output
-System efficiency
-Fail safe action
-Temperature measurement
-Vibration measurement
-Gear behavior
-Oil pressure measurement
-Tool operation check
-Engine and KERS parts durability test
-Energy management
With the RV (real vehicle) bench (1), furthermore, it

was possible to mount an actual Formula One
monocoque and cooling system on the bench and
perform a variety of tests in an actual vehicle
environment in addition to the above power train test
bench environment (Fig. 20). The RA1082 was actually
placed on the RV bench so that countermeasures for
electrical noise that was unique to KERS were completed
before driving on a circuit. Admittedly, because of the
RA1082 monocoque
Engine
Motor
Fig. 18
Fig. 20
Combination examination of engine and motor
Bench examination using RV bench with real

Formula One KERS car
Waveform1 11/25/08 18:25:08

18:24:26.418Grid: 5 s
Motor speed
Engine speed
Assist / Regenerative
Throttle opening
Circuit11 11/25/08 18:25:08

%r R1:, Out Lap (04:25.440),
Distance=2863.4 m
5500
500
1000
11-Start
Start-1
1500
11
1
10-11
5000
4500
10
9-10
98-9 8
4000
1-2
2000
3500
7-8
23
2-3
3-4
4-5
45
5-6
3000
7
18:24:00
18:24:05
18:24:10
18:24:15
18:24:20
18:24:25
18:24:30
18:24:35
18:24:40
18:24:45
18:24:50
18:24:55
18:25:00
18:25:05
18:25:10
Lap 83f
%r R1
nEngine
rThrottleL=
%r R1
%
nKERS
18:25:20
In Lap
%r R1
rpm
18:25:15
rpm
%r R1
MKERSMotorRequestFinal =
Monza course layout
Nm
ATLAS 8.17.3537
ATLAS 8.17.3537
Fig. 19
Circuit mode test on dyno
235
6-7
6 2500

Table 4
Event
Durability test results of dyno

Purpose
Mileage
KE-01 Check environment and first issue parts
238.4 km
KE-01.5 Check track-1 specification parts

KE-02
465.4 km
Track test simulation

(for Jerez circuit)
611.0 km
KE-03 Check track-2 specification parts
748.8 km
KE-04 Check race specification parts
1351.9 km
Check engine gear durability with extreme circuit mode

(test stopped due to schedule)
production schedule, the RV bench was not used with

subsequent vehicles, but power train durability was tested
on the bench prior to circuit driving and mileage
equivalent to more than two race events (1350 km) was
assured with all functional components (Table 4).
4.2. Track Tests
In late April 2008, HRF1 conducted the shake down
of the RA1082, the first machine with KERS, and
subsequently conducted actual driving test four times
with the RA1082 and twice with the RA1089.
RA1082, the first actual vehicle, underwent
modifications and the vehicle was finally complete four
days prior to shake down, which cut deeply into the
originally planned two weeks of system checks, and with
vehicle settings, additional modifications, and the like,
there was in effect a preparation period of only about
two days. The track test members had to deal with the
system operations check and initial troubles in a short
period of time and were not able to get to complete
operation by the prescribed date. For that reason, in the
initial shake down we decided to maintain the electric
current status, and to limit themselves to checks of
system safety and the function of functional components.
At HRF1, we were not able to confirm system operation
by chassis dyno as with the RV bench, so all we could
do was to check system operation by firing-up the engine
in a factory. However, to fire-up the engine required a
large number of engineers and mechanics and a launch
sequence starting several hours in advance. For the sake
of engine durability, moreover, it is not possible to let
it run for long periods with no load, so we could only
get the desired data in a very short period of time, and
it was difficult to do system check of KERS. Ordinarily
fire-up only takes place once or twice before shake
down, but fire-up was conducted dozens of times with
the RA1082. The first shake down was not done under
the maximum load allowed by regulations, but the run
was memorable for being the first KERS actual driving
test among all 10 Formula One teams. At the factory,
we subsequently responded to nonconformities found
during shake down, and a week later conducted output
and regeneration tests on the Silverstone Circuit,
confirming that KERS functioned effectively under
Formula One conditions. The team subsequently made
more enhancements, conducted a private test in July and
566.1 km
took part in a joint test at the Jerez circuit in Spain in

September, thus beginning the first regular run on a
circuit (Fig. 21). The benefits of KERS were verified
at the maximum load under regulations at the Jerez
circuit; the following benefits were confirmed.
(1) As Fig. 22 indicates, lap times were reduced by
approximately 0.4 seconds when assist with energy of
about 400 kJ per lap was applied.
(2) As shown in Fig. 23, car speed increased by 7 km/h
with distance gain of 7.8 m (1.6 car lengths) with
continuous assist of 324 kJ on one straightaway.
Because the amount of tire skidding and aerodynamic
specifications, including speed during assist and road
surface conditions, differed from initial simulation
results, there was some discrepancy in values, but the
effectiveness of KERS was sufficiently demonstrated.
A driver who experienced the full power assist of
KERS commented, I was impressed by how amazingly
236
Fig. 21
KERS car and engineers
83.2
83.1
Lap time [sec]
KE-05
83.0
82.9
0.4 sec
82.8
82.7
82.6
0
100
200
300
400
Assist energy [kJ]
Fig. 22
Lap time against the assist energy
500
Motor power [kW]
fast I approached the hairpin when the assist kicked in.

It makes acceleration from the engine alone feel as if
carrying a heavy weight.
Then the RA1089 was completed in November 2008
ahead of winter testing. However, the system of the
RA1089, mounted with a new VCU as described earlier,
is very different from that of the RA1082, and so noise
frequently resulted in system failures, and together with
bench analysis, day after day was spent in noise analysis
on actual vehicles and discussions on countermeasures.
For the RA1089, the HRF1 engineers, mechanics and
other track test members worked together to conduct
more than 100 fire-ups, more than we did with the
RA1082, with the result that we overcame all the
troubles. This was the result of all the members, from
the UK and Japan, coming together to solve some very
difficult issues.
Then on November 28 th, the RA1089 underwent
shake down on the straight course of Santa Pod in the
UK (Fig. 24), confirming that all systems were operating
normally, and since the judgment was made that the
system could be used on a circuit, the RA1089 and its
equipment were loaded onto a trailer on the evening of
December 4 th as the test vehicle for joint testing on
80
60
40
20
0
-20
-40
-60
-80
Track test results of KERS car

Driver
(chassis)
A. Wurz
(RA1082)
Date
Circuit
29-Apr
Santa Pod
7-May
Silverstone
(short course)
J. Rossiter
(RA1082)
29-Jul
Silverstone
(Stow School)
M. Conway
(RA1082)
16-Sep
Jerez*
19-Sep
Jerez*
13-Nov
Kemble
28-Nov
M. Conway
(RA1082)
A. Wurz
(RA1082)
Mileage
Notes
2.4 km Shake down

Shake down
25.0 km 25 kW assist
35 kW recharge
Shake down
60 kW recharge
60 kW assist
88.6 km
60 kW recharge
60 kW assist
225.8 km
60 kW recharge
A. Davidson
10.5 km Shake down
(RA1089)
Shake down
A. Davidson
(RA1089)
25 kW recharge
*: Full race track
Santa Pod
December 9th at Jerez. On December 5th, however, it was

announced that Honda was pulling out of Formula One
racing, and all activities ceased, which meant that the
shake down of November 28th was the last track test of
Hondas third-era Formula One activities. The record of
KERS circuit track tests is given in Table 5.
5. Conclusion
Assisted by 60 kW for 5.4 sec
310
Without KERS
With KERS
290
Car speed [km/h]
Table 5
270
250
230
Car speed = 7.1 km/h

Without KERS = 283.8 km/h
With KERS
= 290.9 km/h
210
190
170
150
25
26
27
28
29
30
31
32
33
34
35
Lap time [sec]
Fig. 23
Car speed comparison
The effort to develop an electrical hybrid system that

intended to introduce KERS in 2009 produced the
following results.
(1) A compact lightweight hybrid system was developed and
a car was produced with performance to realize maximum
output of 60 kW and 400 kJ of assist on each lap of a
circuit.
(2) Honda became the first in the world, beating out other
companies, to conduct actual driving tests of KERS on a
circuit, and demonstrated faster lap times with KERS
(about 0.4 sec faster with 400 kJ of assist) and the effect
of an overtaking boost (about 7 km/h with assist of 324
kJ).
(3) The team demonstrated the safety of KERS in all
processes in the development of functional components
and the vehicle, as well as in circuit driving.
Afterword
Fig. 24
Launch of the RA1089 with KERS system
The development effort was confusing at first,

because several things were happening simultaneously:
R & D on completely new KERS functional components,
all staff members concerned with KERS the technical
development staff, race engineers, the mechanics, other
site managers, vehicle production staff, and vehicle
electrical equipment staff in the UK and Japan
receiving training and mastering high voltage systems,
cooperation with the FIA on safety, and working to
revise regulations. But the fact that the development took
place in such an unusually short period of time (nine
237
months from the start of development until first driving,

and 15 months until shake down of the racing specs
vehicle), is a tribute to the joint development system
with the HRF1 members, overcoming language barriers
and national borders and showing mutual respect, and to
the joining of minds of everyone concerned, including
cooperating manufacturers, those involved in distribution,
business travel and translation, and local staff. This is
something all members are proud of.
Reference
R&D Technical Review, 2009, F1 Special (The Third
Era Activities), P. 95-100
Author
Masataka YOSHIDA
238
Masato KITA
Hirofumi ATARASHI
Descriptions of Material Technologies
High-performance Shell Bearing

from New Material
Kiyoshi ITO*
Makoto ASAMI*
Hiroshi KOINUMA*
ABSTRACT
A material that combines high thermal conductivity, seizure toughness, and high strength was developed by
enhancing hardening methods and fine homogenization in the deposit phase of a Corson alloy. This material was used
as the back metal for shell bearings, it was used in parts as a high strength high thermal conductivity shell bearing,
and it achieved consecutive use in events under one-engine, two-race rules.
1. Introduction
Table 1
Shell bearings for use in Formula One racing engines

require sliding durability to a high PV value (surface
pressure P x peripheral velocity V: 2000 MPa m/s or
above) in order to sustain the inertial force and explosion
force experienced under high speed conditions. When
concerns about insufficient durability existed, there were
sometimes cases when engine power performance would
be suppressed in order to assure durability.
Figure 1 shows the cross-sectional structure of shell
bearings. Conventional shell bearings have a threelayered structure. On the other hand, shell bearings for
use in Formula One racing require sliding durability that
is enhanced to an ultimate degree, and therefore Cu alloy
back metal with a two-layered structure that emphasizes
heat drawing performance was adopted. With this
structure, the material properties of back metal Cu alloy
have a substantial influence on shell bearing durability,
and it was therefore decided to aim for development of
a new Cu alloy.
2-layer metal
3-layer metal
Pb-plating
Overlay
Pb-plating
Cu-alloy
Lining
Steel
Cu-alloy
Back metal
Fig. 1
Cross-sections of each metal
Main properties of several Cu-alloys
Electric conductivity
UTS
IACS (%)
MPa
BeCu50
45-50
690 - 800
BeCu25
25
>1000
65
490
CF-2 (Cr-Cu alloy)
40-45
690
NC50 (Corson alloy)
Material
0.2%YS
MPa
590
2. Development of New Cu Alloy for Use in

Shell Bearings
A Cu alloy for use in shell bearing back metal should
have the following three properties in combination:
(1) Thermal conductivity: Efficient cooling of heat generated
by sliding
(2) Sliding performance: Inhibit seizing against nitriding
crank shaft
(3) High strength: Assure contact pressure with connecting
rod big end
2.1. Selection of Alloy
Table 1 (1) shows the main properties of typical Cu
alloys. Cu alloys with high thermal conductivity as
represented by electric conductivity (IACS%) are Be-Cu
alloys (hereafter BeCu), Cr-Cu alloy (hereafter CF-2),
and Corson alloy (hereafter NC50)(2).
These Cu alloys were subjected to a test of
toughness against seizure with nitriding steel using
the sliding rig test of Daido Metal Co., Ltd. The

240
results are shown in Fig. 2. NC50 has greater seizure

toughness than BeCu alloy, which makes an excellent
sliding material for use in a bush and is commonly
found in that application. NC50 is a Cu alloy of the
deposit strengthening type that has Ni, Si, and Cr as
main additive constituents. The deposit phases (Ni 2Si,
Cr 3Si) are thought to suppress its seizure against steel.
Material development was carried out using NC50, a
material that achieves a balance of high thermal
conductivity and seizure toughness, as the base.
2.2. Balance of Thermal Conductivity and Strength
The use of NC50 material as shell bearing back
metal to assure contact pressure against the
connecting rod big end under high PV conditions
necessitates strength.
Greater strength can be achieved by increasing the
deposit phase, but increasing the quantity of additive
elements results in a reduction in thermal
conductivity. Instead of changing the quantity of
additive elements, therefore, it was decided to seek
higher strength by adding hardening methods
(extrusion method, multi drawing method) to the
process. In addition, the heat treatment was optimized
(by multi heat treatment) in order to achieve a finer,
homogenized deposit organization, and thermal
conductivity was increased. Relative to NC50 (the
base material), the developed material NC50ES
achieved a 30% increase in strength and a 12%
increase in electric conductivity (Fig. 3).
3. Confirming the Effectiveness of Shell

Bearing Parts
3.1. Shell Bearing Rig Test
Prototype shell bearings were fabricated with
NC50ES material as the back metal. These were
subjected to performance evaluation using a shell bearing
tester(3). Figure 4 shows the results.
The horizontal axis shows the flow rate of the oil
supply from the shaft to the shell bearing, while the
vertical axis shows the shell bearing temperature
(measured by a thermocouple embedded inside the shell
bearing). It was confirmed that NC50ES has a shell
bearing temperature 3-5C lower than that of
conventional materials. This is conjectured to be the
effect of increased thermal conductivity in the back
metal material.
3.2. Engine Tests
Figure 5 shows the state of damage to the shell
bearing sliding surface after conducting engine endurance
testing. The condition of the remaining overlay on the
shell bearing surface (No. 1 cylinder on the connecting
rod beam side) is shown three-dimensionally. There is
more remaining overlay (area) on the NC50ES shell
bearing. It is speculated that when the temperature of the
shell bearing as a whole is reduced, this also causes the
temperature of the overlay to decrease, resulting in an
increase in wearing toughness. This material also made
it through durability testing in modes matching the most
Bearing temperature (C)
15
10
5
0
Fig. 2
BeCu25
CF-2
NC50
155
150
145
: Current material
: NC50ES
5C
140
135
130
125
Test condition
Speed
10000 rpm
Load
10 kN
Oil temp 100 C
120
200
Fig. 4
3C
300
400
500
Oil flow rate (cm3/min)
Seizure toughness of Cu alloys
70
600
Result of test on bearing tester
Good
Cr-Cu alloy
60
IACS (%)
Seizure load (MPa)
20
BeCu50
Current material
NC50ES
50
BeCu50
NC50ES
: 20.0
12%
NC50
Overlay thickness ( m)
: 17.5 20.0
: 15.0 17.5
40
: 12.5 15.0
30%
: 10.0 12.5
30
: 7.5 10.0
: 5.0 7.5
BeCu25
: 2.5 5.0
20
400
600
800
1000
#1UP-side
0.2% proof stress (MPa)
Fig. 3
Position of NC50ES material
Fig. 5
241
0 2.5
#1UP-side
Overlay thickness after engine endurance test
High-performance Shell Bearing from New Material
demanding high-load conditions on the circuit, and

produced satisfactory results without adhering, seizing,
or cracking.
4. Conclusion
The new Cu alloy NC50ES was developed for use
as shell bearing back metal. Shell bearing parts were
fabricated from the material.
The reliability of the shell bearings was enhanced
with respect to the high-speed, long-distance assurance
of Formula One engine regulations. The shell bearings
achieved continuous use in events under one-engine,
two-race rules as well as straight-end 19000 rpm
performance enhancement.
References
(1) http://www.yamatogokin.co.jp, 2009/03/27
(2) JP, 3563315, B (2004)
(3) Machida, K., Takahashi, S., Ueshima, H., Fujiki, K.,
Proposal of New Bearing Simulation Rig Tester for High
Efficiency Engine Bearing Development, Honda R&D
Technical Review, Vol. 18, No. 2, p. 131-139
Acknowledgement
This occasion is taken to express the warmest
gratitude to Miyoshi Gokin Kogyo Co., Ltd., and
Yamato Gokin Co., Ltd., which cooperated generously
in the development of this material.
Author
Kiyoshi ITO
242
Makoto ASAMI
Hiroshi KOINUMA
Development of Hollow Crankshaft
Kiyonobu MIZOUE*
Yasushi KAWAHITO*
Ken MIZOGAWA*
ABSTRACT
A hollow crankshaft of light weight and high quality was realized through the use of friction welding. The oil
passage in the journal was made a hollow structure, and this had the effect of increasing average oil pressure by 500
kPa (reducing the pressure drop), reducing the pulsation amplitude by 400 kPa, and raising the minimum oil pressure
by 600 kPa. The connecting rod bearing clearance was also adjusted and confirmed to yield 1 kW of performance
enhancement and over 1200 km of durability as well as curbing bearing friction. Although this new crankshaft was
never used in a race, this developed technology has provided the prospect of the first hollow crankshaft of 10 kg or
less for practical use in the Honda Formula One.
1. Introduction
The durable reliability of the crankshaft is essential
in the high speed and high power development of
Formula One engines. Accordingly, this part is also
considered to require lower friction and lighter weight
in the pin and journal rotating portions as well as higher
stiffness with respect to torsion resonance. The
development of a lightweight, high stiffness hollow
crankshaft has been underway since the end of 2003 as
a horizontal deployment of the hollow structure of the
connecting rod in other parts. Meanwhile, the
International Automobile Federation (FIA) included a
provision in its regulations that no welding would be
permitted between the front and rear main bearing
journals from 2006. This would make it impossible to
use a welded hollow crankshaft in cars entered in FIA
races. Honda therefore established the structure
manufacturing technology for a hollow crankshaft in
2005, and set out to install it in a racing vehicle for use
that year.
The creation of a hollow crankshaft required joining
technology capable of providing stable strength and
precision in the junctions that are of greatest importance.
It also required manufacturing technology capable of
assuring quality by enhancing the accuracy of machining
on junctions as well as by removing burrs and the like.
joining methods were selected. The manufacture of a

high-quality, high-precision V10 hollow crankshaft
involved enhancing the accuracy of friction welding
(hereafter FW) on five journals and implementing quality
assurance regarding residual burrs at joins, and
deformation from machining and nitriding. Figure 1
shows the hollow structure and an external view of it.
2.1. Hollow Structure Specifications
The location and optimum shape of joins in the
hollow were subjected to CAE verification, from which
the joins at the center of the journal, as shown in Fig.
2, were selected.
The hollow pin with a diameter of 34 mm (journal
wall thickness: 6 mm) opened the way to satisfying the
strength requirements.
The bending stress, caused by explosion force, is
2. Establishment of Hollow Structure Joining

Technology and Quality Assurance
In order to create this hollow structure, the joining
locations and strength feasibility were verified and
243
Fig. 1
Section of hollow crankshaft
Development of Hollow Crankshaft
mitigated by the hollow structure, so that the stress was

lower than in a solid. Furthermore, the inertial force was
little influenced by the hollow structure, and the hollow
pin with 34 mm diameter had lower stress than the solid
pin with 33 mm diameter.
2.2. Joining Technology and Strength Verification
The three manufacturing techniques of electron beam
welding (EBW), diffusion bonding, and FW were
subjected to comparative evaluation by rig testing of
fatigue strength. FW, which yielded join strength equal
to the strength of the base material, was decided on for
the specifications.
2.3. FW of Five Journals
The basic conditions were set at a speed of 1900 rpm,
pre-force of 10 tons, a pre-heat time of 9.5 sec, an upset
force of 30 tons, and upset length of 9 mm. The surface
roughness of the join surfaces was enhanced and acetone
degreasing was done during the cleaning prior to joining
(Fig. 3).
Key points in determining the FW conditions
included shaping the flash form to rise up as much as
possible and ensuring that the media is kept from being
worked in.
The chucks for gripping the workpieces for the series
of five joins were boremachined together to align their
center lines, enhancing the accuracy of butting by the
rotating side as well as the clamping stiffness on the
fixed side. The equipment peculiarity was addressed
using adjusting shims to correct for eccentricity of axis.
With the join conditions established, the series of five
FW joins (joins made while undergoing 180 degree
phase reversal) was made, achieving an offset of 0.35
mm at the largest with respect to the target offset of 0.5
mm (Fig. 4).
2.4. Quality Assurance of Hollow Portion

The hollow portions joined by FW also serve as oil
passages, making it necessary to conduct thoroughgoing
quality assurance regarding the removal of burrs. First
of all, the FW conditions were set so as to stabilize the
flashes and control their shape. As shown in Fig. 5,
scale, machining burrs, and spatter are removed after
nitriding by barrel finishing with vibration using steel
spheres and cleaning by cavitation. All were then
checked by microscope and guaranteed. Figure 6 shows
the nitriding quality of the hollow interior. Although the
central portion of the flashes were not nitrided, the joins
and the hollow portion were confirmed to be uniformly
nitrided. Rig testing of the 2.5 cylinder joined items
confirmed strength equal to the base material.
2.5. Heat Treatment and Manufacturing Completion
Accuracy Assurance
Deformation from heat treatment was addressed by
using crankshaft pin milling items as dedicated jigs and
setting conditions by corrective hardening. When the
offset exceeded 1.0 mm, the bend was trued up. For
working accuracy, a base adjustment eccentricity center
was used and uneven thicknesses in the finish of the
hollow portion were minimized. The target for wall
thickness of 6 0.75 mm in the journal hollow portion
was achieved with 6 0.3 mm and a maximum uneven
thickness of 0.55 mm.
Fig. 4
Friction bonding
FW bonding
Fig. 2
Journal joint
Rotation
Hold
Fig. 5
After cleaning
Pre-stress
Flash
Flash
Upset
Rear
Front
Fig. 6
Nitride of FW section
Flash
Fig. 3
FW bonding
244
3. Results
A weight reduction of 0.8 kg was achieved, going
from 6.55 kg (10.3 kg with counterweight) for the solid
crankshaft to 5.75 kg (9.5 kg with counterweight) for the
hollow, and making this the first Honda Formula One
crankshaft to realize a weight of 10 kg or less. The
effects of enhanced lubrication with the hollow
crankshaft include recovery of 500 kPa in the
instantaneous minimum oil pressure, which had shown
negative pressure because of oil pressure pulsation, and
prospects for durability.
4. Conclusions
The technology was established for manufacturing a
hollow crankshaft that would satisfy the properties
required for use in a high-speed, high-power Formula
One racing engine.
The crankshaft was made hollow using a highly
reliable FW technique, and production engineering
capable of guaranteeing crankshaft accuracy as well as
the cleanliness to use the hollow portion for oil passages
was established.
It has been verified that the developed technology
reduced weight 8% relative to the conventional
crankshaft and had the effect of oil pressure pulsation
reduction that could not be obtained with the
conventional structure, yielding the prospect of durability
for practical use. Regulations prohibiting welded
crankshaft structures went into effect in 2006, and this
hollow crankshaft was never used in actual competition.
Acknowledgements
This occasion is taken to express the warmest
gratitude to Takayuki Ohnuma and Nobuki Matsuo of
Honda Engineering Co., Ltd., for their cooperation with
FW joining technology and quality assurance.
Author
Kiyonobu MIZOUE
245
Yasushi KAWAHITO
Ken MIZOGAWA
Development of Metal Matrix Composite Piston
Kazuhisa ISHIZUKA*
Katsuhito KUSAKAI*
Nobuyuki IMAI*
ABSTRACT
A metal matrix composite that can be expected to have outstanding strength at elevated temperatures was applied
to reduce the piston weight. In composites, it is generally difficult to cope with both strength and toughness. To solve
this issue, a powder alloy was applied using the mechanical alloying process, and the manufacturing process and surface
treatment were optimized. This achieved a 16% weight reduction compared to the conventional material, AA2618. In
addition, this enabled the Formula One engine speed to be increased by 400 rpm.
1. Introduction
Table 1
Reducing the inertial weight of reciprocating systems

is the most important subject to increase engine speeds,
and it was necessary to develop new materials to realize
lightweight and high strength parts for Formula One
pistons.
Aluminum alloy metal matrix composites (MMC)
strengthened by ceramic dispersion can be expected to
have high specific strength, high stiffness, and better
characteristics at elevated temperatures. However, it was
necessary to overcome the issues that toughness and
ductility are generally low, and strength anisotropy which
comes from the restrictions on the forging process. To
solve these issues, a powder alloy was applied using the
mechanical alloying process, and the characteristics were
enhanced by modifying the forging and heat treatment
processes.
2. Developed Technologies
2.1. MMC Materials (AMC225XE) for Pistons
Table 1 compares the physical properties of MMC
and the conventional material, AA2618(1). MMC uses
AA2124 as the base alloy, and adds SiC with a particle
size of 3 m in a 25% ratio by volume as the dispersed
ceramics. The SiC is dispersed and alloyed with the base
alloy powder for the prescribed time using a high energy
mill. Figure 1 shows a metallographic image of the
MMC, and a mapped image of the carbon from the SiC
obtained using EPMA. Here, a countless number of
submicron size particles can be observed, dispersed
uniformly between large particles with several m. These
submicron particles are thought to be the SiC crushed
by the mechanical alloying process and dispersed within
Material
AA2618
MMC
Properties of piston material
Modulus
GPa
74.6
113.0
Density
g/cm3
2.76
2.89
CTE
ppm/K
22.6
15.5
CTC
W/mK
165
129
the matrix. This material realizes high fatigue strength

at elevated temperatures and better toughness and
ductility by not only precipitation of the base alloy and
the law of mixture, but also dispersion strengthening
mechanism of the nanosize particles.
2.2. Forging and Heat Treatment Technologies
In general, powder metal materials are worked after
sintering to achieve good mechanical properties. In early
days, pistons were forged using previously extruded bars,
but there was the issue that anisotropy in the piston roof
was serious, and the expected characteristics could not
be obtained. The issue of anisotropy was successfully
resolved by forging pistons directly from HIP (Hot
25 m
1 m
Fig. 1

246
Metallograph of MMC and distribution of carbon
Isostatic Pressing) materials for which 3-dimensional

isotropy can be expected.
The required forgeability was achieved by the lowspeed hydraulic forging press. High cooling rate during
quenching was ensured by using thin materials with a
near-net shape, which realized good mechanical
characteristics. Application of near-net materials also
helped reduce machining of MMC, which is a difficultto-cut material.
2.3. Surface Treatment
The lower coefficient of thermal expansion due to
SiC dispersion leads to an increase in the clearance
between the piston and the cylinder. This causes an issue
in securing reliability of these sliding parts. This issue
was overcome by applying an electroplated hard film and
a resin coating containing dispersed hard particles to the
skirt and top land, which secured the required durability
and reliability.
4. Conclusion
A high strength piston material was successfully
developed by optimizing the composition and
manufacturing method of MMC manufactured using the
mechanical alloying method. Application of this material
reduced the piston weight by 16% compared to the
conventional material AA2618, and achieved an increase
in the engine speed of 400 rpm.
Reference
(1) http:/www.amc-mmc.co.uk/
3. Effects on Performance
Figures 2 and 3 show the mechanical characteristics
of a test sample taken from the roof section of the piston
material. Both high fatigue strength and ductility were
achieved in the piston operating temperature range of
200 to 300 C.
Use of this material realized lightweight and high
strength pistons that were applied to races from the first
GP of 2004.
Specific fatigue strength

(Mpacm3/g)
90
80
AA2618
70
MMC
60
50
40
30
20
150
Fig. 2
200
350
250
300
Temperature (C)
Specific fatigue strength of developed material

60
Elongation (%)
50
AA2618
40
MMC
Author
30
20
10
0
0
Fig. 3
50
100 150 200 250

Temperature (C)
300
Elongation of developed material
350
Kazuhisa ISHIZUKA
247
Katsuhito KUSAKAI
Nobuyuki IMAI
Development of Lightweight Titanium-aluminide

Piston Pin
Sadami MINATO*
Yasunori ONAHA*
ABSTRACT
An intermetallic titanium aluminide material displaying increased fracture toughness and fatigue strength, and a
process enabling the formation of the material into components, have been developed. Exploiting the materials low
specific gravity and high modulus of elasticity, it was employed to manufacture a piston pin, enabling the achievement
of a 17% reduction in weight and a 28% increase in rigidity against a conventional nitriding steel piston pin. The
developed piston pin has contributed to increased speed, power and long-distance reliability in V10 engines, which
have been required to complete two race distances since 2005.
1. Introduction
In addition to possessing high specific stiffness at the
upper limit of Formula One material regulations and
excellent high-temperature strength, titanium aluminide
(TiAl) also displays excellent fatigue strength at ambient
(or room) temperatures. For these reasons, manufacturers
have been attempting to extend the use of the material
to the main reciprocating components, which are the
subject of a constant quest for weight savings. However,
there have been concerns over the low fracture toughness
and the quality of the European materials employed in
engine valves(1), and other significant issues have arisen,
including insufficient resources for development due to
oligopolistic supply.
The aim of the project discussed in this paper was
to develop an original high-quality TiAl material by
balancing fracture toughness with strength, and to
contribute to the achievement of increased engine speed
by reducing reciprocating mass through the application
of the material in piston pins.
2. Developed Technology
2.1. Material Design
In order to increase the strength and fracture
toughness of TiAl, the first important step is to use an
extrusion process to refine the coarse lamellar
microstructure produced by ingot casting. The piston pin
material must be produced at a diameter of 19 mm, and

therefore necessarily possesses an insufficient extrusion
ratio. For this reason, refinement of the microstructure
during working was promoted by adopting a chemical
composition design in which an intermetallic phase
(TiAl), phase (hcp-Ti), and phase (bcc-Ti) coexist
at 1150 C, the final hot working temperature.
In addition, the exploitation of the metastable phase
in the final microstructure, which promotes strength and
fracture toughness, is a factor in the achievement of
increased ductility. Therefore, the chemical composition
was designed to produce large quantities of the phase
material in a stable state at the annealing temperature of
1000 C. Figure 1 shows the results of a comparison
of microstructures and phase volume fractions using
Thermo Calc(2).
2.2. Forming Process
A process of cold crucible induction melting and
continuous casting, in which the molten developed
material is solidified while being subjected to
electromagnetic mixing, is used to obtain a uniform ingot
microstructure.
TiAl possesses a high level of resistance to heat
deformation. Stainless steel sheets are generally used for
canning to control the decline in the work temperature
and the material undergoes hot plastic working between
1200 C and 1350 C(1), (3). For the developed material,
on the other hand, a dedicated antioxidant was selected.

248
The use of this antioxidant in combination with a die

glass lubricant reduces the materials deformation
resistance, enabling hard working at an extrusion ratio
of 58 at a lower temperature than specified above,
between 1150 C and 1200 C. Refinement of the
microstructure during the formation process is promoted
in this range.
Finally, the material is subjected to annealing at 1000
C for 2 hours, followed by forced cooling (Ar gas
cooling) in order to obtain a phase between 20% and
35%. By this means, stable elongation of 2% or more
at ambient temperatures has been balanced with fatigue
(a) Conventional TiAl alloy
strength of 800 MPa or more (rotating bend) at ambient

temperatures. Figure 2 shows the microstructure of the
material (phase volume fraction) and its effect on the
materials mechanical properties under each of the
conditions used in the formation process.
3. Component Specifications
CAE and stress measurements were conducted to
determine the piston pin design in view of the strength
of the piston pin itself and the effect to the piston, and
the component design was changed from the
conventional tube type used with nitriding steel to a solid
type. The solid design increased rigidity in the collapsing
direction by 28% and reduced stress on the piston pin
boss, which had previously represented a reliability issue,
by 6%, while the application of the new material enabled
the achievement of a 17% weight saving. Figure 3
shows comparison results of deformation in CAE
analysis.
In addition, given the decline in hardness of the base
material with the use of TiAl, a high-hardness diamondlike carbon (DLC) coating with a 7 m Cr 2 N film
formed by PVD sputtering as a backup layer was
employed as the surface treatment, enabling adhesion
strength and scratch resistance to be increased.
(b) Developed TiAl alloy

Annealing temp.
Fig. 1
(a) Conventional steel
Extrusion temp.
Fig. 3
(b) Developed TiAl
Piston pin deformation (Catia V5 CAE)
Comparison of phase ratio (Thermo-Calc)
4. Performance Results
Phase volume
fraction [%]
5.0
wb
800
4.0
Elong
600
3.0
400
2.0
200
1.0
0.0
Elongation [%]
Rotating bend
fatigue wb [MPa]
1000
2 (Ti3Al)
20
(TiAl)
40
(bcc-Ti)
60
Ext. temp.
1200 C
1150 C
1150 C
Ext. ratio
58
Annealing
As ext.
As ext.
1000 C
1000 C
Fig. 2
1150 C
Effect of phase structure on mechanical

properties
The solid TiAl piston pin was evaluated in high-load

durability mode in an engine. As an effect of the
achieved weight savings, the level of conrod bearing
wear was equivalent to that associated with a
conventional nitriding steel piston pin, even though the
engine speed had been increased by 200 rpm.
In addition, the high rigidity of the developed
material reduced the level of piston pin DLC wear and
piston boss cracks, which had previously represented a
concern in terms of reliability over long distances.
Figure 4 compares the film thickness of the remaining
overlay layer on the surface of conrod bearings produced
using the developed and conventional materials following
the endurance test. Figure 5 shows a comparison of the
condition of the piston pin DLC for the developed TiAl
and conventional steel following the endurance test.
249
Development of Lightweight Titanium-aluminide Piston Pin
: 20.0 : 17.5 - 20.0

: 15.0 - 17.5
: 12.5 - 15.0
: 10.0 - 12.5
: 7.5 - 10.0
: 5.0 - 7.5
: 2.5 - 5.0
: 0.0 - 2.5
:
- 0.0
[Unit : m]

(18800 rpm/1503 km)
Fig. 4
: 20.0 : 17.5 - 20.0

: 15.0 - 17.5
: 12.5 - 15.0
: 10.0 - 12.5
: 7.5 - 10.0
: 5.0 - 7.5
: 2.5 - 5.0
: 0.0 - 2.5
:
- 0.0
[Unit : m]
(b) Developed TiAl

(19000 rpm/1485 km)
Overlay thickness of connecting rod bearing

after endurance test
DLC wear
(18800 rpm/1503 km)
Fig. 5
(b) Developed TiAl

(19000 rpm/1485 km)
DLC condition of piston pin after endurance test
5. Conclusion
Control of microstructure by means of chemical
composition design and management of the formation
process has enabled the development of a TiAl material
that balances high fatigue strength with excellent fracture
toughness (displaying an elongation of 2% or more). The
use of this material to manufacture a piston pin has
resulted in the achievement of a 17% reduction in
weight, and has contributed to a 200 rpm increase in
engine speed, and, by means of a reduction in the level
of deformation of the pin, to reduced DLC wear and
increased reliability in the piston pin boss section.
Acknowledgments
The authors wish to take this opportunity to offer
their sincere thanks to the staff of the Titanium Metals
Division of Kobe Steel, Ltd., and the Kobelco Research
Institute, Inc., who generously assisted in the
development of the material and product discussed in
this paper.
Author
References
(1) Clemens, H., Kestler, H., Eberhardt, N., Knabl, W.:
PROCESSING OF TiAl BASED ALLOYS ON AN
INDUSTRIAL SCALE, The Minerals, Metals &
Materials Society, Gamma Titanium Aluminides 1999,
p. 209-223 (1999)
(2) http://www.thermocalc.se/, 2009/04/02
(3) JP, Patent Application No. 2000-24748
Sadami MINATO
250
Yasunori ONAHA
Development of Hollow Titanium Connecting Rod
Sadami MINATO*
Ken MIZOGAWA*
ABSTRACT
It is necessary to reduce the reciprocating mass in order to increase the engine speed and power of Formula One
engines. The project discussed in this paper therefore set out to increase the section modulus of the shaft of the
connecting rod while maintaining its rigidity and achieving weight savings. To this end, the diffusion bonding method
was optimized, and a process of manufacturing a hollow connecting rod was developed. The developed connecting
rod is lighter in weight and higher in rigidity than a rod with a conventional I-type section produced by forging, and
has contributed to enabling engines to be increased in speed.
1. Introduction
As one of the main kinetic components enabling the
operation of high-speed and high-power Formula One
engines, connecting rods (conrods) are the subject of a
constant quest for weight reductions and increases in
strength and rigidity. For this reason, titanium alloys
displaying high specific strength were applied in their
manufacture. In 2000, weight savings were achieved
through the use of a -rich + titanium alloy, SP-700(1),
which possesses 25% higher fatigue strength than that
of the formerly used 6A14V titanium alloy. However,
responding to demands for further weight savings
exclusively by means of increasing strength was bringing
materials close to the limit of rigidity design, a situation
which necessitated a new technological breakthrough.
The potential for the use of a hollow conrod structure
as a means of achieving weight savings while
maintaining a geometrical rigidity was therefore studied.
phase, and therefore does not affect the base material by

heating. In addition, titanium displays a high oxygen
solubility limit, so oxide layers easily diffuse and
disappear on titanium surfaces. Diffusion bonding was
therefore focused on, and manufacturing methods for the
component were studied on this basis.
2.2. Mechanism of Diffusion Bonding
Diffusion bonding is a bonding method in which the
temperature of the materials to be bonded is maintained
at 0.7 Tm (Tm = melting point) or more in a vacuum
or reductive gas environment, and pressure is applied in
order to promote diffusion. Figure 1 shows a model of
the diffusion process(2). In the initial stage of the process,
the asperities on the surface to be bonded are deformed
and their close adherence promoted by increasing
pressure and heat. Next, diffusion causes the grain
boundaries at the interface between the materials to
migrate and vacancies to disappear. In the final stage of
the process, the remaining vacancies disappear through
2.1. Study of Method for Hollowing Conrod
A variety of potential methods of realizing a hollow
conrod structure were studied. One suggested method
was to form a hollow shaft extending from the big end
by means of electrochemical or mechanical machining,
which would then be cover-welded using electron beam
welding (EBW), thus forming a hollow structure.
However, this method was unable to resolve the issue
of the strength of the joints. Issues of reduced strength
also arose in the cases of casting and wax soldering.
Diffusion bonding, as employed in the manufacture
of aircraft turbines, involves the diffusion of a solid
251
Fig. 1
Model of diffusion bonding process
Development of Hollow Titanium Connecting Rod
Developed
(hollow)
Rolled plate
(t19)
Blanking
(water Jet)
T-joint piece
Diffusion bonded tensile test piece
1200
60
1000
50
800
40
600
30
400
20
200
10
Hot forging
0.2%YS-T
0.2%YS-L
UTS-T
UTS-L
EL-T
EL-L
RA-T
RA-L
0
DB-Rt 1.6
DB-Rt 6.3
0.2%YS : 0.2% yield strength

UTS : Ultimate tensile strength
EL : Elongation
RA : Reduction area
Microstructure of diffusion bonded area of

connecting rod
Cut
L-joint piece
EL / RA [%]
Fig. 4
25 m
Round bar
( 50)
10
45
Base (no-DB)
Conventional
(solid)
M6
17
10
Diffusion bonded area
Fig. 2
2.5
0
R1
2.3. Conrod Bonding Process

Because the conrod is solution-aged at a temperature
lower than the transformation temperature (870 C) in
order to obtain a predetermined level of strength, the
bonding temperature was set at 830 C, equivalent to the
solution treatment temperature. The maximum pressure
was set at 4.0 MPa, and the diffusion time kept for 5.0
hr. A hot press vacuum furnace owned by Kinzoku
Giken Co., Ltd., capable of independent load control in
16 axes, was employed in the diffusion bonding.
Figure 3 shows the process of manufacture of the
hollow conrod. A rolled sheet is roughly blanked using
a water jet, after which it is machined into a half blank,
forming a hollow shaft. These half blanks are
superimposed and diffusion bonded.
The use of dowels positioned at the big and small
ends controls relative displacement during bonding to
within 0.13 mm at the upper limit of standard deviation.
The amount of crushing in the direction of thickness was
set at 4% of the initial thickness of the material, based
on the height of the carbon stopper plates during hot
pressing.
The level of roughness and cleanliness of the bonding

surface affects the mechanical properties of the bonded
section. Tests were therefore conducted to determine the
effects of these factors using tensile test pieces bonded
by means of two joint types (Fig. 4). Figure 5 shows
tensile properties for different levels of bonded surface
roughness. The level of surface roughness had a
particular effect on the elongation and reduction area of
the T-joint, and was therefore set at Rt1.6 or below in
order to obtain tensile properties equivalent to those of
the base material.
0.2%YS / UTS [MPa]
volume diffusion, and bonding is completed. Figure 2

shows the bonded microstructure realized in the conrod
in this project. A continuous metallic microstructure with
no remaining asperities at the bonding interface has been
obtained.
Fig. 5
Heat treatment
(STA)
Half-blank machining
(dowel at S/B end)
Machining
Acid
cleaning
Effect of surface roughness on tensile

properties of diffusion bonded test piece
Surface treatment
Diffusion bonding
(hot pressing)
Finished product
Work
Stopper plate
Fig. 3
Developed process of manufacture of hollow connecting rod
252
3. Achieved Performance
A hollow structure in which diffusion bonding is
employed in the central section of the conrod thickness
has been developed, as shown in Fig. 6. This has
increased the modulus section of the shaft of the conrod
while enabling thickness to be minimized. Compared to
a conventional I-type section, an 8% reduction in weight,
2.5 times increase in the rigidity of the shaft, and 18%
increase in the rigidity of the circulation of the big end
have also been achieved. In addition, as a result of the
reduction in the load on the conrod bearings, the
potential for a 250 rpm increase in engine speed has
been demonstrated in durability tests in a real engine.
Diffusion
bonded area
Conventional
Fig. 6
Hollow shape
Comparison of conventional and hollow conrod
4. Conclusion
A method of manufacture of a hollow conrod using
diffusion bonding has been developed. The weight
savings achieved enabled engines to be increased in
speed and power, and the technology was introduced to
race engines in 2003.
References
(1) Ouchi, C., Minakawa, K., Takahashi, K., Ogawa, A.,
Ishikawa, M.: Development of -rich + Titanium
Alloy SP-700, NKK Technical Review, No. 65, p. 6167 (1992)
(2) Owczarski, W. A., Paulonis, D. F.: Application of
Diffusion Welding in the USA, Welding Journal, Vol.
60, No. 2, p. 22-33 (1981)
Author
Sadami MINATO
253
Ken MIZOGAWA
Masayuki TSUCHIYA*
Sakae TSUNASHIMA*
Hisashi HASHIMOTO*
ABSTRACT
Titanium aluminide intermetallic compound has a high specific stiffness close to 40 GPa/gcm3, the maximum
allowed under the regulations. The material also has high temperature strength and its use in the cylinder head valves
of Formula One engines has contributed to achieving higher engine speed. Vacuum annealing has been developed to
try to enhance fracture toughness, and a technique has been established to prevent products with micro-cracks from
being shipped if they occur during processing, by optimizing eddy current inspection technology to overcome quality
issues.
1. Introduction
Table 1
Titanium aluminide (below, TiAl) has a Youngs

modulus of 155 GPa, density of 4.05 g/cm3 and great
strength under high temperatures. Substituting with this
material in cylinder head valves (below, valves) that use
titanium has enabled the achievement of higher engine
speed by giving them more stiffness at lighter weight.
However, the fact that this material is an intermetallic
compound means that valve breakage often occurs
because of the low fracture toughness and the quality of
production (Fig. 1), increasing the need for high fracture
toughness and enhanced quality.
Valve blank is produced through a process of twice
extruding with the use of vacuum arc remelting (VAR)
material(1). This blank is produced by only one company
in Europe (2), and negotiations and development that
include quality assurance and process sharing up to the
completion of the valve have progressed in tandem with
the materials application in racing engines with the aim
of achieving higher engine speed.
Fig. 1
Chemical composition of TiAl (atomic%)

Al
Cr
Nb
Ta
Ti
2002
46.5
2.5
1.0
0.5
0.1
Bal
2003-2005
42.0
2.5
1.0
0.5
0.1
Bal
2. Development Technology
2.1. Annealing for high fracture toughness
Since 2003, materials regulations have prohibited
material with specific stiffness of greater than 40 GPa,
so Honda engineers have switched to a composition with
less Al, as shown in Table 1. Because of electrode
dripping during VAR, valves initially have a segregation
structure as shown in Fig. 2, and in 2003 valves were
sorted by quality based on observation of the structure
of all valves. Subsequently, melting rods for VAR were
shortened and a quality sorting technique was established
that used samples showing structural limits.
Broken valve
Fig. 2

254
Segregation structure of valve
On the other hand, X-ray diffraction phase

analysis of the material determined that the structure
consisted of 2-phase (Ti 3 Al) / -phase (TiAl) / phase (bcc-Ti). Because the -phase is a metal phase,
it contributes to fracture toughness, but the amount
is very small and the valve material is made with a
heated extrusion production method, so it is
considered to be a semi-stable phase. Moreover, the
amount of -phase did not stabilize because of
segregation and variance in extrusion conditions when
producing the material.
Annealing was used to stabilize the amount of phase and enhance material elongation. Vacuum
annealing was performed with a soaking heat treatment
at 1000C, in which the three phases of 2 + + exist,
and subsequently the amount of -phase was raised to
more than 20 atomic% by forced cooling with argon gas.
Mechanical characteristics evaluation results are shown
in Fig. 3. This elongation of more than 2% at room
temperatures and rotating bending fatigue strength of
more than 800 MPa at room temperatures, and ensures
both high strength and high fracture toughness in the
valve.
2.2. Micro-crack Scanning Non-destructive Quality
Assurance Technology
During the TiAl extrusion process, the material is
covered with a stainless sheath. For this reason,
during machining both the highly malleable sheath
and the TiAl with low fracture toughness have to be
planed, and there is a possibility of cracks resulting
from the vibration of machining. In addition, there is
residual stress in the material during the extrusion
process, so it was also conceivable that cracks could
occur during finishing. If there are cracks open to the
surface, they can be detected by a dye penetration test
such as fluorescence dye penetration, but in some
cases there are cracks with compression stress that are
not open to the surface, and these can slip through
without being detected by dye penetration testing. As
can be experienced with Co-Ni bolts, there are also
cases where high-strength materials manufactured by
2.3. Effects
In 2002, techniques to make inlet valves lighter
helped cut 6.0 g from the weight of earlier titanium
valves while raising engine speed by 600 rpm, and in
2003, valves became 2.4 g lighter and engine speed 240
rpm faster.
8.0
Ultimate tensile strength
Bending fatigue strength
Elongation
1400
1200
Stress (MPa)
extrusion processing experience cracks with

compression stress in the grinding process.
The technology of eddy current scanning, which
detects changes in eddy current where cracks exist,
was optimized, and non-destructive quality assurance
was conducted. Eddy current scanning is done twice:
after blank machining with the sheath and TiAl
removed; and after complete machining. This
enhanced detection precision. The blank machining
status, from which there is a certain amount of
material yet to be removed, allowed detection of
relatively large cracks and was effective. For eddy
current scanning on valves that have completed
machining, a new sensor probe was developed,
positioning was done by robot arm, the gap between
sensor and valve was optimized, and noise was
prevented by optimizing the signal processing filter.
This made it possible to detect micro-cracks 100 m
deep in actual valves, as well as detecting microcracks 10 m deep in test pieces.
Figure 4 shows the appearance of eddy current
scanning.
7.0
6.0
1000
5.0
800
4.0
600
3.0
400
2.0
200
1.0
Elongation (%)
1600
0.0
AS Extrusion
Fig. 3
Anneal 1000 C 2 hr
Mechanical properties of TiAl
Fig. 4
255
Method of eddy current inspection
3. Conclusion
Using vacuum annealing heat treatment and eddy
current scanning, Honda developed lightweight TiAl
valves with long-distance reliability. This technology was
applied to inlet and exhaust valves to help create engines
able to produce higher engine speed.
References
(1) Publication of unexamined patent application, 200024748 (2000), Japan
(2) http://www.mwracing.eu/
Author
Masayuki TSUCHIYA
256
Sakae TSUNASHIMA
Hisashi HASHIMOTO
Development of Titanium Hollow Valve and Study

of Sodium-potassium Valve
Taihei SANADA*
ABSTRACT
Valve specifications with the largest hollow area possible while still assuring the strength of welded areas were
established, achieving a weight reduction of 4.4 g per valve, and applied to races. In addition, a high-quality oxidation
resistant process was established for filling titanium valves with sodium-potassium, and technology that cools the high
temperature areas of the valve by 100C or more was developed. These cooling effects enabled expansion of the hollow
area, further reducing the weight by 4.8 g per intake valve and 4.6 g per exhaust valve. Engine 2-race event durability
tests were completed, but the intake valve experienced a drop in output due to the effects of sodium-potassium heat
exchange, and some issues with exhaust valve durability also remained.
1. Introduction
Formula One regulations prohibited the application of
intermetallics to engine parts from 2006. Therefore, the
titanium-aluminum (TiAl, density 4.05 g/cm3, Youngs
modulus 155 GPa) valves had to be changed to titanium
alloy valves (in case of intake valve, density 4.65 g/cm3,
Youngs modulus 114 GPa).
To reduce the weight and increase the stiffness of
titanium alloy (Ti) valves, and increase the engine
rotations, it was essential to make the valve head and
stem hollow. In addition, filling Ti valves with sodiumpotassium was investigated to further reduce the weight
by expanding the hollow space.
thickness of the non-welded area was determined by

valve fatigue tests to assure a parts strength equivalent
to that of a solid valve head. Figure 2 shows the test
results. The residual thickness requirement was set at
1.27 mm or more.
Hollow
Non-welded area
Plug
EBW
2. Development Contents
Fig. 1
2.1. Development of Hollow Valve Structure

The valve head was hollowed by drilling out the
valve head, lightly fitting and electron beam welding
(EBW) a plug, and performing stress-relief heat
treatment.
Figure 1 shows a photograph of the valve head
section. EBW was performed from the combustion side
of the valve head, and melted the periphery of the plug
to a sufficient depth. The heat-treated microstructure of
the welded area was destroyed and the fatigue strength
dropped to 30% of the non-welded area, so the residual
257
Section of hollow head of intake valve
1.2
Fatigue strength ratio
to solid valve head
Lightweight, high-stiffness Ti valves were developed

by creating a hollow valve structure and establishing a
technology for filling it with sodium-potassium.
EBW
1.0
0.8
0.6
0.4
1.27 mm
0.2
0.0
0.0
Fig. 2
0.5
1.0
1.5
Non-welded area thickness (mm)
2.0
Fatigue test results for welded valve head
Development of Titanium Hollow Valve and Study of Sodium-potassium Valve
The valve head was hollowed to the greatest extent

that satisfied the residual thickness requirement. Together
with hollowing the stem, this achieved a weight
reduction of 4.4 g compared to a solid valve.
2.2. Development of Titanium Sodium-potassium Valve
The sodium (Na) and sodium-potassium used as
coolants are active, so they readily oxidize in air. In
addition, titanium has a lower standard free energy of
oxide formation than that of Na and potassium (K)(1), so
when the inside of a Ti valve is filled with oxidized
sodium-potassium, the inner walls of the valve oxidize
and the strength drops. Therefore, a process for filling
titanium sodium-potassium valves with non-oxidized
sodium-potassium materials is necessary.
Figure 3 shows a photograph of the sodiumpotassium exhaust valve section. The valve head and
stem have a hollow structure, and the sodium-potassium
is sealed by an inner plug and a stem end plug.
Figure 4 shows the sodium-potassium filling facility
and filling operation. Before sodium-potassium filling,
vacuum gas exchange is performed in the glove box to
replace the air inside the hollow area with argon (Ar)
to prevent oxygen from entering the hollow area. The
oxygen density was set at 10 ppm or less. Na-78wt%K
(NaK), which is liquid at room temperature, was selected
as the sodium-potassium material, and the chances for
contact with oxygen were minimized by filling the inside
of the valve directly from the bottle via a filling nozzle.
The filling amount was set at 50% of the hollow volume
in consideration of fluidity inside the hollow area.
A press machine that fits the inner plug and stem end
plug was located inside the glove box, and sealed the
valve after NaK filling. Finally, the stem end was laser
welded and completely sealed in a normal atmosphere.
Inner plug
Stem end plug
Fig. 3
Inside of
box
Dew
point
meter
NaK
nozzle
Valve
blank
Glove box
Fig. 4
NaK filling facility and filling operation
Comparison of current and NaK valves

Intake valve
Temperature
(C)
Fatigue strength
(MPa)
Valve weight (g)
Exhaust valve
Current
NaK
Difference Current
NaK
Difference
718
544
-174
840
737
-103
227
615
+130%
126
211
+67%
35.1
30.3
-4.8
32.7
28.1
-4.6
3. Engine Test Results

3.1. Temperature Reducing Effects of Sodiumpotassium
To confirm the valve head temperature reducing
effects, a quenched heat-resistant steel chip was
embedded in the center of the valve head, and the valve
temperature was estimated from the change in hardness
after engine operation. Table 1 shows the results. The
intake valve temperature was reduced by 174C, which
increased the material fatigue strength by 130%, enabling
per-valve weight reduction of 4.8 g. The exhaust valve
temperature was reduced by 103C, which increased the
material fatigue strength by 67%, enabling per-valve
weight reduction of 4.6 g.
3.2. Engine Performance and Durability
Motoring head tests confirmed that the sodiumpotassium intake valve increased the rotation upper limit
by 500 rpm, but the output dropped by 5 kW in the dyno
test. As the volumetric efficiency dropped, it is
hypothesized that the valve head port side and stem
temperatures rose due to the heat exchange of the
sodium-potassium, which warmed the intake air and
resulted in a drop in the intake air filling ratio.
The sodium-potassium exhaust valves finished a 2race event as an engine durability test, but hairline
cracks were evident in the welded area of the stem
end. Since the submission deadline for engine spec
homologation that would freeze the specifications was
approaching and there was insufficient time to
implement crack countermeasures, the exhaust valve
was not applied.
4. Conclusion
Section of exhaust NaK valve
Vacuum
gauge
Oximeter
Table 1
(1) Valve-head welding technology was established that does

not lower the parts strength, and hollowing achieved a
weight reduction of 4.4 g compared to a Ti solid valve.
The developed hollow valve was applied in races from
the first race of 2006.
(2) A high-quality oxidation resistant process was completed
for titanium sodium-potassium valves. Cooling effects
of 100C or more in high-temperature areas of the valve
were confirmed, enabling weight reduction of 4.8 g per
intake valve, and 4.6 g per exhaust valve.
(3) Sodium-potassium filling was confirmed to provide
sufficient cooling effects. However, output dropped in
some cases, and one factor for this is thought to be a
drop in volumetric efficiency in the intake valve.
258
Reference
(1) Japan Society of Corrosion Engineering: Fusyoku
bousyoku handobukku, Maruzen, p. 32 (2000) (in
Japanese)
Author
Taihei SANADA
259
Development of DLC Coating on Camshaft and

Rocker Arm
Naoaki ITO*
Masaomi YONEHARA*
Kazushige YAKUBO*
ABSTRACT
Among the Formula One engine valve train parts, the camshaft and rocker arm are subject to severe sliding
conditions and constant demands for higher rotating speeds, so it was necessary to increase endurance reliability.
Therefore, diamond-like carbon (DLC) coating films that strengthen the surface were developed, the film compositions
and hardness balance were optimized, and the coating films were applied. As a result, friction was reduced by a total
of approximately 5 kW, and endurance reliability was achieved that enables continuous camshaft use in 4-Race events.
1. Introduction
The camshaft and rocker arm are Formula One
engine valve train parts that are subject to severe sliding
conditions. The sliding conditions in the boundary
lubrication area reach a contact pressure of 1 GPa or
more, and a PV (contact pressure x sliding velocity)
value of 20000 MPa m/s or more. Figure 1 shows the
position of the sliding conditions on a Stribeck curve.
To secure sufficient endurance reliability under these
conditions, it was necessary to cover the parts with a
tough coating film that does not experience scuffing,
even when the oil film breaks, and does not wear under
high speed and high contact pressure sliding conditions.
For this reason, DLC coatings have been used since
2002.
Engines are becoming increasingly high power and
high speed, a pneumatic valve return system (J-VLV
mechanism) using an orifice has been applied, and

Formula One regulations also prescribe the continuous
use of a single engine in 2-Race events, so further
increases in durability were required. This increased the
need to develop a DLC coating for the camshaft and
rocker arm.
2. DLC Specifications
2.1. Approach towards DLC Layer Configuration
The sliding surfaces of both the cam robe and the
rocker arm slipper were treated with DLC, and this was
confirmed to increase wear toughness and reduce friction
for both parts. Figure 2 shows the film configuration
concept. To realize uniform contact, the top layer forms
a low-hardness running-in layer. In addition, the DLC
layer that serves as the main sliding layer achieves a
balance between hardness and toughness, and the
Surface condition
Control of solid-solid contact by surface smoothing
Friction
Boundary lubrication area
Running-in top layer

Accommodation of surface roughness, swelling and
parts alignment
Sliding layer
Compatibility of scuffing toughness and
wear toughness
Mixed
lubrication
Bonding layer
Multi-layer structure to reduce stress inside coating
Retention of transformation ability
Substrate surface
Rationalization of surface roughness, hardness and
cleanliness
Elastohydrodynamic
lubrication
V/P
Fig. 1
Stribeck curve
Fig. 2

260
Concept of DLC and bonding layer
bonding layer, which has a hardness gradient, is formed

by the spattering method to increase the scuffing
toughness and adhesion strength of the DLC film.
2.2. DLC Treatment of Parts
For DLC to slide well against each other, the film
hardness balance between them must be optimized. For
example, DLC with the same hardness or that have
topcoats that are too hard will attack each other, which
is not desirable.
The combination of DLC specifications was
evaluated and confirmed by basic sliding tests performed
under high contact pressure conditions that simulated the
actual sliding conditions. The results showed that a
hardness balance of approximately 1.3 times in terms of
the main sliding layer hardness reduced friction and
provided high seizing durability. Figure 3 shows the
hardness profiles of IDV38S, made by ICS Corporation,
Running-in layer
and DF1, made by Kobe Steel, Ltd., which were selected

as the combination of DLC specifications.
The DLC treatment for each part was as follows. The
camshaft was treated with the high hardness IDV38S,
and the rocker arm slipper surface was treated with DF1,
which has a lower hardness than IDV38S. This was done
so based on the rule of thumb that when these parts
slide, the camshaft experiences greater damage. Figure
4 shows the parts appearance after DLC treatment and
an image of sliding. Figure 5 shows the DLC film
configurations.
In addition, coating film quality is important for
camshaft and rocker arm sliding, as DLC peeling trouble
occurs for films with quality issues. Therefore, quality
assurance methods were established for film uniformity,
adhesion strength, wear toughness and other items;
quality check samples were created that determine the
allowable number, size and location of pinholes; and the
quality was controlled.
3. Confirmation of Effects
Main layer
35
25
Approximately 1.3 times
20
15
10
-IDV38S (ICS)
-DF1 (KOBE)
5
0
50
150
250
350
450
550
650
750
850
950
An IDV38S-treated camshaft and DF1-treated rocker

arm were introduced in the second half of 2006, and
sliding reliability was assured under conditions of a
contact pressure of 1.37 GPa and a PV value of 22 000
MPa m/s. Constant development led to yearly increases
in the limit contact pressure, enabling high load and high
speed operation at a contact pressure 17% higher and a
PV value 32% higher than that of the combination of
DLC specifications for 2004 (Fig. 6).
Displacement into surface (nm)

24000
1500
Coating hardness of IDV38S and DF1
Contact pressure
PV value
1450
23000
1400
22000
1350
21000
1300
20000
+17%
+32%
19000
1250
PV (MPam/s)
Fig. 3
Contact pressure (MPa)
Hardness (GPa)
30
18000
1200
17000
1150
16000
1100
2004
Fig. 6
Fig. 4
2005
2006-
Change in contact pressure and PV value
Parts after DLC coating
4. Conclusion
Running-in layer
: Hv1000, 0.2-0.5 m
Running in layer
: Hv1000, 0.2-0.5 m
DLC: Hv2500-3000, 1.5 m
DLC: Hv2000-2500, 2 m
Bonding layer: Hv1900, 1 m
Substrate
Substrate
Fig. 5
IDV38S and DF1 layer configuration
As Formula One engines evolve to higher output and

higher speeds, steady progress was promoted to develop
camshaft and rocker arm DLC that can withstand
operation under harsh sliding conditions. As a result,
performance was enhanced by a total of approximately
5 kW. In addition, these cam parts have been used in
4-Race events since 2007 with the aim of reducing costs,
and the DLC coatings were able to assure sufficient
endurance reliability.
261
Development of DLC Coating on Camshaft and Rocker Arm
Acknowledgements
The authors would like to express their deep thanks
to ICS Corporation and Kobe Steel, Ltd., for their
cooperation with the production technology and quality
assurance associated with these parts.
Author
Naoaki ITO
262
Masaomi YONEHARA
Kazushige YAKUBO
Development of Laser Clad Welded Valve Seat
Nobuyuki IMAI*
Hiroaki NISHIDA**
ABSTRACT
Laser clad welded valve seats can expand the flexibility of layout compared to conventional press fit valve seats.
The slot form, the shape and amount of the powder to be supplied, and the laser output parameters were optimized
and balanced, quality assurance requirements were established, and the valve diameter was expanded by 1 mm. This
achieved an increase in engine power of 6 kW at high water temperatures. In addition, the continuous use of one
engine over the distance for two Formula One race events was supported by enhancing the powder materials.
1. Introduction
The valve seat (hereafter, seat) pressed fit require
sufficient wall thickness around the seat press fitting area
of the cylinder head (hereafter, head), and also
stiffness of annular rigidity of the seat itself. For this
reason, the flexibility of the combustion chamber layout,
including the valve diameter, valve pitch, and port angle,
was limited. In addition, the area between the seat and
the head is crimped, so heat transfer is insufficient
between the valve, seat and head. This development

realized a large diameter valve layout by establishing
laser clad welded seat technology that forms the seat
directly on the head, and also aimed to reduce weight
by reducing the head wall thickness and increase engine
power by enhancing cooling performance.
Figure 1 shows the press fit seat port section, and
Fig. 2 shows the laser clad welded seat port section.
Laser clad welding enabled a thinner water passage wall
thickness.
2. Establishment of Laser Clad Welded

Seat Technology and Quality Assurance
The form of the slot used to clad the powder, the
diameter of the powder to be supplied, the manufacturing
conditions for welding the powder, and the quality
requirements were set.
Fig. 1
2.1. Laser Clad Slot Form

Figure 3 shows the laser clad slot form. The powder
is collected in the center in a symmetrical form. When
the clad thickness differs at the center and both edges
of the slot, the welding time also differs. This may result
in non-welds at the bottom of the slot, so the slot bottom
form was an arc and the slot walls formed an obtuse
angle. In addition, excess powder may likewise produce
a difference in thickness, so the slot height was set at 4
mm to prevent unnecessary powder from collecting.
Layout of press fit valve seat
Thin
wall
Thin
wall
Fig. 2
Layout of laser clad welded valve seat
2.2. Powder Supplied for Laser Clad Welding

A copper alloy in the form of a spherical powder
with a diameter of 80 5 m was used as the clad
powder. The powder was supplied using a vibration
powder feeder, and supply was controlled at an accuracy

** Aircraft Engine R&D Center
263
Development of Laser Clad Welded Valve Seat
of 0.005 g/s. This enabled the manufacture of

homogeneous laser clad welded seats.
To support the continuous use of one engine over the
distance for two Formula One race events, seat wear
resistance was enhanced by mixing stellite and tribaloy
powder into the copper alloy powder. Figure 4 shows
the transition in the development specifications. Efforts
were made to simultaneously achieve wear resistance and
fracture toughness in accordance with changes in the
valve train system specifications.
2.3. Laser Clad Welding Conditions
The welding process is an important process for the
laser clad, and is controlled by the laser output and the
amount of powder. In particular, the start point overlaps
with end point, so start area of laser clad form must be
inclined. Therefore, cladding start point was performed
by gradually increasing the amount of powder according
After work
Combustion chamber
Valve seat
to the form, and increasing the laser output power in

steps. Figure 5 shows this image drawing.
2.4. Quality Assurance of Laser Clad Welded Seats
Low quality in the start area may result in seat
defects and lead to engine blowout or reduced engine
power.
Non-welds can result from an excess powder amount
or insufficient laser output, and brittleness can result
from an insufficient powder amount or excessive laser
output. Figure 6 shows examples of these cases. Nonweld and embrittlement layer limit standards of quality
were prepared by analyzing the materials of seats in
which defects actually occurred.
Limit standards were judged by analyzing the cutting
planes of lot samples. The non-weld limit standard was
no pinholes of diameter 0.3 mm or more, and the
embrittlement layer limit standard was no embrittlement
0.4 mm or thicker as obtained by measuring the hardness
of the start area at the two points 1 mm and 4 mm from
the start point. Figure 7 shows the limit standard of
quality.
Symmetry
Clad area
Work
Line
Obtuse angle
Non-weld
Length regularity
Port
Fig. 6
Cylinder head
Fig. 3
Laser clad welding slot form
Dilution brittleness
Limit quality examples
Clad (Cu alloy)

Start point 1 mm
Cylinder head (Al alloy)
Impact stress
Good
Cu alloy +
stellite 5%
Hardness (Hv 100 g)
800
Cu alloy
Cu alloy +
tribaloy 5%
Cu alloy+
stellite 7%
4 mm
700
600
500
Critical zone = Hv500
400
300
200
Distance from start point

1 mm (good)
4 mm (good)
Bad sample
100
0
-0.4
Abrasion resistance
Fig. 4
Fig. 7
End area of laser clad
Cylinder head
Fig. 5
Start and end form of laser clad
0.2
0.4
0.6
0.8
Seat side
Distance from interface (mm)
Valve seat development corresponding to

higher loads
Start area of laser clad
-0.2
Head side
Quality standard for dilution brittleness
3. Results
Laser clad welded seat technology increased the
flexibility of the layout around the valve seat, enabled
expansion of the valve diameter by 1 mm, and achieved
a 6 kW increase in the engine power at a high water
temperature of 120C.
264
4. Conclusion
Laser clad welded valve seat manufacturing
technology was established and applied to Formula One
races from 2004. Hard particle powder was added and
quality standards were established, which helped achieve
the durability and reliability required to support the
continuous use of one engine over the distance for two
Formula One race events.
Acknowledgements
The authors wish to express their deep thanks to
Tosei Electrobeam Co., Ltd., for cooperation in
establishing the powder supply conditions.
Author
Nobuyuki IMAI
265
Hiroaki NISHIDA
Takeshi MUNEMURA*
Hiroshi YAMADA*
Takayuki OHNUMA**
ABSTRACT
This paper discusses the use of titanium exhaust pipes for Formula One race cars in weight reduction. Newly
developed processes for titanium material including heat treatment providing an acicular structure, oxidation
resistant coating, sandwich press forming, and investment casting manufacturing have achieved a 20% weight
reduction. This specification was not applied to race cars but provides technology for lightweight titanium exhaust
pipes.
1. Background and Objective
2. Technology Developed
Exhaust pipes are mounted on the upper part of

Formula One race cars, and their component weight
is high. Therefore, weight reduction has a large effect
on enhancing chassis performance. Normally Inconel
625 (specific weight of 8.5 kg/cm 3), which has good
high-temperature strength and workability, is used for
exhaust pipes. In this research, titanium alloys, which
have low specific weight (specific weight of 4.5 kg/
cm 3 ) , were selected as substitute material, in
association with a variety of development efforts such
as controlling the micro structure, enhancing
oxidation resistance, and establishing casting and
press forming method to reduce the weight of exhaust
pipes.
2.1. High Temperature Strength

Figure 1 shows the structure of an exhaust system.
The exhaust pipe is made up of a primary pipe, collector,
and tail pipe. The materials used are KS Ti-9 plate
material manufactured by Kobe Steel, Ltd. and
investment casting Ti-64.
The strength of titanium alloys is lower in high
temperature environments exceeding 600C. In order to
enhance high temperature strength and creep resistance,
heat treatment providing an acicular structure was carried
out at 1000C on the final process and structural control
was performed. As shown in Fig. 2, specific fatigue
strength is higher than Inconel 625 at temperatures lower
than 800C and so material replacement was feasible.
Front
Rear
Collector
50

(MPa/g/cm3)
Tail pipe
High temperature press forming

(Ti-9 plate t = 0.7)
Investment casting
(Ti64 t = 1.0)
Primary pipe
Ti-9 acicular
30
20
10
<Temp. range of actual parts>

Tail and primary Collector
0
550
High temperature press forming

(Ti-9 plate t = 1.0)
Fig. 1
Inco625
40
Structure of exhaust pipe
600
Fig. 2

** Honda Engineering Co.,Ltd.
266
650
700
750
800
Temperature (C)
850
900
950
2.2. Oxidation Resistance

The oxidation of titanium alloys accelerates in high
temperature environments so an oxidation-resistant
coating became necessary. An internal dispersion process
of aluminum was developed, enabling a titanium
aluminum layer to be formed on the surface, and
providing good oxidation resistance characteristics for
atmospheric exposure at 950C. Figure 3 shows
processes for applying both a stable thick oxidationresistant coating over the entire surface, including the
inside of the pipes, and heat treatment for material
structure control. A 30 m stable titanium aluminum
oxidation-resistant coating as shown in Fig. 4 was
achieved.
2.3. Workability
In order to reduce thickness and weight and enhance
design freedom in the shape, press formed parts and
investment casting parts were welded together.
2.3.1. Press forming
As the elongation of titanium alloys at room
temperatures is small, hot press forming of 600C or
higher was necessary, under which the material
elongation is equivalent to that of Inconel material.
However, the heat capacity of the thin plate material is
low and the temperature dropped quickly after it has
been taken off from the furnace, so cracking could not
Welding
Chemical
milling
Annealing
Heat treatment
for diffusion
Fig. 3
be avoided. Press forming was enabled through

sandwiching using stainless steel plates on both sides to
hold temperature constant (Fig. 5). In order to achieve
both holding temperature and press dimension accuracy,
a thickness of 1.0 mm was used for the stainless steel
plates. These stainless steel plates prevent direct contact
between the die and the titanium alloy plates and so they
also help to control galling.
2.3.2. Investment casting
There had been no major accomplishment in thin wall
casting of a complex shaped collector and lost wax
casting is unable to meet the short term delivery of shape
modifications required for racing. The requirements for
casting were met using a Rapid Prototyping (RP) model,
and by applying the LeviCast method (suction method
by negative pressure). A thickness reduction by chemical
milling was also utilized, and together these methods
made it possible to reach the target plate thickness of
1.0 mm and a 30-day delivery time. Figure 6 shows a
summary of the process and Fig. 7 shows the external
appearance of the parts.
Master RP model production
(polystyrene)
3D data
Removing RP model
burning shell
Fig. 6
Aluminum
plating
Casting
Molding
HIP
Chemical
milling
Process of casting
Heat treatment
providing acicular
structure
Process of coating and heat treatment

Fig. 7
Casting collector and primary transitional pipe
3. Results
TiAl anti-oxidation layer
This development achieved a weight reduction of

1400 g (20%) per exhaust pipe assembly (Fig. 8).
Ti base metal
20 m
Fig. 4
Cross section of anti-oxidation coating
SUS t1.0
Ti-9 t1.0
SUS t1.0
Die (FCD)
Fig. 5
Die (FCD)
Rear
Front
SUS t1.0
Ti-9 t1.0
SUS t1.0
Die (FCD)
Process of press forming
Fig. 8
267
Full titanium exhaust pipe assembly
4. Conclusion
The exhaust pipe for the 2007 specification had a
unique layout: the exhaust pipe extending toward the
engine in the forward direction, then loops back toward
the rear. The loop-back area was broken during a track
test and so this was not used for the race. However, in
the 2006 specification (normal type; straight toward the
rear of vehicle), the exhaust system based on this
technology was used for 1200 km, confirming the
durability of the technology.
This has shown that weight reduction is feasible by
converting the exhaust pipe to titanium and optimizing
the component shape.
Author
Takeshi MUNEMURA
268
Hiroshi YAMADA
Takayuki OHNUMA
Development of Lightweight and Low-friction Resin

Materials for Oil Pump Rotor
Yu MURAI*
Masayuki TSUCHIYA*
ABSTRACT
The total weight of the scavenge and feed oil pump rotors fitted in Formula One V10 engines was previously
more than 1 kg, leading to demands for weight savings and friction reductions. A polyethyl ethyl ketone material and
a polyimide material have been used to manufacture the scavenge pump outer rotor in a two-stage injection molding
process, and a polyethyl ethyl ketone extrusion material has been used to manufacture the feed pump outer rotor.
This has resulted in a weight saving of 465 g and a friction reduction of 1.8 kW.
2. Product Development
2.2. Development of Manufacturing Methods

Injection molding not only enables the achievement
of high strength, but also obviates the need for the
cutting process, thus reducing costs. However, because
the outer rotor is thick, it was a challenge to achieve
dimensional accuracy using normal injection molding,
given the large surface sink. A two-stage process was
therefore applied in which the inner part of the outer
rotor is molded first, after which the outer part of the
outer rotor is molded by enveloped casting. The twostage injection molding method enables different
2.1. Material Evaluation

The materials used to manufacture pump rotors are
required to display mechanical strength, dimensional
stability, wear resistance, and sliding performance at oil
temperatures of around 120 C. Candidates for the resin
material were selected from among high-strength and
high temperature resistance super engineering
thermoplastic resins and thermosetting resins. These were
a polyethyl ethyl ketone (PEEK) injection material
reinforced with 30% carbon fiber (CF), a PEEK
extrusion material reinforced with 30% CF, a
thermoplastic polyimide (PI) injection material reinforced
with 30% CF, and a phenol resin injection material
reinforced with 40% glass fiber (GF). High-temperature
tensile tests and high-temperature tensile fatigue tests
were conducted on test pieces formed into ASTM#1
dumbbells. The results of the fatigue tests are shown in
269
160
= 0 B50 (MPa)
Because dry sump lubrication is employed in Formula

One engines, two oil pumps are used: A feed pump to
increase pressure, and a scavenge (scav) pump to
scavenge oil. Both pumps are trochoid pumps. One set
of rotors for the feed pump and seven sets of rotors for
the scav pump are fitted in each V10 engine at the
commencement of development, the total weight of
which was previously more than 1 kg. The oil pump
rotors have conventionally been manufactured from steel
and aluminum (Al) sintered metal. The project discussed
in this paper developed resin rotor materials in order to
achieve weight savings and reduce friction.
Fig. 1. The materials selected were the PEEK injection

material reinforced with 30% CF and the PI injection
material reinforced with 30% CF, which displayed high
strength in the temperature range in which the rotors
would be used.
Fatigue strength
1. Introduction
PEEK+CF30%
injection
Working temperature
140
120
PI +CF30%
injection
100
80
Phenol+GF40%
injection
60
40
PEEK+CF30%
extrusion
20
0
0
Fig. 1
50
100
150
Temperature (C)
Fatigue strength of resin materials
200
Development of Lightweight and Low-friction Resin Materials for Oil Pump Rotor
Inner: PEEK CF30%

(wear resistance)
Teeth surfaces
Molding
Outer: PI CF30%
(dimensional stability)
1.2
Friction torque (Nm)
materials to be used for the inner and outer parts. The

PEEK material with its good wear resistance was
employed for the inner part and the PI material with its
good dimensional stability was employed for the outer
part, enabling the issues of the surface sink and wear
resistance to be resolved simultaneously. In addition, the
method has enabled the accuracy of the teeth on the
outer rotor to be controlled to within 20 m, and the
teeth machining process can therefore be omitted (Fig.
2). The inner rotor is manufactured using a WE54
magnesium alloy extrusion material in order to achieve
weight savings. A weight saving of 420 g (53%) was
achieved against Al sintered materials in the case of the
scav rotor.
The coefficient of thermal expansion of the resin
materials varies with their method of manufacture, and
they display strong anisotropy. For the injection
materials, the coefficient of thermal expansion was
measured in two directions to set the dimensions. For the
extrusion materials displaying a high level of variation
between lots, the coefficient of thermal expansion of
different lots was measured in two directions in order
to set the rotor thickness and the diameter dimensions.
1.0
Conventional Al-PM rotor
0.8
0.4
Developed rotor
0.2
Outer: PEEK-PI
Inner: WE54
0
3000
Fig. 3
4000
5000
6000
7000
Pump speed (rpm)
8000
9000
Results of friction torque measurements
4. Conclusion
Resin oil pump rotors were developed, with the
following outcomes.
(1) Scav pump rotor
The outer rotor is manufactured from PEEK and PI in a
two-stage injection process, and the inner rotor is
manufactured from WE54 magnesium alloy, resulting
in a weight saving of 420 g and a friction reduction of
1.8 kW. The new rotor has been employed in races since
the first race of 2004.
(2) Feed pump rotor
The outer rotor is manufactured from a PEEK extrusion
material and the inner rotor is manufactured from the
conventional Al sintered metal, resulting in a weight
saving of 45 g. The new rotor has been employed in races
since the 16th race of 2004.
Outer and side surfaces

Machining
Fig. 2
10% DOWN
0.6
Acknowledgments
The authors wish to take this opportunity to thank all
the members of staff at Nichiei Co., Ltd., who made
helpful proposals concerning the molding of the parts
discussed in this paper, and who cooperated in the
development of the manufacturing method for the parts.
Developed outer rotor manufactured using

two-stage injection molding
3. Performance and Effects

An oil pump rig test was performed at an oil
temperature of 100 C using an injection molded resin
scav rotor. As Fig. 3 shows, the results of the test
indicated a 10% decline in torque. An engine test
showed a decline in friction of 1.8 kW. No mechanical
damage or significant performance declines were
observed following engine durability tests.
There were concerns that the use of resin for the feed
rotor would increase rotor wear due to increased
pressure. The wear resistance of a variety of materials
was tested in oil pump rig tests. The lowest amount of
wear was displayed by a rotor with an outer rotor
manufactured using the PEEK extrusion material, and an
inner rotor manufactured using an Al sintered material.
In addition, no significant performance declines were
observed following engine durability tests using a rotor
manufactured from these materials. A weight saving of
45 g (52%) was achieved in the feed pump.
270
Author
Yu MURAI
Masayuki TSUCHIYA
Research on Coolant Using Nano Additives and

Heavy Water
Yu MURAI*
ABSTRACT
Research was done on coolants with the objective of increasing automobile radiator cooling efficiency and
contributing to aerodynamics by enhancing the thermophysical properties of engine coolants.
The research confirmed the effectiveness of adding nano additives to enhance heat conductance, as well as using
heavy water in solvent to increase specific heat of the coolant. Because evaluations of cooling performance in actual
engines have not shown a clear difference from conventional coolants using water, a precise heat transfer measurement
technology was established to clarify the factors affecting heat transfer, and from this, directions were established for
the development of engine coolants.
1. Introduction
Heat transport capacity =

Specific heat per unit volume
Formula One engine coolants have conventionally

used water. Coolants circulate between the engine and
radiator to radiate heat. Increasing the radiators cooling
efficiency makes it possible to reduce the surface area
of the radiator duct aperture and lower aerodynamic
resistance. Additionally, by making the radiator itself
smaller, it is possible to design in ways that reduce
weight and enhance aerodynamic characteristics, factors
that could shorten lap times.
The target of the research was to enhance cooling
efficiency by the equivalent of one rank of duct aperture
surface area (equivalent to reducing coolant temperature
by 4C) by enhancing the coolants thermophysical
properties. This paper introduces research that, based on
a patent(1) concerning heat transport fluids from research
on mass-produced automobiles, sought to raise the
coolants heat transfer coefficient by producing
experimental coolant whose thermophysical properties
have been enhanced by adding nano additives at a high
concentration and using heavy water.
2. Materials Development
The heat transport capacity of coolants is generally
calculated from the specific heat per unit volume and
from the heat transfer coefficient, as shown in Eq. (1).
From Eq. (3), which is found by substituting Eq. (2) into
Eq. (1), it is supposed that raising thermal conductivity,
raising specific heat and lowering viscosity will cause
heat transport capacity to increase.
Heat transfer coefficient
(1)
(2)
Substituting (2) into (1),
Heat transport capacity =
[Cp ] [ 1/6 (Cp )1/3 U1/2 2/3]
(3)
Pr: Prandtl number; Cp: specific heat; : density;

: dynamic viscosity
U: flow velocity; : thermal conductivity
The properties of the coolant experimentally
produced are shown in Table 1. Based on a heat
transport fluid patent, the aim was to increase the
coolants thermal conductivity by adding cup-stacked
carbon nanotubes (below, CS-CNT). Coolant DW178
was test-produced in which dispersant was used in water
to disseminate CS-CNT at a rate of 10 wt%, and the
coolant was tested in a wind simulator(2), but the radiator
became clogged and prevented evaluation. The cause is
believed to be that the heat tolerance of the dispersant
was insufficient in the coolant temperature zone of
100C and up in which coolants are used in Formula
One racing.
The aim in this research was to increase thermal
conductivity by not using a dispersant whose heat
tolerance was in question, and to increase specific heat

271
Research on Coolant Using Nano Additives and Heavy Water
Material
Additive
Thermal properties of coolant
Coefficient of heat transfer (W/m2K)
Table 1
Thermal Specific Kinetic

Density
viscosity
Dispersant conductivity heat
W/(mK)
J/(cm3K)
mm2/sec
g/cm3
H2O (Base)
0.61
4.2
0.7
1.00
DW178
CNT 10%
ROSO3-Na
0.74
4.0
1.4
1.06
H2O+Al2O3
Al2O3 30%
0.73
4.0
1.5
1.27
D 2O
0.60
4.6
0.9
1.11
D2O+Al2O3
Al2O3 30%
0.79
(+30% )
4.5
2.5
(+7.7%) (+270%)
1.41
16000
Conduction
14000
12000
10000
H 2O
H2O+Al2O3
D 2O
D2O+Al2O3
Convection
H 2O
8000
6000
4000
D2O+Al2O3
2000
0
0
0.5
1.0
1.5
2.0
2.5
3.0
Flow velocity (m/sec)
by using heavy water. Al 2O 3 nano additive (below,

Al2O3), formulated by a sputter rapid cooling technique,
was used as an additive since it offers better dispersion
in water as compared to CS-CNT. To increase specific
heat, heavy water (below, D2O) was used that has 10%
higher volume specific heat than ordinary water (below,
H 2 O). Adding Al 2 O 3 to D 2 O at a rate of 30 wt%
increased thermal conductivity by 30% and specific heat
by 7.7%. Viscosity, however, also increased by 270%.
Three formulations were tested in the wind simulator:
Al 2 O 3-added H 2O, Al 2O 3 -added D 2O, and D 2 O. The
cylinder head temperature fell by 2 - 5C, but at the
same time engine output dropped by 4 - 8 kW, there was
no correlation between test results, and reproducibility
was low. Therefore, a technique was devised that
evaluates the coolant, including the impact of the flow
of each coolant, by separating heat conduction within
solids and the heat conduction on the fluid/solid
boundary.
3. Stand-alone evaluation results

A liquid-circulating heat exchange testing system was
built as shown in Fig. 1. Figure 2 shows results of
measuring heat transfer coefficient dependence on flow
velocity. The development specifications suggest that
cooling performance will be enhanced as compared to
water in the flow velocity zone of 0.5 m/sec and below,
but above 0.5 m/sec, cooling performance will decline.
It is predicted that since the flow velocity in the cylinder
heads water channels is 2 - 6 m/sec, and that inside the
radiator tubes is 1 - 2 m/sec, cooling efficiency will
decline if this is used in Formula One coolants.
Based on the above, at low flow velocities, the
impact of the liquids thermal conductivity will be great,
but at high flow velocities, the impact of viscosity will
Fig. 2
Relationship between heat transfer and flow

rate
be great. In other words, at low flow velocities, heat

transfer within and among substances is dominant, but
at high flow velocities, heat transfer from movement of
the substances themselves is dominant. This made it
clear that an automotive engine coolant must be
formulated to strike a good balance between increasing
thermal conductivity and reducing viscosity.
4. Conclusion
The newly developed nano fluid increases thermal
conductivity, but at the same time causes an increase in
viscosity because of the effect of interactions between
additive particles, indicating that cooling efficiency
would decline at the flow velocities at which Formula
One engine coolants are used. Therefore, engine coolants
need to be formulated to strike a good balance between
increasing thermal conductivity and controlling increases
in viscosity.
References
(1) JP, P2008-201834, A (2008)
R&D Technical Review, F1 Special (The Third Era
Hot coolant; 80 C
Thermo
couple
Stainless steel
Thermo
couple
Author
Channel 9x50x1-3 mm
Aluminum or Stainless
2-10 mm
Channel 9x50x1-3 mm
Thermo
couple
Stainless steel
Thermo
couple
Cold coolant; 7 C
Fig. 1
Overview of heat transfer coefficient

measurement apparatus
Yu MURAI
272
Daiki KONAGAYA*
Tsutomu TANAHASHI*
Takashi TANAKA*
ABSTRACT
Two types of steel, a high pitting-toughness material displaying excellent performance in relation to the gear tooth
fatigue characteristic and a high yield strength material displaying excellent tooth bending toughness, were developed
for use in Formula One gearboxes. Optimization of the alloy composition increased pitting toughness, contributing to
the achievement of a 740 g weight saving by enabling the thickness of the gear teeth to be reduced. The high yield
strength gear material displayed a yield strength in excess of 1900 MPa, and was developed to ameliorate the issue
of tooth root fatigue damage originating in bending of the gear teeth under the excessive input forces characteristic of
Formula One racing. From 2008, regulations were changed to stipulate the use of a single gearbox for four race events,
and the new material contributed to the achievement of increased long-distance reliability.
1. Introduction
Table 1
Chemical composition
(mass%)
At the commencement of the development program

for Hondas Third Formula One Era, gears manufactured
by Xtrac from a material corresponding to SNCM815VA
were employed in the BAR Honda Formula One
gearbox. However, in 2001, there was a frequent
occurrence of tooth pitting on the final drive gear. The
need for a high pitting-toughness gear material therefore
increased, and attention was focused on the development
of a new material.
Bending of the gear teeth due to extremely high input
forces, as for example at the start of a race, also
represented an issue. Changes in Formula One
regulations in 2008 stipulated the use of a single gearbox
for four race events, and a development program for a
high yield strength gear material was therefore conducted
in order to increase tooth bending toughness. The target
value for tooth bending toughness was set 20% higher
than the target for high pitting-toughness material, and
the target value for 0.2% yield strength was set at 1900
MPa and above (40% higher than the target for high
pitting-toughness material).
2. Development of High Pitting-toughness

Material (LBHD-2E)
2.1. Developed Technology
In order to increase the pitting toughness of gear
teeth, it is important that the hardness of the tooth
surface should not decline (i.e. the tooth surface should
not soften) even at high temperatures(1). Table 1 shows
C
Conventional steel
SNCM815VA
Developed steel
LBHD-2E
Si
Mn
Ni
Cr
Mo
0.12- 0.15- 0.30- 4.00- 0.70- 0.150.18 0.35 0.60 4.50 1.00 0.30
0.30
1.50
0.35
2.00
1.50
0.75
the chemical composition of the developed steel, in

which the amount of Si, Cr, and Mo, elements that
increase resistance to temper softening, was optimized.
A triple melt process was used to manufacture the steel
in order to reduce the incidence of inclusions, which can
be the origin points for tooth pitting, enabling production
of a high-cleanliness material.
2.2. Effects on Performance
Figure 1 shows the surface hardness profile of the
conventional and developed steels after heating to 300C
following carburization. The decline in surface hardness
is controlled in the developed steel, LBHD-2E, in
comparison to the conventional steel, SNCM815VA, and
it displays a higher resistance to softening at high
temperatures.
Figure 2 shows the results of a gear tooth pitting test
using race gears. LBHD-2E displays more than twice the
pitting toughness of SNCM815VA in terms of life
cycles. LBHD-2E was employed in races from 2004
onwards, and enabled the achievement of a weight
saving of 740 g through the reduction of the thickness
of the gear teeth.

273

Table 2
800
Chemical composition
Hardness (HmV)
Conventional steel
(mass%)
Developed steel
700
LBHD-2E
Conventional steel
LBHD-2E
600
Developed steel
250STF5
500
400
0.0
Si
Mn
Ni
Cr
Mo
0.30
1.50
0.35
2.00
1.50
0.75
0.52
2.50
0.70
1.80
0.80
0.40
SNCM815VA
Table 3
0.5
1.0
1.5
Comparison of mechanical properties
2.0
Distance from surface (mm)
3600
3400
3200
3000
2800
2600
2400
SNCM815VA
2200
2000
1800
1600
1400
1200
1000
1.E+04
1.E+05
EL
%
Core
hardness
42 HRC
SNCM815VA
1029
1308
18.6
LBHD-2E
1332
1795
15.9
52 HRC
Developed steel
250STF5
2137
2493
6.5
59 HRC
14
LBHD-2E
12
LBHD-2E
250STF5
10
8
6
4
+19%
1.E+06
1.E+07
0
0.75 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15
Impact load (ton)
1.E+08
Number of failures
Fig. 2
TS
MPa
Conventional
steel
Comparison of hardness profile of

conventional and developed steel after
300 C heating test
Deformation ( m)
Contact stress [MPa]
Fig. 1
0.2% YS
MPa
Fig. 3
SN diagram of gear tooth pitting
3. Development of High Yield Strength Gear

Material (250STF5)
3.1. Developed Material
The development of a high yield strength material
was conducted with high-Si steel(2) as the base material,
given that it enabled the achievement of a balance
between high yield strength and high toughness, and was
a material for which carburization could be used to
harden the surface layer. The amount of C added was
increased to achieve high hardness in the matrix, and the
amount of Si added was increased to increase resistance
to softening. Because Si also promotes grain oxidation,
low pressure carburizing was selected as the carburizing
method. Direct quenching was employed during the
carburizing treatment and temperature conditions were
optimized in order to prevent grain growth. Table 2
shows the chemical composition of 250STF5, the
developed steel, and Table 3 shows its mechanical
properties. The material satisfied the target of 1900 MPa
or higher for 0.2% yield strength, reaching a figure of
2137 MPa.
3.2. Effects on Performance
F i g u r e 3 shows impact loads and bending
deformation in tooth bending tests conducted on a pair
of gears. Compared to the previous steel, LBHD-2E, the
Result of tooth bending test
load at which the teeth commenced bending was 19%

higher for the developed steel.
Tooth bending tests in an actual gearbox also showed
that the load at which the teeth commenced bending was
24.8% higher in the developed steel as a percentage of
input torque. 250STF5 was employed in races from
2008.
4. Conclusion
A high pitting-toughness material was developed, and
was employed for ratio gears in races from 2004. This
helped to enable the thickness of the gear teeth to be
reduced by 20%, contributing to the realization of a
weight saving of 740 g in the gearbox as a whole. The
material was later also employed for the final gears and
the dog rings, contributing to the achievement of
increased reliability against gear tooth pitting.
A high yield strength gear material was developed
that exceeded development targets, achieving a yield
strength of 2137 MPa, 60% greater than that of the
previous steel. When the developed steel was employed
in a gearbox, the load at which the teeth commenced
bending was increased by 24.8%. This material was used
for the 1st gears and shaft from the opening race of 2008,
and contributed to the achievement of increased
reliability over long distances, as necessitated by the
274
regulation stipulating the use of a single gearbox for

four-race events.
Acknowledgments
The authors wish to take this opportunity to thank the
staff of Daido Steel Co., Ltd., for their generous
assistance in materials development and gear tooth
bending tests during the course of the projects discussed
in this paper.
References
(1) Hanyuda, T., Nakamura, S.: Properties of Newlydeveloped Pitting Resistance Steel for Gears,DPG1,
Electric Furnace Steel, Vol. 71, No. 1, p. 59-64 (2000)
(2) Kurebayashi, Y., Hatano, A.: Influence of Silicon and
Vanadium Contents on Fatigue Properties of Carburizing
Steels, Electric Furnace Steel, Vol. 67, No. 1, p. 34-43
(1996)
Author
Daiki KONAGAYA
275
Tsutomu TANAHASHI
Takashi TANAKA
Development of DLC for Transmission Gears
Tsutomu TANAHASHI*
Yasunori ONAHA*
ABSTRACT
A diamond-like carbon (DLC) coating has been developed for the transmission gears to increase their transmission
efficiency. Using the carbon sputter method increased topcoat hardness and suppressed the formation of interfaces
within the coating film, which enhanced durability under high surface pressures. This DLC was applied to the shift
gear, final gear and bevel gear, and contributed to reducing friction. In addition, it secured the reliability needed to
enable continuous use in the 4-Race events that Formula One regulations prescribe.
1. Introduction
Transmission loss is mainly due to losses from the

bearing and gear, and gear sliding loss accounts for a
large percentage of this. Therefore, DLC coatings that
have good friction characteristics are thought to
effectively increase transmission efficiency. However,
the transmission gears (hereafter, gears) have a high
surface pressure load, so the DLC used by engine valve
train parts lacked sufficient durability and could not be
applied.
Therefore, to increase transmission efficiency by
utilizing the DLCs good friction characteristics, the
following targets were newly set, and a DLC coating for
gears was developed.
(1) A coating film composition that can withstand surface
pressure loads up to a maximum 2.2 GPa, enabling
continuous use in the 4-Race events that Formula One
regulations have prescribed since 2008
(2) Establishment of a coating method that forms a uniform
film over the complex gear tooth shape
To increase adhesive strength between the coating

film and the substrate, which is an issue under high
surface pressures, the substrate surface roughness was
finished to approximately 0.1 Ra, and surface
contamination (oil components and minute corrosion)
was eliminated by enhanced cleaning. As a result, this
has resolved issues rooted in the substrate surface
condition.
The nano-indenter hardness of the topcoat was
increased to 60 GPa to enhance the wear resistance.
From there down to the substrate, the coating film
consists of an interlayer metal-carbon coating film
(hereafter WC-C) with a hardness that changes
gradually until the interlayer joins with the chrome (Cr)
adhesion layer. Figure 1 shows the DLC coating film
composition.
The carbon sputter method was used continuously
from the DLC topcoat to the WC-C. Hydrocarbon gas
is generally used in many cases, but continuous sputter
enabled the formation of interfaces to be suppressed,
which prevented interlayer peeling under high surface
pressures.
2. Development Concept
Engine valve trains that use a DLC coating perform
reciprocal motion, so there is an oil film break point with
a sliding speed of 0 m/s. This means that a hard bonding
layer is required below the coating film to enhance scuff
resistance.
On the other hand, the gear sliding environment has
a high maximum surface pressure of 2.2 GPa, but the
sliding speed does not go to 0 m/s, and there is no oil
film break such as in a valve train. For this reason it was
thought that a bonding layer is not required for the gear
DLC, so efforts focused on developing a topcoat and
coating film composition with an emphasis on wear
resistance.
276
DLC
WC-C
Substrate
(a) FIB-SIM image
Fig. 1
DLC
(topcoat)
Cr
WC-C
(interlayer)
Cr (adhesion)
1 m
0.6 m
Substrate
(b) Schematic diagram
DLC cross section
In addition, gears have an uneven shape (especially

the bevel gear), so uniformity of the coating thickness
between the tooth tip and the tooth root is an issue.
However, efforts were made to secure a uniform coating
thickness by adjusting the bias voltage and work
placement inside the DLC chamber. Figure 2 shows the
coating thickness reduction ratio in the direction towards
the tooth root, using the coating thickness at the tip of
the bevel gear tooth as the reference. This shows that
the developed DLC has a more uniform coating
thickness compared to the conventional DLC.
120
Developed DLC
100
Conventional DLC
(%)
80
60
Tip side
40
Flank center
20
Root side
Fig. 2
Flank center
Root side
4. Confirmation of Effects
Pin on disk seizing limit tests and actual transmission
tests were performed. In the pin on disk seizing limit
tests, oil was applied, wiped off, and then the load at
which the friction coefficient rose due to seizing
measured. As shown in Fig. 3, the seizing load increased
by 40%. This is thought to be due to an increase in the
Seizing pressure (MPa)
300
Friction
Shift gear
Bevel gear
Final gear
0.8 kW
1.3 kW
1.2 kW
adhesiveness between layers. In addition, engine tests

using an actual transmission verified an increase in both
performance and durability. Friction was reduced by a
total of 3.3 kW as shown in Table 1, and a transmission
efficiency of 97.0% was achieved.
Figure 4 shows the condition of the bevel gear and
the final gear after running in a 4-Race event. The figure
indicates that the damage at the tooth roots is miniscule
(indicated by the arrows in the figure), and that there is
no issue with durability.
A DLC coating for transmission gears has been

developed that has good durability under high surface
pressures. This DLC was applied to the shift gear and
the final gear from 2007, and to the bevel gear from
2008. Friction was reduced by a total of 3.3 kW, and a
transmission efficiency of 97.0% was achieved. In
addition, the developed DLC was confirmed to have
sufficient durability for use in 4-Race events.
Coating thickness distribution at tooth flank
350
Friction reduction of each DLC gear in

actual gear box
5. Conclusion
0
Tip side
Table 1
Acknowledgements
The authors wish to express their deep thanks to ICS
Corporation for their cooperation with the hard DLC
condition settings and gear coating production process.
Reference
Lubrication: semi-wet
Velocity: 0.2 m/s
(1) Suzuki, H., Ikenaga, M.: Jireidemanabu DLC

seimakugijutsu, Nikkan Kogyo Shimbun, p. 44-50 (2003)
(in Japanese)
250
+40%
200
150
100
Conventional DLC
Fig. 3
Developed DLC
Result of pin on disk test (DLC disk-DLC pin)
Author
(a) Bevel gear
Fig. 4
(b) Final gear
DLC after 4-race event
Tsutomu TANAHASHI
277
Yasunori ONAHA
Development of High Thermal Conductivity

Material for Heatspreader
Yu MURAI*
Masatoshi OKUMURA*
ABSTRACT
The DC-DC converter is key to developing small, lightweight power control units for Formula One hybrid systems.
A new material consisting of a silver-impregnated diamond powder compact, was developed in order to enhance the
heat dissipation performance of the DC-DC converter. The coefficient of thermal conductivity of the developed material
is three times that of a conventional aluminum material. The use of the silver-diamond composite material in a
heatspreader has reduced the temperature increase of the switching elements by 33%, offering a good prospect for the
extension of continuous boost from 1.4 to 6.6 seconds, the maximum regulation figure.
1. Background and Aims
Table 1
From 2009, the use of the Kinetic Energy Recovery

System (KERS), a hybrid system for Formula One
vehicles, is allowed by Formula One regulations. The
KERS systems enable 60 kW of motor power and 400
kJ of regenerated energy to be used per lap to provide
boost-up. Extension of the time for which continuous
boost can be provided will help in overtaking on the
straights, and will increase the degree of freedom in
planning race strategies.
The KERS power control unit (PCU) consists of a
DC-DC converter (VCU) that controls voltage, and an
inverter that controls motor current. The heat load is
highest on the switching elements in the VCU, which
have a high operating frequency. The project discussed
in this paper set out to enhance the heat dissipation
performance of the VCU heatspreader in order to
increase motor output to the regulation upper limit of 60
kW and enable 6.6 seconds of continuous boost (full
boost).
Material
Properties of materials for heatspreader

Thermal
conductivity
W/mK
Coefficient of
thermal expansion
10-6/K
200
630
23.4
5.5
A6063
Ag-CD
g/cm3
Volume fraction
of diamond
%
2.7
6.0
62
Density
plates. If the thermal expansion coefficient of the

heatspreader was similar to that of AlN, the buffer used
to relieve heat stress could be removed, and thermal
resistance would be reduced (Fig. 1). As Table 1 shows,
the production of a material from a composite of
diamond and silver (Ag), both materials with high
coefficients of thermal conductivity, enabled the
achievement of a coefficient of thermal conductivity
more than three times higher than the conventional
material. In addition, the volume fraction of diamond,
which has a coefficient of thermal expansion of 1.1 x
10-6 K-1, has been set at 62% in the composite material,
giving it a coefficient of thermal expansion of
5.5 x 10-6 K-1, close to that of AlN (Fig. 2).
Chip
Solder
Chip
AlN
2.1. Study of Material

In order to enhance heat dissipation performance, the
heatspreader material would be required to display a high
coefficient of thermal conductivity and a coefficient of
thermal expansion similar to that of the insulating plate
of a circuit board. Aluminum nitride (AlN), a material
with good thermal conductivity, is used for insulating
278
Brazing
Al buffer for heat stress relief
Al 6063
Solder
AlN
Solder
Heatspreader
Ag-CD
Pin fin
Former spec.
Fig. 1
Developed spec.
Cross-section of power device
2.3. Development of Component

The diamond composite material displays a high level
of hardness, making cutting and grinding extremely
challenging. Water jet machining was used to drill the
through holes used to fit the case and to cut the exterior,
procedures which do not require high precision. Electric
discharge machining was employed to produce the
grooves used for the mounting of the thermocouple, the
taper holes and the blind holes, for which precision is
essential.
In addition to reducing the coefficient of thermal
expansion of the heatspreader, specifications were
employed for the potting resin, which gave it a low
coefficient of thermal expansion of 10 x 10 -6 K -1. No
performance decline was observed following tests of
thermal shock resistance in power cycle tests.
4. Results
Heat dissipation performance was evaluated in
thermal resistance tests of prototype power modules
using heatspreaders manufactured from Al6063 and the
Ag-CD material.
In a comparison using the same finless shape, the
temperature increase of the switching elements was 35%
lower in the power module using the Ag-CD
heatspreader than in the module using the conventional
Al6063 heatspreader. As Fig.4 shows, when the
developed material was applied in the manufacture of a
heatspreader that also incorporated a change in shape, the
temperature increase of the switching elements was
reduced by 33%. This indicates a solid potential for the
extension of continuous boost time from 1.4 to 6.6
seconds per lap (full boost) at a cooling water
temperature of 80 C or below, as assumed for an entire
racing season.
180
170
Chip temperature (C)
2.2. Manufacturing Method for Material

A fine crystalline diamond powder with a mean grain
diameter of 0.1 mm was used in order to enable maximal
exploitation of the characteristics of diamond, which
possesses a high coefficient of thermal conductivity. The
diamond powder was pressurized inside the mold to
increase the filling factor. The metal was impregnated
into the material and a heat treatment was applied to
prevent distortion, following which the compact was
demolded. Figure 3 shows the microstructure of the AgCD material.
Al 6063 with pin fin

1.4 sec
160
150
120
Ag-CD plane
110
6.6 sec
100
Regulation limit
90
Thermal conductivity (W/(mK))
800
80
CTE
Good
700
Potting resin
10 ppm <
TC
Good
500
Ag
400
Fig. 4
Go
Cu
AlN4.5
Former
6063
SiC
10
15
20
25
Coefficient of thermal expansion (10-6/K)
Fig. 2
Thermal conductivity and CTE of materials
0.1 mm
Chip temperature after boost
An Ag-CD material displaying a higher coefficient of

thermal conductivity and lower coefficient of thermal
expansion than a conventional Al material was
developed. The use of the developed Ag-CD material in
the VCU heatspreader enhanced heat dissipation
performance by 33%, offering good prospects for the
ability to provide full boost for 6.6 seconds per lap for
an entire racing season.
Diamond powder
Author
Ag matrix
Fig. 3
5. Conclusion
Other materials
for heatspreader
Si3N4
Al
Si AlN
100
od
300
200
Boost time (sec)
Developed
Ag-CD
600
33%
Chip temp. upper limit
140
130
Microstructure of Ag-CD material
Yu MURAI
279
Masatoshi OKUMURA
Development of High Induction Stator Core
Takehiro MIYOSHI*
ABSTRACT
Stator core iron loss was reduced by 60% by optimizing the thickness and microstructure of iron-cobalt alloy,
which shows a high induction, and by developing a new adhesive and surface treatment technology. A manufacturing
process was developed that satisfies the required magnetic characteristics and dimensional accuracy, and a new
production system was constructed for laminated parts. The newly developed stator enabled a 20% reduction in the
motor size.
Application of hybrid systems to Formula One racing

cars was planned from 2009, and the increased weight
and effects on aerodynamics of adding motor-related
parts was an issue. Reducing the size and weight of
motor parts is essential for maintaining competitiveness,
and the aim was to reduce the stator core size by
applying a high induction permendur (Fe 49%, Co 49%,
V 2%, hereafter FeCo materials). FeCo materials are
used in the military, aerospace and other fields, but the
steel type specifications are limited and there is also a
long manufacturing lead time. Higher induction and
lower iron loss were pursued for racing use, and efforts
were also made to establish material specifications and
a stator core manufacturing method that satisfy the
required dimensional accuracy and productivity.
2. Developed Technologies
2.1. Material Development
Figure 1 shows the target material property values.
Compared to Fe-Si (Base), existing FeCo materials
(original) have high iron loss, so the target was set at
an iron loss equal to or less than that of Fe-Si. Iron loss
is classified into the two types of eddy current loss and
hysteresis loss, and eddy current loss is proportional to
the square of the thickness, so eddy current loss can be
reduced by reducing the thickness(1). However, reducing
the thickness raises the issues of a lower lamination
factor, lower rolling and pressing quality, and lower
productivity due to an increased number of layers. The
iron loss values of cold rolled materials with a thickness
of 10 m to 150 m were measured under conditions
of 1.0 T and 400 Hz. The magnetic characteristics were
measured using a toroidal-shaped laminated ring with an
outer diameter of 25 mm and an inner diameter of 15

mm. Figure 2 shows the structure after rolling, and Fig.
3 shows the magnetic characteristics. The lowest iron
loss value occurred at a thickness of 50 m. Reducing
the thickness lowers the eddy current loss, but at the
same time the structure becomes finer, resulting in an
increased hysteresis loss. This is thought to be why the
total iron loss is not reduced at a thickness of 50 m or
less. The lamination factor drops when lamination is
performed with a thickness of 50 m, so the thickness
was set at 100 m in consideration of the balance
between the lamination factor and iron loss.
An insulation layer is needed between the steel sheet
layers to reduce eddy current loss, so thinness, insulation
performance, and productivity are demanded. It was
difficult to coat thin plates with a zinc phosphate
insulating film for Fe-Si, and the required insulation
performance could not be obtained. Various insulation
treatments were evaluated, and the results showed that
eddy current loss was most reduced by the oxide film.
This film was formed by adding a heat treatment process
after magnetic annealing.

280
2.4
Induction (T)
1. Introduction
FeCo
(original)
Target
2.0
Base
1.6
1.2 Amorphous
0.8
Fig. 1
Thin
Fe-Si
Low loss
Fe-Si
High TS
Fe-Si
Fedust
10
15
20
Iron loss (W/kg)
Target properties
25
30
Calking and welding produce strain that leads to a

rise in iron loss, so the laminated materials were fixed
using adhesive. This adhesive required adhesive strength,
thinness, heat resistance and oil resistance, so a newly
developed heat resistant adhesive with low viscosity was
applied. Figure 4 shows the insulation film layers. The
adhesive layer together with the insulation layers on both
sides have a combined thickness of just 3 m, which
enabled a stator lamination factor of 97% or more.
2.2. Heat Treatment and Lamination Process
The heat produced by magnetic annealing causes the
fixing jig to adhere and deform, so deformation was
suppressed by adding a permeable ceramic. This work
method enabled simultaneous heat treatment of multiple
stators.
FeCo materials show order-disorder transformation by
magnetic annealing, resulting in different dimensions
before and after treatment. Press accuracy was achieved
by using a die that is adjusted for heat deformation.
Securing flatness of the edge surfaces after adhesive
bonding was difficult, but parts accuracy was achieved
by increasing the jig accuracy, use of vacuum debubbling, and optimizing the core-turning and lamination
procedure.
3. Effects on Performance
Copper wire was wrapped around a core yoke, and
the DC BH characteristics and iron loss value were
measured to confirm the magnetic characteristics of an
actual stator core. Figure 5 shows an exterior view of
the test sample. The induction B50 was 2.0 T or more,
and iron loss was 12 W/kg or less at W10/400, which
met the targets. A motor using the newly developed
materials achieved 15% higher torque with a size 20%
smaller than that of a motor using Fe-Si materials.
Fig. 5
BH measurement sample
4. Conclusion
t = 150 m
t = 10 m
50 m
50 m
Fig. 2
Acknowledgments
25
Iron loss (W/kg)
Microstructure
(1) Using existing FeCo materials as a base, the thickness

and structure were optimized and a new surface treatment
was developed. This simultaneously achieved high
induction and low iron loss.
(2) A machining process for ultra-thin laminated materials
was developed that both satisfies dimensional accuracy
and has high productivity.
(3) The motor size was reduced by 20%.
Total loss
The author wishes to express his deep thanks to

related parties at Jeftec Co., Ltd., for their cooperation
in establishing the manufacturing process for this part.
20
15
Hysteresis
10
Reference
5
Eddy current
0
0
50
100
150
(1) Matsuoka, H.: Wakaru Denjikohan, Nippon Steel

Corporation, p. 40 (1985) (in Japanese)
Thickness ( m)
Fig. 3
Iron loss and thickness
Author
2 m
FeCo
Oxide layer
Adhesive
FeCo
Fig. 4
Surface treatment and adhesive layer
Takehiro MIYOSHI
281
Weight Reduction Study for Side Panel

of Formula One Monocoque
Yasuhiro YAMADA*
ABSTRACT
This paper describes the research conducted on Carbon Fiber Reinforced Plastic (CFRP) for the side panel of a
Formula One monocoque.
The elongation ratio of a matrix resin was increased, thus increasing intrusion load 30% by load distribution. This
enabled weight reduction by 1.1 kg for only a segment of the side panel.
In addition, weight reduction research was further conducted on single-sheet side panel, an alternative option to
aluminum-honeycomb sandwich side panel, and the resulting technical issues were clarified.
1. Introduction
A several-mm-thick material called CFRP is used for
the Formula One monocoque. CFRP is made by laying
up multiple 100 m-thick prepreg sheets, which is a
carbon fiber fabric impregnated with matrix resin, and
hardening the layered sheets. In addition, sandwich
structures in which an aluminum honeycomb is
sandwiched between CFRP increased strength and
rigidity, and weight reduction is further required.
The Federation Internationale de lAutomobile (FIA)
sets regulations for intrusion resistance in side impacts
to help ensure side panel safety(1).
The FIA updates side panel regulations every few
years in order to enhance safety, and each time the
monocoque weight is increased.
Weight increases are expected to continue in the
future following regulation updates, so the mechanism
of intrusion resistance was studied, a matrix resin for
side panels was developed, and weight reduction research
using optimized panel structures was initiated.
affixing bolts, is penetrated at a rate of 2 1 mm/min.

The FIA stipulates three regulation parameters: intrusion
load, energy absorption, and fracture morphology. In
2005, all three of these parameters were increased to the
following: intrusion load: during the first 100 mm of
displacement by the penetrator, the load must exceed 250
kN; energy absorption: during the first 100 mm of
displacement by the penetrator, energy absorption must
exceed 6 kJ; and fracture morphology: after the
penetrator displaces 150 mm, there should be no damage
to the border or to the fixture.
Fig. 1
Side panel area of monocoque
2. Side Panels and Intrusion Regulations

In Fig. 1, the side panel is shown as the shaded
portion. The side panel takes up over 50% of the surface
area of the monocoque, thus it is a component that
benefits greatly from weight reduction technology.
Figure 2 shows the setup for the intrusion test of the
side panel.
For the intrusion test, a test panel is used to simulate
a laminate on the side panel section of the monocoque.
This test panel, which has a 550 mm x 550 mm outer
dimension framed with a 25 mm wide metal frame for
282
Fig. 2
Intrusion test setup

Table 1
Side panel regulation transition
Destructive
Maximum load; Energy absorption;
morphology;
first 100 mm of first 100 mm of
after 150 mm
displacement (kN) displacement (kJ)
displacement
Year
2001-2004
150
2005-2006
No destruction
to frame
Additional extra
secondary panel
250
2007-
Others
Each racing team is conducting research to clear

these regulations by weight reduction applications.
Moreover, since 2007, the FIA has ordered the addition
of FIA-designated, 6.2 mm-thick laminates to portions
of the exterior side panels that satisfy the above
requirements(2) (Table 1).
3. Technology Elements
The impact of carbon fiber-type, panel structure,
aluminum honeycomb density, matrix resin of CFRP was
confirmed.
3.1. Carbon Fibers
Intrusion tests were conducted on CFRP made from
commercial carbon fibers. The carbon fiber series
T1000G manufactured by Toray Industries, Inc. was
confirmed as offering the highest tensile strength and
best intrusion resistance among available carbon fibers.
3.2. Panel Structures (Sandwich Panel and Single
Sheet)
Formula One monocoque is designed to receive the
highest intrusion load when the inner wall of the panel
is penetrated. Figure 3 shows intrusion test results for
the sandwich panel and single sheet. For the sandwich
panel, the top panel that corresponds to the outer wall
panel of the monocoque was penetrated with 142 kN of
load. Then, the bottom panel that corresponds to the
inner wall panel of the monocoque was penetrated with
261 kN of load. Also, the single sheet was penetrated
with 200 kN of load, 61 kN less than the bottom panel
of the sandwich panel, despite being 1.3 mm thicker than
the bottom panel of the sandwich panel. Furthermore, if

the 10 mm thickness of the honeycomb is subtracted, the
displacement of the single sheet is found to be 4 mm
less than bottom panel penetration.
For the sandwich panel, the top panel and aluminum
honeycomb is considered to disperse the load in the inplane direction, so less stress would be generated at the
bottom panel under the same load. In other words, the
load distribution was expected to be an integral technical
element for increasing intrusion resistance.
3.3. Aluminum Honeycomb Density
With sandwich panels that use low-density aluminum
honeycomb for weight reduction, the intrusion load is
decreased. That is, honeycombs with low density have
low compressive strength and shear strength, and
consequently, is unable to distribute the load acting on
the top panel over a wide area of the bottom panel (Fig.
4). Technology is required that supports reduced
intrusion loads for low-density honeycomb.
3.4. Matrix Resin
Pre-impregnated matrix resin, used as CFRP
materials, takes full advantage of the special
characteristics of carbon fiber, including high strength
and high elastic modulus. However, the high strength
and high elastic modulus characteristics of CFRP were
expected to adversely affect intrusion resistance because
of low load distribution. Therefore, CFRP trials were
made with two kinds of epoxy resins, one with a high
elastic modulus and another with a high elongation ratio.
Then, intrusion resistance tests were conducted on the
side panel. As anticipated, the results showed that matrix
resin with high elongation ratio has greater intrusion
resistance (Fig. 5).
Thus, it was confirmed that low-density honeycombs
and matrix resin with excellent load distribution
Shear destruction of honeycomb
Jig
Top panel
Bottom panel
Stress distribution area
300
Bottom panel
10 mm
(honeycomb)
Fig. 4
Side panel section before bottom panel intrusion
250
4 mm
Single panel
300
61 kN
250
Top panel
150
Load (kN)
Load (kN)
200
Sandwich panel
(Top 1.3 mm / Al honeycomb 10 mm
/ Bottom 2.7 mm)
100
200
High elongation matrix

High modulus matrix
150
100
Single panel (4.0 mm)
50
50
50
100
Displacement (mm)
150
Fig. 5
Fig. 3
Intrusion test results of sandwich / single panel
283
20
40
60
80
100
Displacement (mm)
120
140
Intrusion test results of high modulus/

elongation matrix CFRP panel
Weight Reduction Study for Side Panel of Formula One Monocoque
properties are a necessary technical element for side

panel weight reduction.
4. Research on Matrix Resin for High

Intrusion CFRP
An aluminum honeycomb with a density of 10 lb/ft3
has been used for the side panel and weight reduction
was conducted by using 8 lb/ft3 of material. Research
was then conducted under the concept of compensating
for reduced intrusion loads due to decreased load
distribution by increasing the elongation ratio of the
matrix resin.
By optimizing the resin components, RH421 epoxy
resin with a 4% elongation ratio manufactured by Nagase
ChemteX Corporation was obtained for the final
specification. Figure 6 shows the mechanical properties.
An intrusion resistance test was conducted on a test
sandwich panel that used RH421 and an aluminum
honeycomb with a density of 8 lb/ft3. The results showed
Commercial grade
(estimated)
50
40
30
20
2
Elongation (%)
Fig. 6
Research was conducted on further weight reduction

with the goal of conforming single-sheet side panels to
regulations without using aluminum honeycombs.
Physical properties of trial epoxy resin
Section
55
300
Bend.,
compression
Tension
Cycom2020 panel
Fig. 7
Intrusion load (N)
250
RH421 panel
Panel appearance after intrusion test
200
150
100
Fig. 9
Bending
modulus
Compression
modulus
Bending
strength
Fig. 8
45
3.5
Compression
strength
Intrusion load efficiency
0
Tensile
modulus
Shear
strength
Intrusion load
50
Tensile
strength
Shear
modulus
50
Intrusion load efficiency

(kN/mm 100)
Tensile strength (MPa)
Final specification
RH421
5. Research on Single-sheet Side Panel
5.1. Verification of Composite Results for Low Elastic

Modulus Layers
Figure 9 shows the intrusion load for single-sheet
CFRP. When the panel thickness was increased, the
intrusion load increased but the intrusion load efficiency,
obtained by dividing intrusion load by the panel
thickness, decreased. This was thought to be caused
because load distribution was decreased together with the
increase in panel rigidity.
Functional materials similar to the aluminum
honeycomb of the sandwich panel were researched to
help distribute the load. Single-sheet side panels were
made for trial purposes (Fig. 10). For the panels, an
organic fiber layer with an elastic modulus less than that
of carbon fiber was arranged in between the CFRP layers
or on the CFRP outer layer. As a result, the intrusion
70
60
a 30% increase in intrusion load over the existing

Cycom2020(3) product specification.
Figure 7 shows the panel intrusion holes. The
Cycom2020 specification shows breaks at sharp angles
around the intrusion hole area as opposed to the gentle
bends exhibited by the RH421 specification.
It is thought that, because of reduced CFRP bending
and the reduced compression elastic modulus (Fig. 8),
bending and compression stress concentration, which are
CFRP weak-points, did not occur and these stresses were
successfully converted into tensile stresses. Thus, the
high elongation ratio of the matrix resin allowed weight
reduction of the aluminum honeycomb and the increased
intrusion load to be achieved simultaneously.
4.0
4.5
5.0
5.5
Panel thickness (mm)
6.0
Intrusion load and intrusion load efficiency of

single panel
Inner
Center
Outer
RH421/T1000G (developed)
Cycom2020/T1000G
(conventional ; Cytec)
Low modulus layer
Mechanical properties of CFRP
Fig. 10
284
Example of low modulus layer arrangement
load was increased more than for the single-sheet CFRP

panel, with the exception of one specification, even
though there was variation depending on the difference
in the kind of organic fiber and arrangement (Fig. 11).
280
Intrusion load (kN)
260
Weight reduction
Good
240
ne
d li
loa RP
n
o
CF
usi
Intr single
of
220
200
180
5500
Fig. 11
6000
Honeycomb sandwich panel

Single CFRP panel
Low modulus layer single panel
6500
7000
Panel weight (g/m2)
7500
6. Conclusion
10000
Energy absorption (J)
Honeycomb sandwich panel

Single CFRP panel
Low modulus layer arranged single panel
8000
6000
4000
0
30
Fig. 12
No delamination
to frame
35
5.2. Issues of Single-sheet Low Elastic Modulus

Composites
Figure 12 shows the remaining issues related to
energy absorption and the fracture morphology for the
side panel. As the amount of the energy absorption
increases the intrusion displacement increases; however,
because single sheets do not undergo a two-stage
intrusion like the sandwich panel, the amount of the
energy absorption is low. Also, when intrusion
displacement increases, delamination is generated at the
organic fiber layer and the carbon fiber layer interfaces
and extends to the edge of the panel (Fig. 13). Because
of this fracture morphology, regulations have not been
achieved. Additional control technology for interface
delamination is required in single-sheet manufacture for
side panels.
8000
Effect of low modulus layer arrangement for

intrusion
2000
Delamination
to frame
40
45
Intrusion displacement (mm)
50
55
Energy absorption of low modulus layer

single panel
The technology that maintains intrusion resistance

with sandwich panels having low-density aluminum
honeycomb was established by producing a high
elongation ratio CFRP matrix resin.
Side panels with RH421 applied allowed the bottom
panel to be thinner, and in combination with a lowdensity aluminum honeycomb achieved a total weight
reduction of 1.1 kg. These specifications were officially
approved by the FIA.
An actual test using a real vehicle with these side
panels was performed and verified with no decrease in
driving stability performance.
On the other hand, it was clear that interface
delamination control technology is necessary regarding
single-sheet construction that will allow further weight
reduction.
References
(b) Side view
(1) Federation Internationale de IAutomobile (FIA): Side

intrusion test procedure, Appendix to the 2007 FIA
Formula One Technical Regulations, Article 18. 6, p.
34-35 (2006)
(2) Federation Internationale de IAutomobile (FIA):
Additional side intrusion panels, Appendix to the 2007
FIA Formula One Technical Regulations, Article 15. 4.
7, p. 26-27 (2006)
(3) http://www.cytec.com/
(a) Panel appearance
Author
Delamination
Low modulus layer
CFRP
Iron frame
(b) Side view
Fig. 13
Appearance after intrusion test of low

modulus layer single panel
Yasuhiro YAMADA
285
Development of Brake Caliper Production Process

with High Strength Al-Li Material
Akihiro YANASE*
Hiroshi YAMADA*
ABSTRACT
A production process was developed that uses high-strength Al-Li materials to reduce the brake caliper weight.
Anisotropy of material strength due to metal flow was eliminated with 3-axis multi forging, and an 8% increase in
fatigue strength was achieved by T6 heat treatment with 3-stage aging. This is expected to increase the service life to
4 times that of current calipers.
Reducing a vehicles unsprung weight contributes to

enhancing its vehicle dynamics, so brake caliper weight
reduction is demanded. A caliper development project
was started with the aim of reducing the weight by 150
g per caliper (600 g per vehicle). As the current caliper
materials, the 2099 materials of aluminum-lithium (AlLi) alloy is used, which are high stiffness, low density
aluminum materials. However, these materials have
anisotropy of material strength, which posed design
restrictions. Therefore, a forging process was established
with the aims of eliminating anisotropy of material
strength, and a heat treatment was developed with the
aim of increasing strength.
2.2. Forging Process for Eliminating Anisotropy of

Material Strength
Anisotropy of material strength could not be
eliminated by 1-axis forging of the 2099 extrusion bar.
Therefore, 3-axis multi forging was applied, which is an
effective method for eliminating anisotropy of material
81
Youngs modulus [GPa]
1. Introduction
Regulation
2090
79
8090
2099
77
75
2618
73
2.5
2.6
2.7
2.8
2.9
Density []

286
Fig. 1
Modulus and density
380
Maximum stress L [MPa]
2.1. Selection of Materials

Figure 1 shows the material properties of the Al-Li
alloy, and Fig. 2 shows the anisotropy of material
strength from the fatigue strength in the longitudinal (L)
and transverse (T) directions of the materials. The fatigue
strength data of 2618 materials for the piston is also
shown for comparison reference.
Under Formula One regulations, caliper materials are
prescribed as aluminum materials with a Youngs
modulus of 80 GPa or less. As the current material, a
2099-T83 extrusion bar of Al-Li alloy is used that
displays low density, high strength, and high Youngs
modulus. However, this material has anisotropy of
material strength, and the low T direction strength
restricted the design.
2099
ext. bar
360
340
8090
forged
320
300
Target
2090
forged
280
260
2618
240
240 260 280 300 320 340 360 380
Maximum stress T [MPa]
Fig. 2
Fatigue property
strength. Before die forging, 50% upset forging from the

three X, Y and Z axes was performed to eliminate
anisotropy of material strength by a hydraulic press.
Figure 3 shows the set forging process.
2.3. Heat Treatment Conditions for Increasing Strength
The 2099 materials are a deposit strengthening type
alloy, with deposit phases (Al 3Li) and T 1 phase
(Al2CuLi) deposit phases. Cold compression to introduce
a dislocation that serves as the deposit phase core is an
effective means of increasing strength, but this is
difficult to apply to calipers, which have a complex
shape. Heat treatment with multi stage aging is also
effective in generating a core for the deposit phase(1), so
T6 heat treatment with 3-stage aging was developed with
solution treatment for 2 hours at 549C, followed by
aging (treatment) for 48 hours at room temperature, 24
hours at 120C, and then 100 hours at 165C as the heat
treatment conditions for the 2099 forged materials.
The status of each deposit phase was analyzed using
a transmission electron microscope (TEM) and
differential scanning calorimetry (DSC), and the
following mechanisms of deposit strengthening were
confirmed: 1) formation of a GP zone by room
temperature aging, 2) phase deposition by 120C
aging, and 3) T1 phase deposition with the GP zone as
the core by 165C aging.
T6 heat treatment with 3-stage aging. Test pieces were

sampled in the L and T directions of the 2099 forged
materials as shown in Fig. 5, and tensile-tensile fatigue
tests were performed.
Figure 6 shows the results. The current 8090 and
2099 extrusion bars exhibit anisotropy of material
strength, with the fatigue strength differing according to
the location from where the test piece samples were
taken. However, the developed 2099 forged materials
introduced in this paper had the same fatigue strength
regardless of sample location, so anisotropy of material
strength was eliminated. In addition, an 8% increase in
fatigue strength was confirmed relative to the T direction
of the 2099 extrusion bars, which had been a design
restriction. As a result, the strength in the weakest area
of the part was increased, giving prospects for a fourfold increase in durability.
2.4. Evaluation of Material Properties of 2099 Forged

Materials
Prototype 2099 forged materials for calipers were
manufactured using the 3-axis multi forging process and
Fig. 5
Sampling regions
Maximum stress at 80 deg C [MPa]
380
2099-T83 ext. bar
z
Heat
treatment
y
x
Die forging
Rotation y
1st upset
Rotation z
2nd upset
3rd upset
Target
340
320
300
8090 forged
T6 (2006)
2099 extrusion
T83 (2007)
2099 multi-forged
3 stage T6
Caliper forging process

Fatigue property of caliper body materials
(110)
3. Conclusion
(Al3Li)
T1 (Al2CuLi)
Fig. 4
8% UP
360
Fig. 6
(100)
280
3-axis multi forging

Fig. 3
TEM photograph of 2099 after 3 stage T6

( x 150000 )
This development established production process

technology that eliminates anisotropy of material strength
by 3-axis multi forging, and increases fatigue strength
by 8% compared to the current materials by T6 heat
treatment with 3-stage aging. Converted to service life,
this increased durability by 4 times.
287
Development of Brake Caliper Production Process with High Strength Al-Li Material
Reference
(1) Romios, M., Tiraschi, R., Parrish, C., Babel, H. W.,
Oqren, J. R., Es-Said, O. S.: Design of Multistep Aging
Treatment of 2099 (C458) Al-Li Alloy, Journal of
Materials Engineering and Performance, p. 641(2005)
Author
Akihiro YANASE
288
Hiroshi YAMADA
Descriptions of Race Management Technologies
Development of Lifecycle Management System

for Racing Parts
Hidetada TANIGUCHI*
Shigeo MIYAJIMA*
Katsunori TAGUCHI*
Osamu WATANABE**
Masataka NODA*
ABSTRACT
In racing operations, parts development and supply must be constantly reviewed in a short time with regards to
changing circumstances in order to simultaneously promote upgrading and supplying of parts with the aim of
maintaining competitiveness relative to other teams. For this reason, a part lifecycle management system was constructed
to increase development speed and effectively utilize resources. This system aimed for centralized management of
part plans and results data for all stages, from investigation of part design specifications to drawing issue, production
plans, arrangement and purchasing, procurement and manufacturing, receiving inspection, inventory control, assembly,
usage history, and disposal. Centralized data management enabled instant sharing of information across the Racing
Division, and provided prospects for creating an environment that enables accurate and prompt decision making even
at racing sites.
1. Introduction
The Racing Division of Hondas Automobile R&D
Center (Tochigi) developed, assembled and supplied
engines, developed advanced chassis technologies, and
also developed and supplied internal gears and gear
shifting parts for the gear box to the Honda Racing
Formula One Team (HRF1). Engines were also
assembled and supplied by Honda Racing Development
Ltd. (HRD) in England, and in addition to parts
development and supply to race sites from Tochigi,
locally procured parts were supplied from HRD.
In racing operations, development speed is demanded
to maintain competitiveness, and this assumes a unique
operations flow (F i g . 1 ). Parts with multiple
specifications are constantly designed, arranged and
tested, then just before assembly the test results are
reviewed to select the optimum specifications from
among the current inventory and parts scheduled for
delivery, and the selected parts are assembled into
products and supplied. In addition, arrangements must be
made concurrently for development parts and racing
supply parts, and assembly specifications also change
frequently. This means that a diverse and large volume
of inventory is essential in securing a constant stock of
parts that might be used.
Assembly specifications and inventory conditions
change constantly, including support for races that are
held every other week, and it is also necessary to
accurately record part assembly results and usage
histories to provide feedback on trouble information.

Parts shifting and plan changes based on tests and race
results also had to be shared between Tochigi and HRD,
so speed and accuracy of information were also
demanded.
In addition, the recent application of regulations
limiting engine usage had tended to increase the need
to manage parts with the same part number by quality
rank, and to perform individual part lifecycle
management.
The system constructed for production car
development operations was also used to manage racing
parts until recently, but no fundamental changes had
been made to suit the environmental changes and
operation contents unique to racing operations.
In the operational flow of production car
development, the prototype assembly specifications and
assembly schedule are determined, and then the parts
resources needed for assembly are secured before
purchasing (Fig. 2). For this reason, the constructed
conventional system lacked the flexibility required by
racing operations, and constantly changing information
could not be accurately reflected.
To overcome this issue, the information needed by
each department was provided separately using other
system ledgers and tools, and had to be collated
manually based on experience or by using large numbers
of people. In addition, when the need for coordination
arose between Japan and England, which are separated
by both physical distance and time difference, atypical

** FUJITSU LIMITED
290

Provisional assembly
planning
Drawing
Drawing
Drawing
Purchasing
Purchasing
Purchasing
Assembly
planning
Manufacturing
Manufacturing
Manufacturing
Assembly
Test
Assembly
planning
Inventory
Assembly
Re-assembly
planning
Inventory
Assembly
planning
Inventory
Assembly
Re-assembly
Test
Race
Assembly
planning
Drawing
Work flow
Assembly
Race
Decision
Purchasing
Availability Information
Fig. 1
Assembly
planning
Drawing
Purchasing
Fig. 2
Manufacturing
Inventory
Racing operations
Manufacturing
Inventory
Assembly
Test
Production car development
and detailed exchanges via telephone or e-mail consumed

much time and manpower.
As a result of the above, it was sometimes difficult
to make optimum decisions promptly, finding
replacements for lacking parts took time and effort, and
excess inventory sometimes occurred, all of which were
factors hindering the effective use of development
resources.
2. Development Policy
The newly developed system aimed to share accurate
parts management information to enable the timely and
accurate decisions needed for racing operations. In
addition, construction of an environment enabling the
maximum utilization of development resources was also
expected as a result.
To achieve this, the target was set as centralized
management of constantly up-to-date plans and results
for all stages, from the investigation of part design
specifications to drawing issue, production plans,
arrangement and purchasing, procurement and
manufacturing, receiving inspection, inventory control,
assembly, usage history, and disposal.
The specific development indexes were determined as
follows.
(1) Faster verification of assembly plan feasibility
(2) Increased purchase order speed
(3) Enhanced parts management information
Construction of the new system required that the
operations flow be restructured, so members selected
from all related departments thoroughly verified the work
flow, and reviewed operational roles to achieve the
optimum overall balance.
Due to the nature of racing operations, speed and
flexibility are demanded, so the division between
contents that should be decided by humans and areas that
can be left to the system was thoroughly discussed, and
the specifications were determined.

In addition, measures were taken to promote accuracy
and maintain up-to-date information, including systems
that provide tangible effects, such as letting not only
people make decisions based on centrally managed
information, but also letting people in charge of entering
information view the necessary information.
An advance project was commenced in Spring 2006,
and work began to understand the current status of all
operational issues involving racing parts management at
Tochigi. Based on these results, operational roles were
redefined starting from Spring 2007, and construction of
the new system began from the start of 2008.
3. Function for Verifying Assembly Plan

Feasibility
An integrated system was constructed that links and
verifies the control number (unit number) information
assigned to each assembled product (A), assembly
schedule (B), assembly part configuration sheet (C),
arrangement, purchasing and delivery information for
each part (D), inventory information (E), assembly
results (F), and usage history (G) (Fig. 3).
The system uses alert displays to automatically
indicate which part impedes the assembly flow when the
assembly plan for each unit number is changed, and all
related persons can share information instantly. This lets
people make comprehensive judgments regarding
responses to various issues, and these decision results are
reflected to the system, enabling the latest plan
information to be shared.
The individual functions that comprise this system are
described below.
3.1. Assembly Part Configuration Management for
Each Completed Product
An assembly part configuration sheet function was
291
developed as a core function for verifying plan

feasibility. This function was created by extracting the
plan management know-how accumulated by the Racing
Division, and managed the parts configuration for each
completed product. This function generates the original
data that is the basis for instructions to all subsequent
processes, and features the ability to change the
configuration even after parts are purchased and
manufacturing starts.
Assembly part configuration management for each
unit number was achieved by linking data such as noted
below (Fig. 4).
(1) Registration of parts by each specification
Multiple specifications are defined for each function
unit, and the components used in each specification are
registered. This enabled the selection and use of arbitrary
parts (selected parts) from among multiple types.
Operation team
3.2. Delivery Management of Purchased Parts

Functions enabling the swift provision of delivery
date information were strengthened based on split
delivery and delivery date changes according to the
supply capability of suppliers, which is common in
racing operations.
(1) Multiple delivery date response management
The system that links suppliers and purchase order
information was modified and a function was developed
so that multiple delivery date responses can be registered
Inventory
Assembly
A
Model
event
unit number
(2) Registration of specifications applied to unit numbers

The combination of specifications to be applied to
each function unit is registered for each unit number.
(3) Assembly part configuration sheet
The data from (1) and (2) above are linked to create
a list of all parts required to assemble each unit number.
B
Assembly plan
Test team
Assembly
F
Part issue
G
Mileage
record
Assembly
result
Designer
C
Drawing issue
Part registration
Specification application
Arrangement and
purchasing order
D
Arrangement and
purchasing order
Assembly plan feasibility monitor
Fig. 3
Assembly plan feasibility monitoring
Part registration for specifications

Piston
Part
Quantity
Common
L1
N5
Specification application for unit No.

Funnel
M
Round1
Region
Bench
R001
R002
Performance
Piston
L1
10
N5
20
Funnel
E1
10
E2
10
F1
memo
F2
Quantity
Assembly part configuration sheet
Part
Assembly
date
2007
06/01
2007
06/02
2007
06/10
Performance
Unit No.
R001
R002
Quantity
Piston
L1
Piston
N5
Quantity
10
14
20
22
Common
Common
20
28
10
14
10
14
20
28
E1
Funnel
E2
Funnel
F1
F2
Common
Common
Fig. 4
Assembly part configuration sheet
292
for a single purchase order.

(2) Confirmation of delivery date response statuses
Various supplier delivery schedule statuses can be
confirmed, such as delivery date responses not yet
received, delivery date delay responses, or changes to
delivery date responses. In addition, information
concerning delivery dates (fixed delivery dates) agreed
upon between the person making the purchasing order
and the supplier can be shared.
3.3. Process Progress Management for Manufactured
Products
When manufactured parts are completed by
machining materials that pass between the in-house
Manufacturing Division and multiple suppliers,
unfinished parts may be transferred directly between
suppliers to shorten the schedule. Process progress
management formerly relied on the experience of the
person in charge, and was performed by understanding
the status of only representative priority management
parts. Therefore, management was systematized and the
following functions were developed to increase
information accuracy and enable the sharing of
information.
(1) Process progress management
Process design and machining process schedules and
results can be displayed for all parts.
(2) Understanding progress using receipts
Unfinished parts do not necessarily move in
purchasing order units. Therefore, progress status was
understood by creating receipts and loading the receipt
information to the system when unfinished parts are
received by the next process.
3.4. Reserves Used to Calculate Arranged Quantities
To calculate the proper amount of parts required
(arranged quantity) amidst frequently changing
circumstances, it is necessary to compare the quantity
required by the assembly plan with the quantity
remaining to be ordered, current inventory, the quantity
scheduled for assembly, and other factors. Therefore, a
reserves function was developed to obtain the proper
arranged quantity.
First, the free inventory is obtained using Eq. (1)
(Fig. 5).
Inventory
quantity
Planned parts
selection & issue
Future inventory
(arranged but not yet ordered
+
ordered but not yet received)
Free inventory
Allocation for planned

assembly units
(planned but not yet
reserved )
Current inventory
Reserved inventory
(reserved for immediate
assembly work, shipping, etc.)
Fig. 5
Free inventory
X=(B+A)-C-D
(1)
X: Free inventory
B: Current inventory in parts warehouse
A: Future inventory (Sum of quantity arranged but not
yet ordered and quantity ordered but not yet received)
C: Reserved inventory (Quantity ready and awaiting
issue from inventory)
D: Allocation for planned unit numbers not yet reserved
(Quantity required by unit number plans but not yet
reserved for issue from inventory)
The final arranged quantity can be determined by
experienced personnel by taking into account irregular
factors such as the feasibility of reusing old parts based
on the shared part management information described
hereafter, and the number of race events remaining in
the season, and then adjusting the quantity with respect
to the free inventory obtained in the manner described
above. This enabled more accurate arrangements.
3.5. Inventory Sorting
Labels noting the unit number information based on
the assembly plan at the time of purchasing order
placement were formerly affixed to parts, but this was
not useful for racing part management in which assembly
plans undergo frequent changes. Therefore, an inventory
sorting function was developed to increase the sorting
speed and accuracy.
This function automatically creates prioritized sorting
proposals for available inventory based on the latest
assembly parts configuration sheets, assembly dates, unit
number purposes, applied specifications, and quality
rank. Humans then make a comprehensive decision to
determine the final sorting destination. This enabled
highly flexible operation.
4. Purchasing Order Related Functions

Various pieces of information are required when
ordering parts, and the links between this information
affect the actual speed required to place purchasing
orders. The speed required to place purchasing orders
was increased by strengthening and enhancing the
functions and processes described below.
4.1. Information Entry Prior to Drawing Issue
In the conventional system, determination of the
arranged quantity, supplier, tentative price and other
purchasing order operations were performed successively
after drawing registration (hereafter, drawing issue).
Centralizing information enabled more accurate advance
information to be shared, which made it possible for
these operations to proceed concurrently prior to drawing
issue (Fig. 6).
(1) Registration of drawing issue schedule
Various pieces of information concerning parts
scheduled for drawing issue are managed. The scheduled
drawing issue date, material quality, base parts and other
information that affects purchasing orders is entered.
293

Sequential operation
Drawing issue schedule
Drawing
Drawing issue
Concurrent operation
Drawing
Drawing issue schedule
Purchase
order
Increased purchase
order speed (2 days)
Drawing issue
Arrangement
preparation
Purchasing
preparation
Fig. 6
Arrangement
Purchase
order
Revised process of purchasing
Coordination issues are prevented by enabling part

numbers to be entered only from the drawing issue
schedule, which is the initial source of information, and
then sharing that information with subsequent drawings
creation and issue.
(2) Arrangement preparation
Information on the arranged quantity based on the
assembly plan, and the department requiring the parts,
is entered.
(3) Purchasing preparation
Suppliers and procurement methods for supplies are
investigated based on the drawing issue schedule
information, and the results are entered.
4.2. Enhancement of Drawing Issue Process
In the conventional drawing issue process,
information sometimes stalled partway through the
information transfer and other processes that relied on
paper-based approval and distribution. In contrast, the
newly developed system enabled instantaneous
information transfer using the following functions.
(1) Compatibility check when issuing drawings
The conventional system had a downstream process
whereby drawings were checked and revised by humans,
but the check function was strengthened so that part
configuration and part history compatibility checks can
be completed in the drawing issue stage.
(2) Specification change notices
Specification change notices can be automatically
created by extracting data from the drawing issue
schedules and advance information.
(3) Issue of specification change notice numbers
Specification change notice numbers are issued
immediately after a drawing issue is approved in order
to strengthen compatibility check functions when
drawings are issued.
5. Enhanced Parts Management

Information
5.1. Inventory Control
Separate ledgers kept outside the system thus far and
information managed by other systems without data links

were integrated to realize the following functions.
(1) Individual part management
Individual part management was enabled by assigning
a unique ID number to each part during the receiving
inspection, and printing or laser-marking a 2D barcode
on the label or directly on the part.
(2) Management by quality rank
This function enabled better inventory control by
ranking parts based on actual measurement results, and
grouping parts into ranks.
(3) Reserved and allocated inventory part list
This corresponds to C + D in Eq. (1), and is a list
of existing parts not yet issued from the inventory, from
among the parts required for assembly of each unit
number according to the assembly part configuration
sheet. When a design change occurs, it is automatically
reflected in this reserved and allocated inventory part list.
(4) Usage history management
Managing the part usage history enabled a more
detailed understanding of part replacement periods based
on lifecycle.
(5) Management of inventory outside the part management
warehouse
Total part quantities and locations can be clarified by
centrally managing the storage location and inventory in
each department, from the time parts are issued from the
part management warehouse until the parts are discarded.
(6) Management by assembly and unit number
Parts can be managed not only by part number, but also
by the arbitrary groupings of assembly and unit numbers,
which are used during storage and transfer (Fig. 7).
5.2. Assembly Results Management
Part assembly results were formerly managed by a
separate system that is not linked with the inventory
control system, and entering the results took time.
However, the newly developed system enabled accurate
recording in a short time. Engines in particular are
composed of approximately 1000 parts, so methods were
devised to efficiently and accurately record the results
for each unit.
294
Unit number Z
Assembly X
Part A : Rank: S Serial: 001

Part B : Rank: A Serial: 102
Assembly Y
Part C : Rank: S Serial: R01

Part D : Rank: S Serial: V07
Fig. 7
Unit number and assembly
(1) Increased assembly results in entry efficiency

The new system enabled quick and accurate entry of
assembly results in a semi-automatic manner instead of
manually, by scanning a 2D barcode of the individual
part ID number. In addition, automatic carryover of
inventory issue information for each part prevented entry
errors and at the same time reduced the data entry
workload.
(2) Easier assembly results verification
Assembly part configuration sheets and assembly
results can be compared, enabling instant understanding
of differences and easy comparison of different assembly
results.
(3) Link with manufacturing quality information
Serial number information is received from a separate
system that manages manufacturing quality information,
and linked with individual part ID numbers. This enabled
detailed information on manufacturing quality to be
referenced from the assembly results.
part in unit number units, and enabled identification of

up to eight different color-coded alerts for easy
understanding of the status at a single glance. In
addition, the layout provides launch buttons for each part
to enable the necessary measures to be taken according
to the alert type.
(2) Operation intensive type
The purchasing order information creation window is
used constantly, so all necessary operations should be
realized in a single window. Therefore, the layout was
divided into three areas, with basic information displayed
in the upper part, and the purchasing order form and
Alert
Launch button
Arrangement status
for each part
Fig. 8
Assembly plan feasibility monitor
6. System Application Technologies

The new system is not a customization of a
commercially available package, but was constructed
through dedicated development in consideration of
maintainability, expandability and feasibility of the
requirements unique to racing operations such as
described thus far.
When developing a system, the policies and goals of
systemization must be commonly held, or individuals
will devise arbitrary requirements, making unification of
specifications difficult. Conversely, excessive
standardization may affect ease of use. An open
application framework was used as the base for the
system, but feasibility of operations was given priority,
and construction aimed for a balance between
standardization and individuality.
The technical contents applied to each level are
described below.
6.1. Approach towards User Interface
Basic window configuration patterns such as search
and lists were provided for the user interface, but the
layout of each window was investigated thoroughly
together with users to realize demands for operability
that differs for each operation. As a result, the following
window configuration patterns were created to match
individual requests.
(1) Monitor panel type
The assembly plan feasibility verification window
shown in Fig. 8 displays the arrangement status for each
295
Basic information
Order form and

detailed information
Fig. 9
Purchase order entry
Unit No. list
Parts list
Fig. 10
Receiving confirmation

Table 1
Common functions
Common functions
Explanation
Usage pattern
a. Display
Table column
customization
Structure expansion
Memo
Enable each users preferred

display, structure and additional
information
Sorting or selecting data on users

demand
b. Multi
location
Multi language
Multi location
Switch displayed language for

menus, drop-down lists, etc.
Usable by multiple offices, both

domestic and overseas
c. External
interface
Excel input
Excel output
Bulk operation by using a

spreadsheet is possible.
Mass data inputs and outputs

essential for fast operations
d. Security
Menu privilege
Data privilege
Can be set for each user based on

their role and screen mode
Privileges control based on

department and/or title
and development efficiency. Table 1 summarizes the

common function technologies applied from the
viewpoints of (a) display, (b) multi location, (c) external
interface, and (d) security.
detailed information displayed in the lower part of the

window (Fig. 9).
(3) Simultaneous list display type
Both the unit number list and the parts list must be
checked at the same time, so two areas with separate
page switching control were implemented in the upper
and lower parts of the receiving confirmation window
shown in Fig. 10.
6.3. Approach towards Data Model

The data model did not use the data structure of the
conventional system, and was instead newly designed to
realize centralized management of information. Design
departments create design drawings and part
configurations in part number units, but the basic
information unit for ordering departments is the
purchasing order form. In addition, after parts are
received, important parts must be identified and managed
individually. Therefore, a database was constructed that
links the assembly part configurations, purchasing order
forms, and individual parts handled by each department
6.2. Common Function Technologies

Common functions were split from the operations
logic and implemented on the framework to increase user
convenience and system development efficiency.
Frequently-used functions were made into common
functions that could be accessed from any window, and
this enabled system developers to focus on the
operations logic design, which increased both quality
Application specification
Arrangement
Unit001
Unit002
Unit003
May/20/2009
Jun/10/2009
Jun/20/2009
Part A
Part B
Part C
1
2
Part A
Part A
Part A
May/10/2009
Part B
May/30/2009
Part C
Jun/10/2009
Purchase order
E0002
Part C
Unit003
Part B
Part C
Jun/10/2009
Jun/10/2009
Received and inventory
Part A
Unit003
0030
Unit001
Part B
Jun/1/2009
Part C
Unit003
May/10/2009
May/30/2009
E0003
E0001
Part C
May/10/2009
0020
Part A
Part B
Part B
0010
50
Unit001
Unit002
Part A
50
E0002
Quantity
Unit002
Agreed
delivery date
E0001
Serial
No.
Part A
Requested
delivery date
Part
Part
Unit002
Ordered
quantity
Mileage
Order
No.
Unit No.
Unit001
Arranged
date
Part A
Test
Serial
No.
Arranged
quantity
Part
Part
2
E0003
E0010
1
1
2
E0011
E0012
Serial
No.
E0001
May/10/2009
0010
E0002
May/10/2009
0010
Part A
E0003
May/10/2009
0010
Part B
E0010
Jun/1/2009
0020
Part B
E0011
Jun/1/2009
0020
Part B
E0012
Jun/1/2009
0020
Jun/10/2009
0030
Data model
296
Order
No.
Part A
Fig. 11
Received
date
Part A
Part C
Quantity
in terms of arranged quantity, received quantity, and

assembly results (Fig. 11). This enabled each department
to handle information in the units suited to the operations
performed by that department, and organically linking
these units of information enabled information spanning
the entire part lifecycle to be instantly understood.
links with a separate system that manages manufacturing

quality information enabled swift and accurate decisions
to be made regarding defective parts and whether to
exclude parts from the same manufacturing lot.
7. Confirmation of the Effects of System

Application
As of 2006, HRF1 was already collectively managing

racing sites and factories for non-engine parts, using the
part usage history management function and 2D barcodes
with an originally developed system based on
commercially available software known as the ERP
(Enterprise Resource Planning) Package. The starting
point for this system development project was the
finding that HRF1 was able to respond immediately to
inquiries from Tochigi regarding gear box parts, exhaust
pipes, and other parts.
Centralizing part lifecycle management required
wide-ranging investigations and reviews to optimize
overall operations. Approximately 3 years were needed
from the start of advance investigations in 2006, but the
know-how generously provided by HRF1, thorough
discussions by project members involved, and the clear
sharing of goals and targets across the Racing Division
enabled completion of the system.
This system was designed for direct application to both
Tochigi and HRD, and aimed for centralized management
of engine parts. This system was not applied directly to
HRF1, but preparations were underway to enable links
with chassis part and gear box part information, such as
a part number read-as function and measures to avoid
duplication of key management numbers, to enable future
links with existing HRF1 systems.
The indexes of faster plan feasibility verification,
increased purchase order speed, and enhanced part
management information are themes that are also
constantly pursued in the development of mass production
cars. Application of this system to the development of
production cars will be investigated in the future.
The decision to withdraw from Formula One was

made immediately prior to application of this system,
and the system has not been used for actual operations.
However, the system was still completed and integrated
tests were conducted. The results confirmed that each
department could carry out operations under an
environment in which all data related to part
manufacturing and supply was centrally managed, and
that timely and accurate decisions could be made, which
was the goal of development.
7.1. Faster Verification of Assembly Plan Feasibility
Formerly, the feasibility of the overall assembly plan
first became apparent when persons in each department
compared the information held separately by each
department. This made a process to investigate
countermeasures essential, and checking the location of
a quantity of specific parts for this sometimes took as
much as two days. However, this system enabled instant
identification and information sharing for the part which
impedes the assembly flow. In addition, the latest
information is centralized and can be easily referenced,
which is thought to enable swift decisions even when
coordinating between Japan and England.
7.2. Increased Purchase Order Speed
Accurate advance information enabled advance
preparations to proceed concurrently, which together
with enhancements to the drawing issue process,
shortened the time required from drawing issue to place
a purchase order by two days. In addition, centralization
of information can also be expected to prevent confusion
and duplicate work due to differences in information.
7.3. Enhanced Parts Management Information
Inventory control from the time parts are entered into
the inventory until disposal can be performed not only
in part number units, but also by individual parts
themselves, assembly, unit number, or quality rank. This
enabled an accurate understanding of inventory quantities
as well as measures for part excesses and shortages
according to the application. In addition, understanding
part replacement periods based on individual part usage
history information can also be expected to help
understand the part quantities required and achieve
proper inventory levels.
Assembly results records up to and including actual
work such as at a racing circuit can also be shared in
real time, simplifying the process of checking which of
the selected parts have actually been used. In addition,
8. Conclusion
Author
Hidetada TANIGUCHI
Katsunori TAGUCHI
Shigeo MIYAJIMA
Osamu WATANABE
297
Masataka NODA
Activity Reports
Looking Back on Assembly Activities

Racing Spirit
Tsuyoshi ISHIHARA*
Atsushi NORIHARA*
Naoto SUNAKO*
Motoaki ISHIZAKA*
ABSTRACT
Numerous technologies were developed and put to use in races during Hondas third-era Formula One activities.
This paper introduces assembly activities from development tests to the racetrack.
1. Introduction
Hondas third-era Formula One assembly activities
began with several goals: no dropping out of a race
because of human error; no supply delays; reflect
development issues in blueprints; and pass down
technological skills. Based on Hondas second-era
experience, these four elements were essential for race
management, and activities were always pursued with a
consciousness of these targets as the path to victory.
This article introduces these activities by dividing
them into those that are in the factory domain, taking
place at Tochigi R&D Center (HGT), which oversaw
development of new technical items and the assembly
and supply of race and test vehicles, and at Honda
Racing Development (HRD); and those in the circuit
domain, which supported race and running test sites.
2. Factory domain
2.1. Startup of Honda Formula One Third Era
The third era started in 1999 by organizing the team
strength so that primarily those members who had
supported the golden era of Honda Formula One in the
second era could pass down their technical skills to new
members, increase development speed and pursue supply
of equipment for races. Passing on everything to the new
members involved a lot more than just writing a manual
and making them read it; it was a lot of work!
This section first gives a simple explanation of how
technological skills relating to the work of the assembly
crew were handed down.
At the development stage, one repeatedly assembles
and dismantles previously developed single-cylinder
engines, demonstration bench engines, durability engines
and running test engines, after which one checks the new
specifications, looks for issues and feeds back results of

countermeasures to try to bring blueprints to fruition.
One assembles an engine built to the latest specifications
and then sends it to the racetrack. Engines used in races
are dismantled, malfunctions are found and then this
goes into the next cycle of activity. In this way, the
assembly domain is always involved, from the beginning
of development until race day.
Although one casually talks about engine assembly,
a Formula One engine is not simply put together using
as-is components that are in inventory. These are very
delicate engines: with the emphasis on performance, they
have only the minimal amount of durability and are
subject to approximately five times the engine vibration
as mass-produced engines, and the slightest error or even
one component quality defect can easily cause them to
break down.
To assemble such an engine, there must be no errors.
Components must be checked assiduously one-by-one so
that not even a fine burr is overlooked, and so that each
component can perform to its maximum extent. The
assembly crew is part of the final process, where one can
touch actual articles and check them with ones own
eyes, and quality must be protected here so that the team
can win the race. Similarly, during dismantling, one
visually senses malfunctions and feeds back information
to bring blueprints to fruition.
Assembly is furthermore required to supply
equipment for development and races with enough speed
to keep up. In fact, the authors cannot remember a single
season in Hondas Formula One activities in which there
was not a need for emergency component replacement
because of malfunction, or no repairs on engines
experiencing trouble. It is always a battle against time,
so one has to work with efficiency, speed and reliability.
To achieve this, the biggest issue is giving new

300
members the ability to spot malfunctions as well as

speed and work precision.
It is hard for new members to learn these things in
a short period of time, and there was a noticeably large
experience gap between them and Hondas second-era
Formula One members. The new members started with
their experience at HGT, got more experience from being
sent to HRD, and finally grew into track mechanics.
These members prepared on average 1000 engines
per year, including new assembly, rebuilds and
reassembly (Fig. 1), with the number of reported
troubles averaging 182 per year, thus contributing greatly
to bringing blueprints to fruition.
2.2. Development of Quality Enhancement Measures
During nine years of activities, the regulations have
changed several times to cut costs: the usage restrictions
on engines have changed so that the guaranteed usage
distance is 1 500 km, up from 300 km; and whereas there
were non-restrictive rules allowing practice, qualifying
and race engines to be changed, this was changed to
require one engine per race-event, and then one engine
per two race-events. The regulations have also changed
in ways that affect race strategy, with a penalty now
being assessed for changing engines. This has increased
the requirements for engine quality and durability.
Originally, our members thought that the old ways
of doing things were sufficient, but extending guaranteed
engine usage distances was a new frontier for our
Formula One activities up to that point, and also
represented a great opportunity to further enhance quality
and durability.
First, members strengthened quality checks of each
component. They anticipated the possibility of impurities
getting into places invisible from the outside, such as oil
passages, and then before assembly used an endoscope
to do a full check. For phenomena such as injuries on
the sleeve, it was decided to do a bench check prior to
shipping, then check piston and sleeve status and look
for impurities with an endoscope immediately before
shipping, and then supply equipment to the racetrack.
Besides this, the number of check items was increased
Motor
Function check engine
Engine for dyno
Engine for track
Engine use restriction regulation changes

Quantity
With no engine use restrictions
1000
Q R
1engine
1GP
1engine
2GP 1engine
2GP (Q, R) 1engine
TEST days restrictions
TEST days restrictions
1170
1141
981
124
229
1033
996
120
847
256
174
134
924
276
305
941
37
65
295
206
118
500
628
250
1003
547
514
246
312
351
04
05
06
457
583
289
03
315
437
657
424
383
304
07
08 - 12/5
00
01
02
2 teams supplied
Fig. 1
2 teams supplied
Record of complete engines and motors

supplied
for all assembly processes and countermeasures were

added to prevent troubles before they could occur, which
strengthened the members efforts to eliminate defects
that were their fault and prevent faulty components from
shipping. This further enhanced members awareness of
quality and made them more mindful of the engines they
themselves assembled.
During assembly, more malfunctions occurred even
as performance was being enhanced. Development had
proceeded on material substitutes to reduce friction of
the air-valve system movable seal, but air leaks during
assembly increased. Every time an air leak occurs, it is
a major issue in terms of supplying equipment for races
on a tight schedule: it takes time to repair, and the crew
has to wait for specifications adopted in response until
it is almost too late for the race. There were also
concerns that an engine might blow out during a race.
The team originally suspected a quality flaw in
components on the periphery of the air valve, but could
not come to a definite conclusion, so they performed
analysis starting at the point where the air leak occurred
and found that a minute amount of dust had gotten
caught on the movable seal face. Under specifications up
to that point, there was no impact even if dust got
caught, so material substitution was believed to be the
cause. Components were cleaned and then checked prior
to assembly, and assembly was done with great care, but
it did not solve the issue and small amounts of dust and
dirt still got caught.
As a countermeasure, impurities that came from
within components and which could not be removed by
ordinary cleaning were eliminated by implementing an
ultrasonic cleaning system. To eliminate airborne dust
and dust from work clothes, a clean room was set up and
workers wore anti-dust clothing over their work clothes
and anti-dust caps over their hair as they assembled airvalve systems. Movable seal faces, which can be
damaged by even a light abrasion, were put on with great
care, and the engine assembly process was nervewracking.
The result was complete prevention of air leaks and
the ability to supply more competitive engines.
2.3. Supply of Motors Made by Honda
In 2007, the FIA presented a regulation permitting
the use of hybrid systems in races starting in 2009, after
which it was decided that the motor would be built by
the assembly crew.
This was an area in which the assembly crew had no
experience, meaning they had to start from the very
beginning. Up to that point, the assembly crew had
overcome many obstacles and felt that it would surely
find a way, but there was much unease with the
transition from things one could tangibly observe to
something electric, which could not be seen at all.
The crews education began with the study of motors
for mass-produced vehicles. Getting advice from experts
on mass-produced motors, the crew began to master this
area.
Actually, even the stage of checking the building
301
process was of greater difficulty than imagined.

Structurally, the job of coiling wire is something that
only one worker can do on one motor. It is difficult to
divide up tasks as one can with an engine and thereby
shorten the production timeline, and when one is
planning to supply equipment for race use, one cannot
wait until just before the race to decide on the best
specifications. Additionally, if some trouble occurs, the
equipment would not be on time, and this increased the
sense of unease.
Being lightweight and compact while providing high
output are absolute conditions for a motor used in racing,
so winding technology was required offering a superhigh space factor. The issue of greatest priority was to
establish a method of winding a coil: specifically, how
to wind it so that it would be within the required size.
An examination and winding experiment were begun,
but Tries 1-3 could not meet the requirements at all (Fig.
2). The coil end was always too high. At this stage,
about 50 hours had been spent on the work. On Try 4,
a jig was devised and experimented with, but as it still
did not meet requirements, some other solution was
necessary. Other attempts were made, such as enhancing
the jig or changing the winding process, and on Try 5 a
breakthrough seemed to be in sight, but while progress
had been made, this still did not meet requirements. With
more innovations, the crew finally succeeded in meeting
requirements with Try 6. It had taken two months from
the start of examination until trial winding.
For the original winding, the crew thought to reduce
work time by implementing equipment and winding
mechanically, but eventually concluded that meeting the
requirements would necessitate a very complex winding
method, and as it was just not possible with a machine,
it had to be done manually.
Once the winding work was over, the coil ends were
molded to fit the coil in the case. At this point, a very
big problem came up. During molding, the wires rubbed
against each other and shorted out. After all the long
hours it took to wind a coil, in the end it was not usable.
The crew understood that the cause was the
relationship between the winding method and molding
load, but since the work was done manually, it was
Fig. 2
Try 1
impossible to make the way the wires come out match

perfectly, so a technique was selected that controlled
load during molding.
While the crew somehow managed to succeed, it took
many staff hours and there were still concerns about
being able to provide the motor on time. After examining
many ideas for reducing work time, about 30 types of
jigs were created.
Eventually, this effort succeeded in reducing winding
work time from about 50 hours to 15 hours, in part
because of overall skill enhancement of the workers.
As a result, the crews goals were in sight for
supplying equipment for durability testing, road testing,
and all races, and supplying system became highly
organized.
Production also got up and running, and the knowhow acquired during production for the 2009 winter
running tests was a very valuable asset. However, Honda
announced it was withdrawing from Formula One and,
unfortunately, this motor was never used in a race.
Motor development that had begun with a lot of
fumbling and no clear idea of how it would turn out
ended up being a source of great confidence and pride
for the assembly crew.
2.4. HRD, where Honda Engines were Born
To offset the risk of distance from the engine
development base HGT, HRD, the base for onsite engine
maintenance (the rebuilding team) and the running test
team during Hondas third-era Formula One activities,
was established to inherit functions from Hondas
second-era Formula One activities and serve as the front
line of support for each circuit in the EU. When first
launched, the site supplied engines during Japans long
holiday periods (May and August), and subsequently its
role expanded, such as making an initial determination
of durability and reliability following running tests of
new-specification engines, making emergency
replacements of rejected lot components, and having
materials analysis members onsite. This allowed them to
make more detailed analyses and assessments of the
situation when malfunctions occurred at the running test
site and thus greatly accelerate development.
Additionally, when supplying engines to two teams the
second time, the scope of the crews responsibility grew
to include all engines for the Super Aguri Formula One
Team (SAF) as well as engines for the Honda Racing
Formula One Team (HRF) (50% of running engines
supplied). The increasing number of engines supplied,
along with the need to deal with malfunctions, made it
necessary to strengthen and expand the assembly system.
Along with this, activities were also transformed by
having one person in residence and persons dispatched
from Japan taking up long-term residence so that it
became their regular work, an arrangement that
supported race activities. Engines prepared by the rebuild
team were handed over to the site mechanics with the
sense of saying, Please take care of this engine, which,
though the words were not actually spoken, bound both
sides in a relationship of trust. Every member of the
302
rebuild team had a sense of pride and self-confidence

that the engines they were shipping out were perfect. The
engine rebuild specialists dispatched from Japan lived
three or four in a home close to HRD during their fiveor six-month stay in the UK. In addition, this provided
a venue for new members to learn the mindset and spirit
of accepting challenges in their day-to-day lives as well
as in races, because these people were sharing a home
with the veterans several times a year, and it provided
the sole place where they could relax while in the UK.
It was HRD that prepared the Honda engine that enabled
Jenson Button to win the Hungarian Grand Prix in 2006.
2.5. Collaboration with Local Members
There were changes to the system in the rebuild field
from Hondas second-era to third-era Formula One
activities. In Hondas second-era Formula One activities,
only persons dispatched from the R&D Center conducted
the work, and they were able to work the same way
where they were dispatched as in the R&D Center,
almost as if the site were an extension of the R&D
Center. In the third era, a number of local staff members
worked side-by-side with members dispatched from HGT
from the time this arrangement was established. The
culture and language were different, of course, and the
authors remember how difficult it seemed to stay in step
with each other. But while fusing our different cultures
and launching a system to build equipment together were
big challenges for us in conducting race activities in the
UK, there was also a great sense of eagerness to do this
as a new experiment in Formula One activities.
The local staff members who were hired were typical
race engine builders who had only assembled parts out
of inventory, and so had no experience with engine
development to deal with the development issues from
the quality side. By contrast, Japanese crew members
acted with a constant sense of responsibility as the final
check-persons in every domain (inspection, processing,
materials). They also were aware at all times that this
place was an R&D center, not an assembly plant, and
as such it was up to them to guarantee the quality of the
finished engines and it would be the assembly crews
fault when defective components went into their
assemblies. When this happened, there were always
discussions about why the defective component was not
noticed beforehand.
The veterans said, Get development information
from the R&D Center and local information about the
circuit, but in any case look closely at the actual engines
and components in front of you. There was never any
wavering from this principle, which was a tradition
handed down from Hondas second-era Formula One
activities. Local staff members were partnered with
persons from Japan who instructed and guided them to
help them develop this way of doing things, so that after
several seasons had passed, the work could finally
proceed smoothly. Local members were given roles with
responsibility and thereby learned a sense of
responsibility, so that the two sides came to understand
each other better and both could absorb new knowledge
from the other.

In 2006, Honda began supplying engines to two
teams for the second time, and it was decided that all
SAF engines would come from HRD. Until then, HRD
had been in charge of preparing Friday practice engines
(P) and race spare engines (RS), but now it would also
be in charge of race engines (R), and this would greatly
transform the arrangement in which HGT assembled and
supplied the R.
Starting that year, all engines from HRD had Made
in UK stamped on them. This boosted the motivation
of local members and was taken as a sure sign of their
independence. The authors recall how this established a
solid relationship of trust between the two sides in
Hondas Formula One endeavors.
2.6. The Unknown Story of How the Winning Engine
was Prepared
In the latter half of July 2006, to make up work
during the summer holidays at HGT, assembly was
planned and carried out at HRD on two engines for the
Hungarian Grand Prix of August 6-8. Preparation of one
engine proceeded smoothly and was completed on
schedule, whereas the other was plagued by troubles.
During assembly, air started leaking from the valve-train
system air-valve seal face and would not stop. Studies
were done on the situations deemed likely to cause this
(damage to the seal face, the interposition of impurities,
deformity and form defects, hardness, etc.), but nothing
could be identified. Repairs were made with a great deal
of work time spent, for example switching out
components from different lots, and the two engines
were finally done with little time to spare. A plan was
organized to assemble and supply Rd13 engines, a total
of three engines including the one assembled at HGT.
The HGT engine, with slightly greater output, was
mounted on Buttons machine, while one of the engines
assembled by HRD went on Rubens Gonalves
Barrichellos machine. As for the remaining engine
assembled by HRD (the one that had an air leak during
the assembly work), it was to be used as the race spare
to the disappointment of the local crew, but then the
Hungarian Grand Prix opened and Buttons engine blew
during practice on the second day. The spare engine
from HRD was quickly mounted on his machine.
Originally slated to start fourth in the qualifying, Button
was given a 10-slot penalty for change of engine and had
to start in 14th place and would have to work his way
up from there. While this would have been fatal under
dry conditions at the Hungarian, which makes overtaking
difficult, the race day began under wet conditions.
Vehicle speed, weather changes, race strategy and
veterans dropping out were among factors that let
Button, starting in 14th place, gradually raise his position
and ultimately finish the race far ahead of the other
vehicles.
At that moment, Honda had won its first third-era
Formula One victory. For HRD, winning the race with
an engine they had built with one British member and
one Japanese member was a great confidence-builder.
303
3. Mechanical Domain
3.1. Supplying Equipment to Two Teams
From the start of Hondas third-era Formula One
activities until Honda began supplying more engines to
two teams in 2001, the number of staff members
increased and, as a result, Honda started supplying two
teams with a staff consisting of 13% experienced persons
and the rest not experienced. For the mechanics, of
course, inexperienced personnel had to be used, but
under the guidance of experienced staff from Hondas
second-era Formula One activities, these people received
practical training in on-site trouble-analysis techniques.
Experience taught us that the ability to evaluate
sudden troubles cannot be learned in a very short time,
so all sorts of means were considered and put into effect
to help the inexperienced staff make up the gap between
themselves and the experienced staff. An attempt was
made to enhance their knowledge by, for example,
reading up on past troubles, conducting simulations of
trouble with an actual vehicle in front of them, and
having groups of young mechanics practicing FMEA
amongst themselves. Sometimes, however, they became
overwhelmed by the atmosphere of the site and the sense
of urgency and were unable to act.
We felt that the ability to judge and act at a
moments notice in extreme circumstances could not be
acquired unless we always had an engine in front of us
and could put our hands on it.
Even as their knowledge and competency increased,
it did not help at the racetrack unless they had tools,
spare parts (including, for example, proven replacement
components, in cases where a malfunction occurred with
a newly introduced component) and so on.
The veterans often said, Preparation occupies 90%
of work.
The most essential thing needed was the practice of
keeping the tools, equipment and components needed at
the site in a neat and orderly way so that they could
quickly be found when in a hurry and nothing would be
missing. This is all very basic. But for the young
mechanics, this was where the work began.
The following relates what happened at one event.
An engine that had been running smoothly suddenly
blew out without warning. The blown engine hurriedly
underwent emergency dismantling by two of us, a
veteran and myself (Ishihara); when we did primary
analysis of the cause, the veteran asked me to pass him
a battery-powered impact wrench, but when I picked it
up, the battery was dead.
This led to a severe scolding from the veteran. I
realized that my negligence in preparation had come to
light in that instant. After that incident, the first thing
we did at the site was to keep a battery charger ready.
The engines supplied to the race track are supposed
to have undergone a very strenuous final check at HGT
or HRD, but all engines underwent another enginespecification check at the hands of the mechanics, with
each bolt checked for torque, harness wiring checked,
etc., time and time again.
Oil-pressure adjustments and air-intake trumpetlength adjustments were made in keeping with the usage
environment at the site.
From time to time, components were replaced to deal
with troubles in time for an event and the crews
constantly tried to extract the maximum performance,
durability and reliability in each race.
On the starting grid of the final round, a number of
final quality checks on the hardware side of the engine
were performed right up to the start of the formation lap,
including checks of whether oil, water, fuel and air
pressure, air and oil consumption and oil and water
temperature were acceptable, whether there were any
leaks, whether all cylinders were firing and whether the
air/fuel ratio was correct.
This is very hard work, as malfunctions have to be
responded to immediately when found.
Sometimes the members were so nervous that their
stomachs hurt and they could not eat until the signal
changed and the vehicle got off to a trouble-free start.
More than actual race results, the mission of the
assembly crew was to make sure the vehicle finished
without trouble: for example, during the race, members
read data from the telemetry system and prepared for an
emergency pit stop if any abnormalities occurred.
3.2. Kitching
On the job at the racetrack, it was necessary to make
sure that all mechanics were correctly producing the
same output. This is referred to as kintaroame, which
is a kind of candy formed with an image in it. The
association implies that it was vital, no matter who cuts
it open and where they cut it, that the result will be the
same.
The two race engines provided for the two drivers
were lined up and arranged in detail right down to the
thickness, position, direction and binding force of the
wire harness and the tie-wrap that anchors the auxiliary
components (Fig. 3).
Deciding on assembly work procedures and
specifications when there are no blueprint instructions is,
in large part, dependent on the experience of the
mechanics.
This is because the heat, vibration and stress on
304
Fig. 3
Completed engine
components are so much more severe than is the case

with mass-produced vehicles. Particularly during the
summer races when it was so hot, the need to take
countermeasures arose often, and when malfunctions
occurred, there was a series of revisions of work
procedures and specifications in keeping with changes
in the environment so that crews on the site could
determine the causes of the malfunctions and make the
correct judgment on how to repair them.
Most particularly, with work procedures and
specifications for engines built to new specifications, the
crew, racking their brains, reflected procedures taken
against past troubles, then verified these procedures in
shake-downs and running tests.
Work procedures and specifications that were found
to be reliable were then compiled in the form of a
manual, including photos taken with a digital camera to
illustrate things that are difficult to explain with text
alone.
Surely, this was a case of seeing is believing.
The crew furthermore created a checklist so that
nothing would escape the attention of all the personnel.
As the years went by, the check items became
subdivided, so that by 2008 there were 120 items.
This final check process before shipping equipment
out is referred to as kitching, a word that is completely
unique to the Honda Formula One world. This word goes
all the way back to the era when the company
participated in Formula Two racing in the 1980s. At the
time, some veterans did not have enough time before a
race to send an engine back to the R&D Center, so they
did an emergency rebuild in the kitchen of the local
company housing unit in England before sending it on
to the race track, and the process was subsequently
named after the rebuild site.
To be a mechanic, first of all one had to be able to
do kitching correctly, and this became the gateway to
success in becoming a mechanic.
During the third-era Honda Formula One activities,
new mechanics were strongly advised by the veterans:
As a mechanic, you are carrying a big Honda sign on
your back. As a representative of the assembly process,
act in ways that will not make us ashamed. Those
words were bracing whenever I heard them. I felt as if
I were an Olympic athlete wearing the rising sun of
Japan on my back.
Go for it! You can do it! Well be rooting for you
on TV!
As the vehicle set out from HRD to the circuit, all
members came out to send it off. This was always a
happy moment.
Formula One is very much a part of European culture
and an event of its noble society. We took on this
technical battle as representatives of Honda from Japan
all the way in the Far East, with a different language and
culture, and as representatives of Japan itself.
3.3. Onsite Engine Dismantling
During the engine startup check on preparation day
during the 2005 Canadian Grand Prix, a lot of air erupted
from the oil tank freezer. Both the P1 pressure (gas

cylinder pressure) and the P2 pressure (engine supply
pressure after regulator pressure adjustment) started
falling and would not stop. When the P2 sub-line was
removed to identify the location of the leak, the leaking
stopped.
The leak was from the engine itself.
With no time to think, the place erupted into an
uproar, with some people rushing to change out the
engine to be in time for the vehicle inspection, and
others acting decisively to hurry and open the cam cover
to analyze the cause.
What the engine showed when it was dismantled was
that a valve cotter had become embedded in a pneumatic
valve return system (PVRS) piston and the seal ring was
protruding very slightly beyond the chamber, causing the
air leak. Results of analysis confirmed a malfunction in
the taper angle of the cotter contact face of the PVRS
piston, creating a looser taper angle than usual so that
the cotter went deeper.
In this state, it was impossible to determine whether
the trouble was caused by a quality defect.
As such, all the crew could do was check the interior
of all five remaining engines. This would take as much
as five hours per engine, and everyone was involved,
including those who had come to hand over their work
to substitute members.
Ordinarily, if one were to open the cover and remove
everything down to the cam, the conventional wisdom
would be to run the engine on a bench and check to see
if there are any oil leaks or other abnormalities. But this
kind of work cannot be done at the racetrack, and neither
can the crew afford any errors. Nonetheless, they
finished in time for the qualifying round, during which
the team earned its first pole position in the season,
giving the crew a taste of having accomplished
something great. With the extreme tension lifted from
their shoulders, the team mechanics shook our hands and
congratulated us with a Good job! Good job!
Naturally, it would be better if there were no
problems, but this experience in particular helped
develop fearless race mechanics and was also reflected
in HGTs engine development blueprints, which
enhanced durability and reliability. Responding to
phenomena as they occur at the actual site, checking
actual articles and getting to know the actual facts are
essential to preventing the same trouble from happening
again; this is the practice we call the three-reality
principles in Honda. The crew relayed to Japan what
they were thinking at the actual site and what was
necessary, and passed that information on to Japan.
We understood the significance of mechanics
because, as in this case, if there is no word from the
actual site, nothing will change.
3.4. Adding an Air Pump
An air-pump system to supply air valves was added
in 2003 as a way of reducing friction in the valve-train
system, and this system was a source of many troubles
for site mechanics.
305
The team requested that the volume of the air

cylinder mounted on the vehicle be reduced. A mounted
air cylinder would certainly seem to have an impact on
aerodynamics, and making it any larger would be
challenging.
An air pump was the way to solve the problem in
order to continue supplying a limitless amount of air,
which is consumed in such great volume.
It was a wonderful idea, as it would also free the
mechanics from having to fill the cylinder with air
onsite. However, at the start of development much more
work was awaiting the crew than they had experienced
up to that point. There was no way it was going to
become easier.
The crew mounted a small compression pump on the
engine and used engine oil to lubricate it. It looked just
as if a radio-controlled cars engine were attached.
At first during development, it was one malfunction
after another, and there was not a single day when it ran
properly and without any trouble.
The air pressure seems off. Hows the regulator
doing? Is the pump working? How about the hose? The
work consisted of a lot of urgent evaluations, because
if the air supply were cut off, the engine would soon
blow.
Simultaneously with air being supplied from the
pump, the engine oil used as lubricant forms oil sludge,
blocking the air channel. The filter intended to remove
this became clogged from the great volume of oil sludge,
so the air channel in the engine front cover also got
blocked.
For the front cover, the crew attached a brush to an
electric drill to clean out the air channel, but this failed
to remove the blockage, so carburetor cleaner was blown
in from the air-pump intake for 10 seconds and left to
sit for five minutes before one minute of firing, a process
that was then repeated a second time. This was done
after both morning and afternoon runs. Components that
had to be dismantled, moreover, were cleaned inside the
trailer. For secrecy, this could not be done in the garage.
The process of cleaning out the acrid, black, slimy
sludge lasted late into the night.
During initial development, filters had to be changed
and channels cleaned out, and there was no way to avoid
stopping the vehicle so much that it impacted not only
the race distance but even vehicle testing.
We still cannot forget the team leaders gathering
around and peering in, then giving us a cold British
stare.
It reminded us that as mechanics, we are the front
line and must bear the brunt of any troubles. As the
engineers worked in the trailer each night under pressure
to organize running data, they were close to that strong
sludge odor, but nobody said, Stop it, it smells terrible!
Everybody was aware of the situation and they put up
with it together.
Naturally, once we all got the idea that weve got
to do something about this, it had a positive effect on
later enhancements. Measures taken included reducing
the temperature of the air blown in to a point where oil
sludge was less likely to form, and also managing the

clearance between pistons and cylinders as a measure
against oil outflow.
3.5. Adding a Quick-Shift Gearbox
In 2003, the first thing the development supervisor
said was, Lets build the worlds greatest gearbox,
which was the start of a project to reduce lap times by
developing an original new mechanism called a quick
shift, which theoretically had no time lag during shifting.
The technology of the system was later transferred to
BAR, initiating the project to put it to use in races.
The development team and we were braced to do
this, but at the time the basis for achieving such
technology had not been established. Those of us who
had specialized in developing and supplying Formula
One engines were also dealing with gearbox
development for the first time, and with no experience
in these matters, we were starting from zero. It was also
a turning point, after which the scope of our
responsibilities expanded.
The British American Racing (BAR) team leadership
expressed disapproval at the start of development, as
they questioned what role Honda would be able to play
in joint development of a gearbox and suggested they
had nothing to learn from us. There was a reason for this
attitude: in the year before the command applied, we
received a chilly response when a malfunction developed
on the driving site in a narrow-shift gear mechanism that
had been suggested by HGT. Taking this experience into
account, we responded this time with greater passion and
speed.
A quick-shift system allows double engagement of
gears: the gear being driven and the next gear to which
the driver is changing. Because of this, during shifting
there is no torque drop. The following year the crew
proceeded to actual driving evaluation, when BAR,
which got to know this mechanism on actual vehicles,
had completely changed their attitude. Many engineers
came to look at how the shift mechanism moved in the
first equipment.
All of them said, Thats crazy, and left. At the
initial running test, after a pit stop, we saw engineers and
mechanics coming by, even climbing up the course-side
fence behind the pit, to hear the sound of gears changing.
In reality, however, a series of severe development
issues occurred. The crew tried all the countermeasures
they could onsite on both the control soft side and hard
side (making manual modifications to change the form),
but each time the same phenomenon occurred at the
same travel distance (35 km).
Although the system had passed the bench tests, there
was now a baffling series of phenomena. From this
situation, the crew proceeded with the priority of
enhancing durability and reliability, with HRD rearranging
components that HGT had sent out, but every time the
logistics took until just before test time. Everyone worked
late into the night to conduct stopgap solutions and change
out the gearbox at the running test site as we endeavored
to somehow stay within the schedule.
306
Gradually we could see the replacement parts taking

effect, and with the start of long runs, drivers began
making such positive comments as:
Shifting was seamless, with no sense of reaching a
plateau. I could even shift up while cornering and the
rear did not show any turbulence, and each time I shifted
up it shrank the distance between me and the car in
front. The new system was proven to reduce lap times
by 0.2 sec/lap compared to a standard gearbox.
Unfortunately, during the 2004 season the system was
unable to meet the 500 km durability requirement, so it
was introduced the following year, in the opening race
of 2005, when it became the first such technology used
in a race. And with that, the worlds greatest gearbox
was complete. At the circuits, other teams refer to this
as seamless shifting.
In any case, after accepting all the facts, the assembly
team had accomplished a transfer of work responsibilities
to BAR, from creating systems and environments for
supplying test equipment to running race operations.
3.6. Adding a J Valve
The J valve system added in 2005 was developed to
reduce valve-train system friction.
In the structure used up until that time, one-way
valves were placed in the air chambers that work as
valve springs at each of the various entries/exits, so that
even if in the worst case the air supply died, the valve
would not fall from its own weight. These one-way
valves were eliminated, however, and each air chamber
was given an orifice to create a continuous passage, so
that if air leaked from one air chamber, it would interfere
with the valves of all cylinders. The crew learned that
the engine quality could not be assured with this new
system as in the past.
The engine assembly method had also changed, but
above all else the mechanics put their energy into
transportation methods and methods of operation from
onsite storage to engine startup, and most especially on
a technique of lubricating the valve stem with oil.
During engine startup there was concern about the
valve stem getting burned, and what caused the most
hardship of all was how to control the amount of oil
inside the engine, which the crew could not see.
During operation, the system banged against the wall
and we were not at all sure it would be OK.
What is the minimum pressure needed to keep valves from
falling from their own weight during transportation?
What is the minimum pressure for valves to track
movement of the camshaft when the crank is being turned
by hand?
What are the temperature and pressure that would make
air leaks unlikely based on characteristics of the PVRS
piston seal?
Can oil be purged from air lines if collected and retained
in air chambers?
What capacity is necessary in large, high-pressure air
cylinders?
Can the engine be shipped by airplane with a large, highpressure air cylinder attached?
How can one check the initial lubrication of valve stems?

These worries were dispelled by testing each item
one-by-one.
In all cases, while engines were being assembled at
HGT, sent to the circuit and returned to HGT again, we
constantly attended to the engines and collected data on
them and finally determined the requirements.
We proceeded with particular caution when the
technology made its racing debut in 2005, and the J
valve proved it had reached completion on the race track
operations side, and it later became the standard.
We continued to collect data after that, to the point
where, by 2008, we had successfully simplified the
technology on the equipment and operations side in large
part.
The technology could then be normally used on an
engine without any of the anxiety as in the early days
of development. The supervisor of J valve development
at the time said something I never forgot:
Whether the J valve will be successful or not all
depends on how its used on the race track. I expect a
lot, especially of the mechanics.
3.7. Extending Engine Mileage
Previously it was possible to change engines on
Friday, Saturday and Sunday, but regulations starting in
2004 required that engine mileage be extended and that
only one engine be used per race. The mileage
requirement was further extended in 2005, so that a
single engine had to be used in two races, starting with
practice. The distance each engine was used went from
300 km to 800 km to 1500 km. Instead of simply
extending the mileage of an engine with the same
specifications, the engine framework specifications
should be altered in the course of steady development
to achieve both durability and reliability on the one hand,
and performance on the other.
The penalties for changing engines were severe: if the
engine was changed because of trouble occurring within
the required usage distance, the driver would be
penalized 10 slots if it were before the qualifying, and
if it were changed after the qualifying, the driver would
have to start in last place.
Before the penalties were imposed, if trouble
occurred, the team could weigh the time it would take
to analyze and repair the defect against the time it would
take to change the engine, and choose whichever was
faster.
If done quickly, an engine can be changed in 30
minutes.
Of course, the pit was set up to deal with such
situations. Nearly all the components concerned with the
engine, including radiators, oil coolers, oil tanks, exhaust
pipes, clutches and hydraulics systems, were installed in
the pit.
At each circuit, there were always practice engines,
qualifying engines and race engines for two drivers, as
well as three spare engines and one engine for a training
car. This meant that there were at least 10 engines in
the garage, and of course not all of them could fit. On
307
older circuits in particular, garage space is limited and

it is a very challenging work environment just to load
and unload engines from their special transportation
boxes.
In a complete turnabout, the change of regulations
eliminated the work of changing engines every day as
well as reducing the total number of engines, so
everyone thought there would be less work to do.
Everyone was wrong.
Now the crew had to check the inside of the cylinders
each day after the engines run, instead of only when an
abnormality occurred. A fiberscope was used to detect
any piston scuffing inside the cylinder, which was
determined by the appearance of the sleeve (Fig. 4).
Another important check item was to detect any
pitting in the spline of the input shaft that couples the
crank tail-end to the clutch housing.
If any pitting were discovered on the spline on the
crank side, it was polished away. If this is overlooked
or ignored, it can start a process that results in the crank
breaking and metal burning as the engine fails to
maintain oil pressure.
Removing pitting marks with a compound and
diamond rasp is a difficult job because the crank is so
hard. The question of whether an engine can be saved
is a heavy burden on the mechanic in charge.
Flushing with oil was also done after each event. The
objective is to get rid of any degraded oil still clinging
to the conrod pin metal and main journal metal. This job
is essential after each race, because experience has
shown that engines with stagnant oil in them for an
extended time will occasionally suffer metal burning,
either in bench tests or on the circuit.
It was also unavoidable that components would have
to be changed that did not assure sufficient durability.
The crew had to use every possible means to meet
the high hurdle of 1500 km.
The engine truly was a lot of work.
At the circuit, although ideally there should be no
more of these jobs to do, they became a regular feature.
Maintenance of an engine to keep it in continued use
depends in many ways on the mechanics competence
and experience, and if just one thing is overlooked, it
Fig. 4
Checking inside of engine
dooms the vehicle to withdrawal from a race.

The assembly crew discussed this a lot with the
design, research and electrical staff. Looking back now,
it seems not so much a discussion as a pathetic plea from
us to them to Please fix this engine right away, and
please get it out to the track so we can use it with
confidence. A major part of this plea was the fact that
each race was contested with a special kind of pressure:
the drivers life was in our hands.
Even aside from the need for continuous
maintenance, the requirement of two races for each
engine put great pressure on mechanics in other ways.
It was no longer simple to decide whether or not to
change the engine, and now the mechanics had to
analyze malfunctions and do repair work with greater
accuracy and speed than ever before.
Accomplishing this created a need for tools with
greater functionality and a wider variety of components
than before.
The amount of equipment and components brought
onsite increased to the point where the crew could
assemble an entire engine from them.
Of course, there were also enhancements in
maintainability and work efficiency.
The veterans were fond of telling us, The race does
not wait; this is a battle with time, and in fact we were
often forced into such situations.
In a sense, the changes in regulations created a
challenge to the mechanics to provide quality assurance
of the engines.
3.8. Launch of SAF
At the end of 2005 it was decided to start supplying
engines to SAF. Five members of the first crew in charge
of this area entered the UK on January 12 of the
following year. At the SAF factory in which they settled,
the rising sun flag of Japan was a big landmark in the
rural area of Leafield west of Oxford, England.
This day was the start of a big challenge.
When the crew arrived, SAF had fewer than 20
employees and no vehicles, equipment or components.
That was the first impression: they dont have anything.
This was quite different from supplying engines to an
existing, already organized team, so the crew had to do
double duty by also launching a team.
Their first goal was to get two vehicles to the final
round grid in the first race of the season. As the crew
gave instructions, everything had to start from the very
beginning: preparing tools and equipment spare parts;
changing the engine auxiliary components (replacing
BAR specifications with SAFs); integrating engine
operating rules (post-arrival, return shipping methods,
warm up, oil management, pre-mounting status, etc.);
engine startup and running procedures; distancecontrolled components; and so on.
Everything started with dialogue.
Having set foot here, the first crew continued
working patiently, keeping its feelings under control.
On January 17, the initial vehicle (the base was an
old Arrows A23 built to 2002 specifications, which had
308
been out of the Formula One world for an unusual four

years) was brought in, after which the team modified the
monocoque to allow Hondas engine to be mounted and
modified the chassis to enable it to pass the FIAs
collision requirements. Then, equipment was introduced
and physical production began.
Ten mechanics were also hired, so that on January
30, vehicle assembly began. Everyone worked on
production, which was all done by hand. There were no
differences among members, and even the chief designer
and technical director were put to work. These people
responded well, and we had no troubles in terms of
communication.
Everybody was helping each other with their roles
and responsibilities, without anyone getting in each
others way. It was all positive and forward-looking, and
with everyone cooperating, there was a strong sense of
teamwork.
Day 32; Shake-down of initial vehicle (SA05-05)
Day 39; Barcelona joint testing (SA05-05): time was
not an issue here, but the vehicle ran 590 km.
Day 58; First race, practice stage, two vehicles began
running
Day 59; Got through qualifying stage with 20th
(SAT) and 21st place (IDE)
Day 60; Two vehicles line up in the final round grid
Having met the first target must have given the crew
a sense of accomplishment, but they were actually quite
cool about it. It seemed as if they took it for granted that
we would be in time for the opening race. Those two
months seemed to go by very slowly, but this was just
the starting line of the season.
Although the team did not bring home any trophies,
it attracted increasing attention. With each race, the look
on team members faces became sterner. There should
be no failures and no delays, so their expressions
changed.
As the races went by, we ran out of spare parts and
team vehicles had to retire more often, emphasizing the
fact that we were not making progress on the vehicle.
Starting from Rd12, we added the Honda gearbox
(quick-shift) and gave technical instruction.
Along with that, we changed the suspension and the
vehicles advanced significantly, which enhanced vehicle
performance and provided valuable racetrack experience.
Prior to qualifying rounds, gear ratios were changed on
three vehicles, including the training car. This work
concerned a gearbox with a new mechanism, and the
SAF mechanics did not have any onsite experience
changing the new mechanism. Two hours before the
qualifying, the three vehicles were lined up for
presentation, and the team had acquired its first
experience working in an actual race. We recognized
again the importance of maintainability and reflecting
enhancements. From there, we began special training to
give the team mechanics more independence.
In 2006, SAF was born as a new team and competed
in all Formula One races.
Everyone who set foot in Leafield that year had a
part in this dream.
3.9. KERS Development

Even the Formula One world is expected to care
about the environment, and it was decided to change the
regulations greatly in 2009.
That is when the kinetic energy recovery system
(KERS) was introduced. It became a matter of some
urgency to develop a hybrid system, a Formula One
version of the INSIGHT, as it were. The reason hybrid
cars are considered so good for the environment is
because of their good fuel economy. And in a race, of
course, competitive fuel economy is a huge advantage
strategically.
The regulations, furthermore, allowed output of up to
60 kW of electric power charged from the motor during
deceleration, with engine assist of up to 400 kJ of energy
released per lap; functionalities like increasing maximum
speed and enhancing acceleration coming out of corners
were expected to reduce lap times. This created a need
to develop a lightweight, compact and efficient system,
and Honda did all the development of the motor, PCU
and battery in-house. Having this development done inhouse was of particular importance to the mechanics as
they headed out to the racetrack.
They would absolutely need a thorough familiarity
with this new system in order to gain the knowledge and
technology to be able to safely and flexibly deal with
malfunctions on the circuit.
If one were to use a purchased item, there would be
no way to know what is inside. In other words, a
purchased item is like a black box, and there is nothing
more to do with it.
In contrast, the crew had no excuse for not
understanding something developed at HGT. Fortunately,
the motor was built by the assembly department.
At first glance, no one looking at this motor
measuring a mere 105 mm x 100 mm would think that
it creates even as much output as one FIT motor.
One can easily imagine that developing this motor
took much more than the average effort.
Everyone on the crew started as a novice in terms
of motor assembly.
It also required a very complex way of winding
completely different from a mass-produced motor, and
was only completed with a great deal of trial and error.
UN rules on handling batteries proved to be a great
obstacle.
Using a battery lithium ion, individual cells were
combined into modules, which were then combined into
packs. The mechanics thought it would be the packs that
they received at the site.
But what was sent from Japan to HRF in England
was a large number of individual cells and a welding
tool to join them together.
It takes at least three months to get permission under
UN rules to ship cell modules and packs, and a new
application is necessary each time specifications change.
This would be fatal in the Formula One world, which
has very short development spans, and we had never felt
how far Japan was from England until that point.
Our goal, nonetheless, was to never give up, and to
309
be the first to do this task successfully, and the only way

to do that was to have the mechanics do the welding in
England. Still, the welding effort had to start from zero.
A mechanic had to learn the special welding
technique in a short time and then bring it to England,
where the welding was done in a corner of the HRF
warehouse. As we struggled with this we paid close
attention to safety.
Each cell has a slight charge to stabilize it, and there
was some concern that during assembly, an error could
electrocute the person. Of course, the people doing the
welding first took a course in electrical safety, and even
went to Hondas Suzuka Factory, where hybrid cars are
mass-produced, to observe there. To be honest, we
thought that electricity was so frightening because it is
invisible to the eye, but as our knowledge grew, we
mastered ways of protecting ourselves from danger.
The mechanics undertook to identify 187
malfunctions and take care of pending items that could
occur by the time the motor, PCU and battery were
mounted to the vehicle.
Not only were there assembly malfunctions from
parts touching each other, but assembly work procedures
and specifications were decided where there had been no
blueprints or instructions relating to reliability and work
characteristics, from putting on the harness to assembly
procedures.
To ensure safe operation at the site, everyone
discussed what was safe and what was dangerous, based
on system diagrams. The team checked each work
process, starting with the preparations prior to engine
startup and continuing through post-cruise dismantling
(Fig. 5).
As this was going on, the Internet started circulating
news and photos of a mechanic with BMW Sauber
getting a shock on a circuit during a test driving. It
shook up everyone and showed that there was disaster
lurking in unexpected places.
Ultimately, it reminded us that, to protect ourselves,
each person had to be aware of his own safety.
The result was mechanics with a wide body of
knowledge and skills and an awareness of safety in areas
completely unlike the past.
Fig. 5
Checking KERS system
In April 2008, actual driving development began with

a shake-down test at Santa Pod in the UK. It was hard
going, since there was a series of issues we had never
experienced before, but with this event the crew
succeeded with a KERS road trial for the first time in
Formula One.
There were subsequently two more shake-downs and
a driving on the Jerez circuit in Spain, and in November
the crew was back in Santa Pod for the fourth
development shake-down, and during these events we
overcame the various issues.
We faced phenomena such as being unable to precharge or unable to connect the main conductor, the
cause of which turned out to be a broken contact point
melting damage prevention circuit when connecting the
main conductor incorporated into the high-voltage circuit
in the battery. Because the battery of the third
development lot was stored under the drivers seat, we
tried connecting the main conductor manually from the
outside by having the driver step out of the monocoque
and removing the seat each time the engine was started.
This connection procedure was complicated and required
much time.
In Formula One development, which constantly
requires more speed, no matter how tiny the part may
appear, this was indeed a serious trouble.
Moreover, the effect of electrical noise on the
communications system prevented the technology from
providing full output, and there was nothing more that
onsite analysis could do in this state. To give priority
to analysis of the electrical noise, the team canceled their
participation in the next Paul Ricard Test, and the
vehicle was hurriedly sent from England to HGT. Of
course, the mechanics also went home with regret in
their hearts.
At Spains Jerez circuit the vehicle made a new start,
the drivers evaluation stating, I still have a lot to learn
about KERS, but I was impressed by how amazingly fast
I approached the hairpin when the assist kicked in. It
makes non-assisted acceleration from the engine alone
feel as if we are carrying a heavy weight.
The fourth development lot system, which was more
lightweight and compact, also suffered frequent issues
with communications line noise, and the mechanics
providing support at HGT took a number of
countermeasures, like adding a condenser to the PCU,
adding ferrite and enhancing locations with defective
grounding.
It was the first time for assembly team members to
touch the PCU. In the meantime, HRF actually
performed more than 100 engine startups and put their
energy into identifying the source of the electrical noise.
The PCU on which measures had been taken reached the
site just barely in time. The electrical noise had been
eliminated.
Then came the December 5, 2008 surprise
announcement that Honda was pulling out of Formula
One.
That marked the end of a solitary 12-month battle the
mechanics had waged in anticipation of a new era.
310
4. Afterword
Looking back on nine years of activities, Hondas
third era of Formula One started with the development
and supply of engines. The effort subsequently responded
to diverse needs, from meeting development
requirements to developing a gearbox, conducting
vehicle running tests, developing KERS and, moreover,
launching a private team. Although Hondas withdrawal
from Formula One ended an era and we were never able
to ascertain the competitive strength of the 2009 vehicle,
we believe that the racing spirit cultivated here will one
day shine again in a new age of racing.
While we did not unfortunately leave a great record
of race success, the guidance given by our veterans
helped to pass down their mindset. We are moreover
thankful for the cooperation of all departments concerned
with development that helped us to carry out our duties.
Author
Tsuyoshi ISHIHARA
Motoaki ISHIZAKA
311
Naoto SUNAKO
Atsushi NORIHARA
Yusuke HASEGAWA*
Shiro HISATSUNE**
Tatsuya KODAMA**
ABSTRACT
In conjunction with the launch of our Formula One project in 1998, Honda Racing Development (HRD) was
established as the British base for administering the Honda Formula One team. Subsequently, plans were modified to
supply engines and joint development efforts with the BAR Team, and HRD came to assume the role of Hondas
frontline European base for timely project and decision-making. This article describes the roles and accomplishments
of HRD.
1. The Roles of HRD

All engine development was formerly carried out at
Honda R&D Co., Ltd. Tochigi R&D Center (HGT),
including design, assembly, testing, and shipping
(including performance checks and packing). Since the
Formula One team, which was a customer for Honda
engines, was based in the UK, and since almost all the
running tests were conducted in Europe, HRD has
functioned as a logistical supply base for providing
engines and parts for races and running tests in a timely
manner. It was the site not only for receiving and
managing parts sent from Japan and then sending them
to the circuit, but also for reassembling engines,
checking shipments, and procuring parts locally.
The technical staff that manages the testing site was
initially made up of mostly of visiting personnel from
Japan, but as ties with the team grew stronger, it became
essential to assign regular personnel to the UK base in
order to enhance communication between the two
facilities. Employees from several sectors, including
Engines, Electrical Components, Materials, Design, and
Assembly, were dispatched to HRD on long-term
assignments, and they performed the pivotal roles of
onsite administration and maintaining ties with the team.
The Formula One Team and HGT were each assigned
roles in chassis development, but in the interests of
sharing technology and enhancing the abilities of the
team, HGT staff members took part in each area of the
teams concerns, including Aerodynamics, VDG,

Composite Materials, and Mechanical Design, and
worked on chassis development as members of the team.
In addition, HRD took advantage of its location to
promote personal interaction with other Europe-based
Formula One teams and suppliers, and gathered technical
information.
2. Location
As shown in Fig. 1, HRD is located in Bracknell,
almost in the center of Berkshire, about one hour west
of London by car. This location is 80 km from the
Formula One teams base in Northamptonshire and 20
km from Heathrow Airport, making it a convenient
location for Formula One-related activities, which require
a lot of travel.
The layout of HRD, comprising an office building
and a factory, is shown in Fig. 2.
In addition to local British employees, half the staff
was made up of Japanese employees on long-term or
temporary assignments, which made HRDs employeemix somewhat different from that of other overseas
locations. Many Japanese employees are there on shortterm assignments, and in addition to working on projects,
HRD also played a major role in managing housing,
hotels, and travel arrangements for the Japanese
employees.
* Fundamental Technology Research Center

** Automobile R&D Center
312
M40
Oxford
M25
London
M4
HRD
Fig. 1
Heathrow Airport
HRD Location (Source: Microsoft AutoRout)
Fig. 2
HRD Factory and office layout
3. Factory Management
Figure 3 is a chronology of activities at HRD. The
Formula One team was BAR at the beginnings, and then
became BAR-Honda, with a 45% investment by Honda.
During the transition to all-Honda-owned HRF1, HRDs
functions also evolved one after another. The division of
labor between HRF1 and HRD can be essentially
described as vehicle body on one hand, and engines on
the other.
Initially, HRDs function was to serve as a contact
point for BAR, FIA and other Formula One associations,
and as a logistics and supply base. However, as BAR
HONDA shifted to Honda Racing Formula One (HRF1),
contact point for BAR became no longer necessary and
the contact point for FIA and other organizations was
also shifted to HRF1.
Management of the factory, which began in 1999,

started with a focus on engine maintenance and parts
supply to testing sites.
Partly due to almost yearly modifications of
regulations and changes in dealing with the second-string
team, HRD expanded its functions beyond those of a
logistics and supply base. From 2005, HRD functions
have been shifted away from only engine and parts
supply and replacement, making HRD an organization
that takes advantage of its location and its mobility for
primary analysis and solution of racing test issues, initial
powertrain checks and settings before running tests, local
parts procurement and other such functions. It also
worked on gathering information by investigating
European technology, which will be discussed later.
Due to modifications of regulations and other factors,
the number of engines supplied decreased or increased
313
as the occasion demanded, but the share of engines

shipped from the UK was increased. The reason for the
increase was a shift of the routine business of shipping
engines to HRD, with the objective being to allow the
Japanese side to strengthen its development efforts. In
fact, an initial ratio of approximately 4:1 (HGT:HRD)
had shifted to 1:3 by 2008, leading to an environment
in which HGT could more easily serve as the main
player in engine development.
Primary analysis and solution of racing test issues,
and powertrain initial checks and settings before running
tests, which were the core of HRDs expanded functions,
took the greatest possible advantage of its location. HRD
conducted bench tests, including initial settings, before
shakedowns of new vehicles or functions, helping to
prevent issues from occurring during initial runs.
As shown in Fig. 3, the increase in engines supplied,
early analysis and solution of issues, and enhancement
in engine quality led to changes in the HRD facility.
HRD was equipped with dynamometers and materials
inspection devices. At first there was only one stationary
dynamometer, but as HRDs role changed, transient
dynamometers were added, and powertrain settings
performed. At the same time, the Materials Department
upgraded its equipment to include such things as a
scanning electron microscope, making it possible for the
department to analyze factors that contributed to issues
at the test site and to make the primary judgments. A
five-axis milling machine and three-dimensional
measuring machine were installed, and a clean room was
introduced in connection with assembly machinery and
inspection equipment. In this and other ways, efforts
were made to enhance parts inspection and quality and
efficiency of the assembly process.
Year
99
00
F1 project
History
These actions facilitated the HRD factorys basic role

as a frontline base for managing races and running tests.
4. Races and Running Tests

Of the 18 races in which the team participated in
2008, there were eight Grand Prix races that required
travel to distant countries such as Japan or China and
10 races in the European round. Some of the teams
participated in running tests in Bahrain, but Honda
conducted all of its tests in Europe, mainly in Spain.
Parts were transported overland by trailer or truck in
Europe, while engines from Japan were shipped to the
UK by air, after which they were transported to the
circuit overland along with the HRD engines.
When there were overseas races, the components that
were normally transported by trailer were packed into a
crate for long-distance shipping commonly called a black
box, and transported by air. Some components were used
only for overseas races and were usually stored at HRD.
In addition to engineers and mechanics, the racing
staff consisted of support personnel, such as truck
drivers. There could be more than 30 persons from HRD
alone, in addition to the team members, making a total
group of nearly 100 persons comprising one team. On
the circuit, this included staff members who prepared
meals for team members and their guests, since each
team had to prepare its own meals.
An important duty of the HRD onsite staff was to
guarantee the reliability and performance of engines on
the circuit, and in order to get the maximum performance
out of an engine, they had to set the engine equipment
and data according to the weather, the features of the
course, and the race strategy, as well as the drivers own
01
02
03
BAR
04
05
BAR HONDA
JORDAN
06
07
08
HRF1
Super AGURI F1
Window function for BAR

Front base of
Window function for

FIA and team party
HRD function
Engine parts supply base
Engine maintenance
Reinforcement
Engine assembly
Logistics
F1 project
Troubleshooting
Powertrain setting
EU technical research
Engine quality enhancement
Local purchase control
Race and track test activity

#1Dyno
#2Dyno
#3Dyno
ONO
ONO
AVL
HRD facility
5-axis machining centre
Laser probe system
Fig. 3
HRD history
314
Material SEM equipment

Clean room
driving habits and preferences.

Running tests were thought of as the final checks for
items that already had been checked for performance by
unit test and bench test. Drivers comments and data
were used to check performance, while checks of race
mileage reliability, temperature, and vibration were
among the tests performed to check conformance to
actual vehicles.
Ignition timing settings were changed in 0.5-degree
increments based on the outside air temperature,
humidity, and engine temperature. The whole process
was painstakingly managed, so that, for example, during
a pit stop in the middle of a race, drivers were told to
change the engine mode if the engine temperature was
too high.
Since shift-timing and straight-end speed affect lap
time and frequency of maximum engine speed, gear
ratios were selected with consideration for the features
of the circuit and the direction of the wind, and in
consultation with the vehicle engineer. Since 2008, gear
ratio exchanges have been limited to one per race, so
ratios have to be selected in advance in anticipation of
the preliminary rounds and the final race. This led to
serious discussions after the practice runs on Friday.
Since cooling performance is directly linked with the
reliability of the hardware, it is important to ascertain
the cooling settings of each part at the assembly stage
of new vehicles in particular. It was therefore adjusted
for throughout the season, depending on the outside air
temperature and use environment. In reality, however, it
was difficult to determine everything beforehand in low
temperature conditions during winter tests, so we often
struggled with onsite settings due to unanticipated
intense heat or unusual weather conditions.
It is usually difficult to determine whether or not an
engine setting is appropriate through data alone, so
settings were carried out in consultations with the driver
and the vehicle engineer, as they determined what points
could be enhanced and what the general settings should
be. For example, when the safety car was introduced, the
lean driving limits for saving fuel were determined by
the driver in the midst of joint runs, and on the basis of
whether engine response would become an issue. The
pedal-to-torque characteristics in a drive-by-wire

situation, as shown in Fig. 4, were tuned for each circuit
based on the drivers comments.
When engine blow and other mishaps occurred,
primary factor analysis was conducted on the circuit, and
the secondary analysis was conducted at HRD, depending
on the urgency of further investigations. HRD often had
to take emergency measures when trouble arose during a
race or during the test runs just before a race.
5. Investigating European Technology

Another responsibility of the onsite personnel was to
gather information about trends in regulations,
competitiveness, and technology, whether through
official or unofficial channels. Since it was difficult for
HGT to gain an overview of this information, it was
considered to be one of the major roles of onsite
personnel to feed back their perceptions of directions in
technology to development.
Programs that took advantage of HRDs location as
a Formula One base in Europe were the Formula One
Engine European Technology Research and
Manufacturers Joint Development Project, which began
in 2006 with HRD local staff playing the central role.
We first contacted Formula One engine builders,
design consultants, parts manufacturers, and simulation
development manufacturers who were based in Europe
and surveyed them, comparing Formula One engine
technology trends in Europe with Hondas engines.
As a result of comparative surveys, it was found that
Formula One engine trends in Europe revolved around
systems with high-lift, narrow-duration camshaft profiles,
lightweight valve-train systems, and cam gear-train
damper systems.
In particular, they led the way in the development of
cam gear-train damper systems, with damper
characteristic specifications being determined with geartrain simulations.
Figure 5 and Table 1 show the names and locations
100
High-speed circuit
90
Grainger Worral
80
Torque (%)
70
IST
60
50
Ilmor
40
30
20
Cosworth
Low-speed circuit
10
0
0
Fig. 4
10 20 30 40 50 60 70 80 90 100
Pedal (%)
CD-Adapco
Pedal-to-torque characteristics
Fig. 5
315
Supplier location (Source: MapQuest)

Table 1
Supplier
Supplier and development items

Company profile
Development item
EU concept
cylinder head assy
Inlet valve (Ti3Al)
(high stiffness / light weight)
High-lift + narrow-duration
cam profile
IST
Consulting company for F1

engine design and material
EU concept
cylinder block assy
Light-weight piston
( 5-axis machine)
Twin oil circuit system
(piston jet - crank)
EU concept air box
8-cylinder LAF sensor
+Real-time mapping
Ilmor
Ex. F1 engine design and

manufacturer
EU style head-port design

High-energy IG coil
Ex. F1 engine design and

manufacturer
Tortional damper for cam gear-train
Grainger Worrall
F1 casting parts
manufacturer
Forced-air chill-casting for cylinder

head
CD-Adapco
CFD simulation development
Port / Expi / Airbox 3D CFD

simulation
Cosworth
of the manufacturers with which HRD carried out its

joint development projects. Most of the development
projects were eventually carried out with manufacturers
in the UK, and this is believed not unconnected to the
fact that Mercedes High Performance Engine, Cosworth,
Ilmor, and other top-ranked racing engine constructors
are based in the UK.
These joint projects allowed HRD to propose the
aforementioned high-lift, narrow-duration camshaft
profile, cam gear-train damper system, and simulations,
and their effectiveness was verified in preliminary tests
conducted with HGT. Subsequently, in 2009, while they
were preparing to manufacture parts with the aim of
adapting this technology for the Formula One engine
project, Honda decided to withdraw from Formula One
racing.
Gear train (+damper) simulation
6. Conclusion
At HRD, which functioned as a frontline base far
away from Japan, both short-term and long-term
personnel encountered difficulties due to the overseas
working environment and unique strictures of racing. As
we struggled on the front lines in direct contact with
races and teams, we learned a great deal and were able
to grow.
We would like to take this opportunity to thank
everyone inside and outside Honda who constantly gave
priority to local programs and supported and cooperated
with us.
Author
Yusuke HASEGAWA
316
Shiro HISATSUNE
Tatsuya KODAMA
Concluding Remarks for Hondas Third-Era

Katsuhiko SUZUKI*
To all the members of the Formula One Project, I

would like to say, Thank you and well done.
To all our friends and colleagues inside and outside
the company who generously gave their support, I must
offer apologies. Our strength was not sufficient to keep
the project going, and so it came to an end before you
could enjoy it to the full. I am truly sorry.
My own involvement with this Formula One project
started on August 31, 2007. That was when I received
a telephone call from Vice President Ikeno of the R&D
Center. I want you to become Operating Officer of the
MS Department (the Motor Sports Development
Department) from September 1, he told me. But
September 1 is tomorrow, I said, and even as I spoke
those words, I remembered that something very similar
had happened to me before.
This was back when I was still an assistant chief
engineer, and it was the afternoon of the day before I
was supposed to leave on an overseas business trip when
a telephone call came from the Manager. He said,
Everything has been called off! Youre going to start
on a special assignment tomorrow, instead! At that
point, all I could say was a very vague, Huh? What?
This had taken me completely by surprise, and I had to
wonder whether they could really do this. If you said this
was a Honda-like way of doing things, then I would have
to agree, so far as that goes.
In 2007, I was General Manager of the Mass
Production Engine Development Division, and at that
point we were involved in some very challenging issues
related to a major mass-production start-up under
extremely demanding conditions. When this latter request
came, therefore, I insisted that I should wear two hats
until March 2008, at least, and they let me
simultaneously remain as Operating Officer of the Mass
Production Engine Development Division and Operating
Officer of the MS Department. As it turned out, it might
have been better if I had concentrated on the MS
Department. It seems to me, looking back at it now, that
the MS Department actually was in greater need of that
level of management and engineering.
When I took responsibility for the MS Department,
I realized that our biggest sore point was communication:
Communication among the Honda Formula One Racing
Team (HRF1), the Honda R&D Co., Ltd. Automobile
R&D Center in Tochigi (HGT), and Honda Racing

Development (HRD);
Communication between the Honda Motor Co. head office
(HM) and HGT; and
Communication among the HGT engine, chassis, gearbox,
assembly, and operations divisions.
Unless communication of this kind is taking place
smoothly, things do not get decided properly. And even
if they do get decided, positive results cannot be
expected unless a consensus is also reached.
What I did, therefore, was to have some long talks
with Ross Brawn, the HRF1 team representative. As the
individuals responsible for respective development in the
UK and Japan, he and I set out to clearly define the roles
and responsibilities of the various divisions concerned.
The result was that the mobilization of our intentions and
technologies regarding the 2009 car proceeded according
to schedule, and is currently contributing to the activities
of Brawn GP, which took over the project.
Although our effort ended before KERS could be
used, we did manage to build up a system with the
worlds highest level of technology in this area, as well.
This includes the super compact 60-kW motor, the PDU
with ultra-fast processing capability, and the laminated
lithium ion secondary batteries. I feel quite certain that
these technologies are sure to be of use to developers
in some form or other in the future.
Next, on the matter of communication with HM, both
President Abe of HRD and General Manager Muramatsu
of the head office Motor Sports Division shared a
common perception of issues with us, and discussed
what needed to be done on both hard and soft aspects
of the project each time something came up. This
included three conferences with HM, HRF1, HRD, and
HGT that took place during the race season.
Regarding the matter of communication in the MS
Department, which is an issue within HGT, I have the
impression that in a time like this, when engine
development is subject to such rigorous controls by the
homologation system, the competition of technology has
shifted over to the comprehensive enhancement of
vehicle dynamics performance. However, the outstanding
talents in the Vehicle Chassis Division at HGT were not
in a position to demonstrate everything they were
capable of doing. I think that what turned this situation

317
Concluding Remarks for Hondas Third-Era Formula One Activities
around for them was the neutral thinking of Ross Brawn.

Just one example is the rear diffuser for the 2009 car.
In the development competition between HRF1 and
HGT, the HGT proposal was adopted. Beyond that, they
have also reinforced their contingent of HGT members
stationed in the UK, and have started actively taking part
in concept meetings for coming vehicle versions. The
knowledge they have cultivated in this way is an
irreplaceable resource that is in the keeping of the
younger HGT members.
This is not just the case with the Vehicle Chassis
Division: the Engine Division has also adopted a special
framework for development. With the engine for the
2009 season having been partially released from
homologation by the International Automobile Federation
(FIA), the Engine Division went from development
emphasizing peak output to development emphasizing
mid- to high-range torque according to lap-time
simulations. They accomplished this in the short timeframe of three months. The first model of their engine
ran on the test bench and generally yielded the
performance they desired. As I see it, realizing the ability
to design engines from the conceptual approach of
overall optimization (namely, lap-time simulation) as
seen in this engine development effort contributed to the
integration of communication by the MS Department and
evolution by the designers. I was also impressed all over
again by the pride and the potential we found in these
engine men. It took them less than two months from
starting development to having their first engine running.
Ever since the seamless shift mechanism was
developed, Honda has constantly held the lead with the
most advanced Formula One gearbox technology. The
lightweight, compact iMAX shift mechanism built into
the shaft that had been scheduled for adoption in 2010
is the ultimate one of this kind. I hope that this kind of
thinking, which pursuits the ideal goal and then achieves
it, will go to work in mass-production, as well.
The Assembly, Procurement, and Operations
Divisions have performed extremely well as unsung
heroes in supporting roles. I feel that they served as my
own two hands. They really wanted to help with
winning, and I could feel the sincerity in our
communication at that level. They have a full
understanding of their respective functions, and they
were developing frameworks for cooperation with HRF1,
such as the dedicated system development project aimed
at realizing smooth logistics.
I was able to spend a very enjoyable year and a half,
thanks to the efforts of all these people, and I am truly
grateful. Thank you very much. Finally, when conditions
take a turn for the better, I fully expect that Honda will
be making a comeback to Formula One racing.
After all, Formula One is the very pinnacle of racing.
It is the ultimate discipline where victory can only be
gained when the elements of technological capability,
full range of speed, and management capability are all
present.
318
Race Records
Specification Summary for Honda's Third-Era Formula One
Chassis, Engine (1999)

RA099 Prototype
Chassis specifications (RA099)
Wheelbase (mm)
3020
Gravitational center hight (mm,
250*
above reference plane)
Main monocoque weight (kg)
54
Tire supplier
Bridgestone
Tire size-front
265/55R13
Tire size-rear
325/45R13
Brake caliper (front and rear)
6 pistons
Brake disk (front and rear)
Carbon, ventilated
Suspension-front
Double wishbone, push rod
Suspension-rear
Chassis electronic control system
Pi
*With driver, without fuel
Engine specifications (RA099E)

Engine Type
90 V10
2995.6
Displacement (cm3)
Cylinder diameter (mm)
94.4
Stroke (mm)
42.8
Weight (kg)
122
Maximum power (kW)
563 Over
Maximum Ne (rpm)
16300 Over
Engine electronic control system
Honda
Honda
Test Driver
Chassis ; RA099
Engine ; RA099E
Jos Verstappen (Netherland)
320

BAR002
Chassis specifications (BAR002)
Wheelbase (mm)
3108
259*
61
Tire supplier
Bridgestone
Tire size-front
265/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi
Gear box specification (BAR002)

Gear box case
Aluminum Casting
Changing mechanism
Xtrac
Forward gear ratios
6 speed, sequential
Clutch disk
Carbon
Clutch mechanism
Pull clutch
RA000E
Engine Type
88 V10
2995.6
Displacement (cm3)
94.4
Stroke (mm)
42.8
Compression ratio
13.4
Weight (kg)
112
Maximum power (kW)
567 Over
Maximum Ne (rpm)
17000 Over
Mileage (km)
400
Honda
Lucky Strike B.A.R Honda

Driver
Chassis ; BAR002
Engine ; RA000E
CARNo.22 Jacques Villeneuve (Canada)
CARNo.23 Ricardo Zonta (Brazil)
321
Ranking
5th
7th
14th

BAR003
Wheelbase (mm)
3183
245*
65
Tire supplier
Bridgestone
Tire size-front
265/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi

Gear box case
Magnesium Casting
Changing mechanism
Xtrac
Forward gear ratios
6 speed, sequential
Clutch disk
Carbon
Clutch mechanism
Pull clutch
EJ11
RA001E
Engine Type
80 V10
2994.1
Displacement (cm3)
95
Stroke (mm)
42.24
Compression ratio
13.4
Weight (kg)
108
Maximum power (kW)
588 Over
Maximum Ne (rpm)
17300 Over
Mileage (km)
400
Honda

Driver
Test & Reserve Driver
B & H Jordan Honda

Driver
Chassis ; BAR003
Engine ; RA001E
CAR No. 9 Olivier Panis (France)
CAR No. 10 Jacques Villeneuve (Canada)
Darren Manning (United Kingdom)
Chassis ; EJ11
Engine ; RA001E
CAR No. 11 Heinz-Harald Frentzen (Germany)
CAR No. 12 Jarno Trulli (Italy)
CAR No. 12 Jean Alesi (France)
CAR No. 11 Ricardo Zonta (Brazil)
322
Ranking
6th
14th
7th
Ranking
5th
13th
9th
15th
19th

BAR004
Wheelbase (mm)
3133
243*
61
Tire supplier
Bridgestone
Tire size-front
265/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi
EJ12

Gear box case
Magnesium Casting
Changing mechanism
Xtrac
Forward gear ratios
6 speed, sequential
Clutch disk
Carbon
Clutch mechanism
Pull clutch
RA002E
Engine Type
94 V10
2994.4
Displacement (cm3)
97
Stroke (mm)
40.52
Compression ratio
12.2
Weight (kg)
111
Maximum power (kW)
640 Over
Maximum Ne (rpm)
18000 Over
Mileage (km)
400
Honda

Driver
DHL Jordan Honda

Driver
Chassis ; BAR004
Engine ; RA002E
CAR No. 12 Olivier Panis (France)
Anthony Davidson (United Kingdom)
Chassis ; EJ12
Engine ; RA002E
CAR No. 9 Giancarlo Fisichella (Italy)
CAR No. 10 Takuma Sato (Japan)
323
Ranking
8th
12th
14th
Ranking
6th
11th
15th

BAR005
Wheelbase (mm)
3194
227*
54
Tire supplier
Bridgestone
Tire size-front
265/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi

Gear box case
Aluminum Casting
Changing mechanism
Xtrac
Forward gear ratios
7 speed, sequential
Clutch disk
Carbon
Clutch mechanism
Pull clutch
RA003E
Engine Type
90 V10
Displacement (cm3)
2994.4
97
Stroke (mm)
40.52
Compression ratio
12.4
Weight (kg)
99
Maximum power (kW)
662 Over
Maximum Ne (rpm)
18800 Over
Mileage (km)
400
Honda

Driver
Reserve Driver
Test Driver
Chassis ; BAR005
Engine ; RA003E
CAR No. 17 Jenson Button (United Kingdom)
324
Ranking
5th
16th
9th
18th

BAR006
Wheelbase (mm)
3131
229*
56
Tire supplier
Michelin
Tire size-front
270/66R13
Tire size-rear
320/66R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi

CFRP
Gear box case
Honda
Changing mechanism
7 speed, sequential
Forward gear ratios
Carbon
Clutch disk
Pull clutch
Clutch mechanism
RA004E
90 V10
Engine Type
2994.4
Displacement (cm3)
97
40.52
Stroke (mm)
12.8
Compression ratio
91
Weight (kg)
662 Over
Maximum power (kW)
19000 Over
Maximum Ne (rpm)
800
Mileage (km)
Honda

Driver
Chassis ; BAR006
Engine ; RA004E
Anthony Daivdson (United Kingdom)
325
Ranking
2nd
3rd
8th

BAR007
Wheelbase (mm)
3085
230*
56
Tire supplier
Michelin
Tire size-front
270/66R13
Tire size-rear
320/66R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Pi

Gear box case
CFRP
Honda
Changing mechanism
(Seamledss shift)
Forward gear ratios
7 speed, sequential
Clutch disk
Carbon
Clutch mechanism
Pull clutch
RA005E
Engine Type
90 V10
Displacement (cm3)
2994.4
97
Stroke (mm)
40.52
Compression ratio
12.8
Weight (kg)
89
Maximum power (kW)
662 Over
Maximum Ne (rpm)
19200 Over
Mileage (km)
1400
Honda

Driver
Chassis ; BAR007
Engine ; RA005E
CAR No. 4 Anthony Davidson (United Kingdom)
326
Ranking
6th
9th
23rd

RA106

until Rd.13 : 3085
Wheelbase (mm)
as from Rd.14 : 3135
231*
57
Tire supplier
Michelin
Tire size-front
270/66R13
Tire size-rear
320/66R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Honda
SA05 (Rd.1-Rd.11)
Gear box specification (RA106)

Gear box case
CFRP
Honda
Changing mechanism
(Seamledss shift)
Forward gear ratios
7 speed, sequential
Clutch disk
Carbon
Clutch mechanism
DPC
SA06 (Rd.15-Rd.18)
RA806E
Engine Type
90 V8
Displacement (cm3)
2395.5
97
Stroke (mm)
40.52
Compression ratio
12.8
Weight (kg)
95
Maximum power (kW)
537 Over
Maximum Ne (rpm)
19600 Over
Mileage (km)
1400
Honda
Lucky Strike Honda Racing F1 Team

Driver
Super Aguri Formula 1

Driver
Chassis ; RA106
Engine ; RA806E
CAR No. 12 Rubens Barrichello (Brazil)
Chassis ; SA06
Engine ; RA806E
CAR No. 23 Yuji Ide (Japan)
CAR No. 23 Sakon Yamamoto (Japan)
CAR No. 23 Franck Montagny (France)
327
Ranking
4th
6th
7th
Ranking
11th
23rd
25th
26th
27th

RA107
Wheelbase (mm)
3165
235*
56
Tire supplier
Bridgestone
Tire size-front
270/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Honda
SA07

Gear box case
CFRP
Honda
Changing mechanism
(Seamledss shift)
Forward gear ratios
7 speed, sequential
Carbon
Clutch disk
Clutch mechanism
DPC
RA807E
Engine Type
90 V8
2395.5
Displacement (cm3)
97
Stroke (mm)
40.52
Compression ratio
13
Weight (kg)
95
Maximum power (kW)
544 Over
Maximum Ne (rpm)
19000 Limit
Mileage (km)
1400
Honda

Driver
Super Aguri F1 Team

Driver
Chassis ; RA107
Engine ; RA807E
Christian Klien (Austria)
Chassis ; SA07
Engine ; RA807E
James Rossiter (United Kingdom)
328
Ranking
8th
15th
20th
Ranking
9th
17th
23rd

RA108
Wheelbase (mm)
3210
243*
55
Tire supplier
Bridgestone
Tire size-front
270/55R13
Tire size-rear
325/45R13
6 pistons
Carbon, ventilated
Suspension-front
Suspension-rear
Microsoft MES
SA08

Gear box case
CFRP
Honda
Changing mechanism
(Seamledss shift)
Forward gear ratios
7 speed, sequential
Clutch disk
Carbon
Clutch mechanism
DPC
RA808E
Engine type
90 V8
2395.5
Displacement (cm3)
97
Stroke (mm)
40.52
Compression ratio
13
Weight (kg)
95
Maximum power (kW)
548 Over
Maximum Ne (rpm)
19000 Limit
Mileage (km)
1400
Microsoft MES

Driver
Super Aguri F1 Team

Driver
Chassis ; RA108
Engine ; RA808E
Alex Wurz (Austria)
Chassis ; SA08
Engine ; RA808E
James Rossiter (United Kingdom)
329
Ranking
9th
18th
14th
Ranking
11th
21st
22nd
2000 Honda Formula One Racing Data Tables

FIA FORMULA ONE WORLD CHAMPIONSHIP
Lucky Strike BAR Honda
ROUND 1 AUSTRALIA GP
QUALIFYING
Melbourne
RACE
5.302km58Laps=307.574km
(DELAY)
ROUND 2 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.325km71Laps=305.909km
(DELAY)
ROUND 3 SAN MARINO GP
QUALIFYING
Imola
RACE
4.693km62Laps=305.609km
(DELAY)
#22 J. VILLENEUVE
#23 R. ZONTA
8th (1'31.968)
16th (1'33.117)
4th
6th
(44.447)
(46.468)
10th (1'15.515)
8th (1'15.484)
R (16LAPS)
9th
(2LAPS)
9th (1'26.124)
14th (1'26.814)
5th
12th
(1LAP)
(1LAP)
10th (1'27.025)
16th (1'27.772)
16th
R (36LAPS)
(DNF)
ROUND 4 GREAT BRITAIN GP
QUALIFYING
Silverstone
RACE
5.141km60Laps=308.356km
(DELAY)
ROUND 5 SPAIN GP
QUALIFYING
6th (1'21.963)
16th (1'22.882)1
Barcelona
RACE
R (21LAPS)
8th
4.727km65Laps=307.323km
(DELAY)
(1LAP)
ROUND 6 EUROPE GP
QUALIFYING
9th (1'18.742)
18th (1'19.766)2
Nurburgring
RACE
R (46LAPS)
R (51LAPS)
4.556km67Laps=305.235km
(DELAY)
ROUND 7 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.367km78Laps=262.860km
(DELAY)
ROUND 8 CANADA GP
QUALIFYING
Montreal
RACE
4.421km69Laps=305.049km
(DELAY)
ROUND 9 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.250km72Laps=305.886km
(DELAY)
ROUND 10 AUSTRIA GP
QUALIFYING
Spielberg
RACE
4.319km71Laps=307.146km
(DELAY)
ROUND 11 GERMANY GP
QUALIFYING
Hockenheim
RACE
6.823km45Laps=307.125km
(DELAY)
ROUND 12 HUNGARY GP
QUALIFYING
Budapest
RACE
3.971km77Laps=306.075km
(DELAY)
ROUND 13 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
6.968km44Laps=306.592km
(DELAY)
ROUND 14 ITALY GP
QUALIFYING
Monza
RACE
5.770km53Laps=306.764km
(DELAY)
ROUND 15 UNITED STATES GP QUALIFYING

Indianapolis
RACE
4.192km73Laps=305.999km
(DELAY)
ROUND 16 JAPAN GP
QUALIFYING
Suzuka
RACE
5.864km53Laps=310.596km
(DELAY)
ROUND 17 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.542km56Laps=310.408km
(DELAY)
POLE
PODIUM
POSITION
(FIRST 3)
#1 M. HAKKINEN
1'30.556
#3 M. SCHUMACHER
#4 R. BARRICHELLO
#9 R. SCHUMACHER
#1 M. HAKKINEN
1'14.111
#3 M. SCHUMACHER
#11 G. FISICHELLA
#5 H. FRENTZEN
#1 M. HAKKINEN
1'24.714
#3 M. SCHUMACHER
#1 M. HAKKINEN
#2 D. COULTHARD
#4 R. BARRICHELLO
1'25.703
#2 D. COULTHARD
#1 M. HAKKINEN
#3 M. SCHUMACHER
#3 M. SCHUMACHER
1'20.974
#1 M. HAKKINEN
#2 D. COULTHARD
#4 R. BARRICHELLO
#2 D. COULTHARD
1'17.529
#3 M. SCHUMACHER
#1 M. HAKKINEN
#2 D. COULTHARD
17th (1'21.848)
20th (1'22.324)
#2 D. COULTHARD
7th
R (48LAPS)
(1LAP)
6th (1'19.544)
8th (1'19.742)
15th
8th
(DNF)
7th (1'16.653)
19th (1'17.668)
4th
R (16LAPS)
(61.322)
7th (1'11.649)
6th (1'11.647)
4th
R (58LAPS)
(1LAP)
9th (1'48.121)
12th (1'48.665)
8th
R (37LAPS)
(47.537)
16th (1'19.937)
18th (1'20.272)
12th
14th
(2LAPS)
(2LAPS)
7th (1'51.799)
13th (1'53.002)
7th
12th
(72.380)
(1LAP)
4th (1'24.238)
17th (1'25.337)
R (14LAPS)
6th
(69.292)
8th (1'15.317)
12th (1'15.784)
4th
6th
(17.935)
(51.694)
9th (1'37.267)
18th (1'38.269)
6th
9th
(1LAP)
(1LAP)
6th (1'38.653)
11th (1'39.158)
5th
R (46LAPS)
(70.692)
#3 M. SCHUMACHER
1'19.475
#4 R. BARRICHELLO
#11 G. FISICHELLA
#3 M. SCHUMACHER
1'18.439
#3 M. SCHUMACHER
#4 R. BARRICHELLO
#11 G. FISICHELLA
#3 M. SCHUMACHER
1'15.632
#2 D. COULTHARD
#1 M. HAKKINEN
#4 R. BARRICHELLO
#1 M. HAKKINEN
1'10.410
#1 M. HAKKINEN
#2 D. COULTHARD
#4 R. BARRICHELLO
#2 D. COULTHARD
1'45.697
#4 R. BARRICHELLO
#1 M. HAKKINEN
#2 D. COULTHARD
#3 M. SCHUMACHER
1'17.514
#1 M. HAKKINEN
#3 M. SCHUMACHER
#2 D. COULTHARD
#1 M. HAKKINEN
1'50.646
#1 M. HAKKINEN
#3 M. SCHUMACHER
#9 R. SCHUMACHER
#3 M. SCHUMACHER
1'23.770
#3 M. SCHUMACHER
#1 M. HAKKINEN
#9 R. SCHUMACHER
#3 M. SCHUMACHER
1'14.266
#3 M. SCHUMACHER
#4 R. BARRICHELLO
#5 H. FRENTZEN
#3 M. SCHUMACHER
1'35.825
#3 M. SCHUMACHER
#1 M. HAKKINEN
#2 D. COULTHARD
#3 M. SCHUMACHER
1'37.397
#3 M. SCHUMACHER
#2 D. COULTHARD
#4 R. BARRICHELLO
1 Raised from 17th to 16th position after a fuel infringement by P. de la Rosa in 9th position caused his qualifying times to be deleted.
2 Raised from 19th to 18th position after a minimum weight infringement by N. Heidfeld in 13th position resulted in his disqualification.
330

CONSTRUCTORS WORLD CHAMPIONSHIP
P.
TEAM (CAR NO.)
TOTAL
AUS
BRA
SMR
GBR
SPA
EUR
MON
CAN
FRA
AUT
GER
HUN
BEL
ITA
USA
JPN
MAL
Scuderia Ferrari Marlboro
170
16
10
13
13
16
10
10
16
13
14
West McLaren Mercedes
152
10
16
16
10
11
16
10
14
13
10
BMW Williams F1 Team
36
3
6
Mild Seven Benetton Playlife
20
20
Benson & Hedges Jordan
17
Arrows
Red Bull Sauber Petronas
Jaguar Racing
10
Telefonica Minardi Fondmetal
11
Gauloises Prost Peugeot
2
7
1
1
2
1
4
3
DRIVERS WORLD CHAMPIONSHIP

P.
No.
TOTAL
AUS
BRA
SMR
GBR
SPA
EUR
M. SCHUMACHER (GER)
108
10
10
10
10
M. HAKKINEN (FIN)
89
10
D. COULTHARD (GBR)
73
10
10
R. BARRICHELLO (BRA)
62
R. SCHUMACHER (GER)
24
11
G. FISICHELLA (ITA)
18
22
J. VILLENEUVE (CAN)
17
10
J. BUTTON (GBR)
12
DRIVER (NAT)
3
3
MON
CAN
3
6
H. FRENTZEN (GER)
11
J. TRULLI (ITA)
11
17
M. SALO (FIN)
12
19
J. VERSTAPPEN (NED)
13
E. IRVINE (GBR)
14
23
R. ZONTA (BRA)
15
12
A. WURZ (AUT)
16
18
P. DE LA ROSA (ESP)
17
J. HERBERT (GBR)
18
16
P. DINIZ (BRA)
19
20
M. GENE (ESP)
20
15
N. HEIDFELD (GER)
21
21
G. MAZZACANE (ARG)
22
14
J. ALESI (FRA)
23
L. BURTI (BRA)
AUT
GER
HUN
BEL
ITA
USA
JPN
MAL
10
10
10
10
10
10
10
10
10
3
2
3
3
2
1
3
2
10
FRA
10
1
1
1
2
1
331

#9 O. PANIS
QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km55Laps=304.865km
(DELAY)
ROUND 3 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
ROUND 4 SAN MARINO GP QUALIFYING

Imola
RACE
4.933km62Laps=305.609km
(DELAY)
ROUND 5 SPAIN GP
QUALIFYING
Barcelona
RACE
4.730km65Laps=307.323km
(DELAY)
ROUND 6 AUSTRIA GP
QUALIFYING
Spielberg
RACE
4.326km71Laps=307.146km
(DELAY)
B&H Jordan Honda
#10 J. VILLENEUVE #11 H. FRENTZEN
#12 J. TRULLI
POLE
PODIUM
POSITION
(FIRST 3)
9th (1'28.518)
8th (1'28.435)
4th (1'27.658)
7th (1'28.377)
7th1
R (4LAPS)
5th2
R (38LAPS)
(62.050)
(72.807)
#2 R. BARRICHELLO
10th (1'36.681)
7th (1'36.397)
9th (1'36.578)
5th (1'36.180)
#1 M. SCHUMACHER
R (1LAP)
R (3LAPS)
4th
8th
(46.543)
(1LAP)
11th (1'15.046)
12th (1'15.182)
8th (1'14.633)
7th (1'14.630)
4th
7th
11th
5th
(1LAP)
(1LAP)
(DNF)
(1LAP)
8th (1'24.213)
11th (1'24.769)
9th (1'24.436)
5th (1'23.658)
8th
R (30LAPS)
6th
5th
(1LAP)
(1LAP)
(85.558)
11th (1'19.479)
7th (1'19.122)
8th (1'19.150)
6th (1'19.093)
7th
3th
R (5LAPS)
4th
(64.977)
(49.626)
(51.253)
10th (1'10.435)
12th (1'11.058)
11th (1'10.923)
5th (1'10.202)
5th
8th
R (0LAP)
(53.776)
(1LAP)
12th (1'19.294)
9th (1'19.086)
13th (1'19.316)
8th (1'18.921)
R (13LAPS)
4th
R (49LAPS)
R (30LAPS)
(32.454)
ROUND 7 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.370km78Laps=262.860km
(DELAY)
ROUND 8 CANADA GP
QUALIFYING
6th (1'16.771)
9th (1'17.035)
12th (1'17.328)
4th (1'16.459)
Montreal
RACE
R (38LAPS)
R (34LAPS)
7th
11th
4.421km69Laps=305.049km
(DELAY)
(1LAP)
(DNF)
ROUND 9 EUROPE GP
QUALIFYING
13th (1'16.872)
11th (1'16.439)
8th (1'16.376)
7th (1'16.138)
Nurburgring
RACE
R (23LAPS)
9th
R (48LAPS)
R (44LAPS)
#1 M. SCHUMACHER
1'26.892
#1 M. SCHUMACHER
1'35.220
#1 M. SCHUMACHER
#4 D. COULTHARD
#2 R. BARRICHELLO
#4 D. COULTHARD
#1 M. SCHUMACHER
1'13.780
#4 D. COULTHARD
#1 M. SCHUMACHER
#16 N. HEIDFELD
#4 D. COULTHARD
1'23.054
#5 R. SCHUMACHER
#4 D. COULTHARD
#2 R. BARRICHELLO
#1 M. SCHUMACHER
1'18.201
#1 M. SCHUMACHER
#6 J. MONTOYA
#10 J. VILLENEUVE
#1 M. SCHUMACHER
1'09.562
#4 D. COULTHARD
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#4 D. COULTHARD
1'17.430
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#18 E. IRVINE
#11 R. ZONTA
#1 M. SCHUMACHER
1'15.782
#5 R. SCHUMACHER
#1 M. SCHUMACHER
#3 M. HAKKINEN
#11 H. FRENTZEN
4.556km67Laps=305.235km
(DELAY)
ROUND 10 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.251km72Laps=305.886km
(DELAY)
ROUND 11 GREAT BRITAIN GP QUALIFYING

Silverstone
RACE
5.141km60Laps=308.356km
(DELAY)
(1LAP)
11th (1'14.181)
10th (1'14.096)
7th (1'13.815)
5th (1'13.310)
9th
R (5LAPS)
8th
5th
(1LAP)
(1LAP)
(68.285)
11th (1'22.316)
12th (1'22.916)
5th (1'21.217)
4th (1'20.930)
R (0LAP)
8th
7th
R (0LAP)
(1LAP)
(1LAP)
#1 M. SCHUMACHER
1'14.960
#1 M. SCHUMACHER
#6 J. MONTOYA
#4 D. COULTHARD
#5 R. SCHUMACHER
1'12.989
#1 M. SCHUMACHER
#5 R. SCHUMACHER
#2 R. BARRICHELLO
#1 M. SCHUMACHER
1'20.447
#3 M. HAKKINEN
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#11 R. ZONTA
ROUND 12 GERMANY GP
QUALIFYING
Hockenheim
RACE
6.825km45Laps=307.125km
(DELAY)
ROUND 13 HUNGARY GP
QUALIFYING
Budapest
RACE
3.975km77Laps=306.075km
(DELAY)
ROUND 14 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
6.968km36Laps=250.831km
(DELAY)
ROUND 15 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.749km
(DELAY)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
ROUND 17 JAPAN GP
QUALIFYING
Suzuka
RACE
5.859km53Laps=310.331km
(DELAY)
13th (1'40.610)
12th (1'40.437)
15th (1'41.174)
10th (1'40.322)
7th
3th
R (7LAPS)
R (34LAPS)
(77.527)
(62.806)
#11 J. TRULLI
#12 J. ALESI
11th (1'16.382)
10th (1'16.212)
5th (1'15.394)
12th (1'16.471)
R (58LAPS)
9th
R (63LAPS)
10th
(2LAPS)
(2LAPS)
11th (1'58.838)
6th (1'57.038)
16th (1'59.647)
13th (1'59.128)
11th
8th
R (31LAPS)
6th
(1LAP)
(64.970)
(59.684)
17th (1'24.677)
15th (1'24.164)
5th (1'23.126)
16th (1'24.198)
9th
6th
R (0LAP)
8th
(1LAP)
(82.469)
(1LAP)
13th (1'13.122)
18th (1'14.012)
8th (1'12.605)
9th (1'12.607)
11th
R (45LAPS)
4th
7th
(1LAP)
(57.423)
(1LAP)
17th (1'35.766)
14th (1'35.109)
8th (1'34.002)
11th (1'34.420)
13th
10th
8th
R (5LAPS)
(2LAPS)
(1LAP)
(1LAP)
#6 J. MONTOYA
1'38.117
#5 R. SCHUMACHER
#2 R. BARRICHELLO
#10 J. VILLENEUVE
#1 M. SCHUMACHER
1'14.059
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#4 D. COULTHARD
#6 J. MONTOYA
1'52.072
#1 M. SCHUMACHER
#4 D. COULTHARD
#7 G. FISICHELLA
#6 J. MONTOYA
1'22.216
#6 J. MONTOYA
#2 R. BARRICHELLO
#5 R. SCHUMACHER
#1 M. SCHUMACHER
1'11.708
#3 M. HAKKINEN
#1 M. SCHUMACHER
#4 D. COULTHARD
#1 M. SCHUMACHER
1'32.484
#1 M. SCHUMACHER
#6 J. MONTOYA
#4 D. COULTHARD
1 O. Panis finished in 4th position, but 25 seconds was added to his racing time for overtaking under a yellow flag, and he was demoted to 7th position.
2 H. Frentzen finished in 6th position, but was raised to 5th position after 25 seconds was added to O. Panis racing time for overtaking under a yellow flag.
332

P.
TEAM (CAR NO.)
TOTAL
AUS
MAL
BRA
SMR
ESP
AUT
MON
CAN
EUR
FRA
GBR
GER
HUN
BEL
ITA
USA
JPN
179
14
16
10
10
16
12
14
10
16
12
12
102
10
10
10
14
80
10
10
Red Bull Sauber Petronas
21
B&H Jordan Honda
19
17
3
1
2
4
3
1
3
3
4
10
3
1
2
2
14
1
Mild Seven Benetton Renault
10
Jaguar Racing
Prost Acer
10
Orange Arrows Asiatech
11
European Minardi F1
7
1
3
4
2

P.
No.
TOTAL
AUS
MAL
BRA
DRIVER (NAT)
M. SCHUMACHER (GER)
123
10
10
D. COULTHARD (GBR)
65
10
56
R. SCHUMACHER (GER)
49
10
M. HAKKINEN (FIN)
37
J. MONTOYA (COL)
31
10
J. VILLENEUVE (CAN)
12
16
N. HEIDFELD (GER)
12
11
J. TRULLI (ITA)
12
10
17
K. RAIKKONEN (FIN)
11
G. FISICHELLA (ITA)
12
18
E. IRVINE (GBR)
13
22
H. FRENTZEN (GER)
14
O. PANIS (FRA)
15
12
J. ALESI (FRA)
16
19
P. DELA ROSA (ESP)
17
J. BUTTON (GBR)
18
14
J. VERSTAPPEN (NED)
19
11
R. ZONTA (BRA)
20
19
L. BURTI (BRA)
21
15
E. BERNOLDI (BRA)
22
20
T. MARQUES (BRA)
23
21
F. ALONSO (ESP)
24
23
T. ENGE (CZE)
25
23
G. MAZZACANE (ARG)
26
20
A. YOONG (MAS)
SMR
6
4
2
ESP
AUT
MON
CAN
EUR
FRA
GBR
10
10
10
10
10
10
10
HUN
BEL
ITA
USA
JPN
10
10
10
10
3
2
2
3
10
10
1
3
GER
1
3
2
3
4
2
1
3
2
1
333
1
2
2
1
1
1

DHL Jordan Honda
#9 G. FISICHELLA
QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
ROUND 3 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)

Imola
RACE
4.933km62Laps=305.609km
(DELAY)
ROUND 5 SPAIN GP
QUALIFYING
Barcelona
RACE
4.730km65Laps=307.327km
(DELAY)
ROUND 6 AUSTRIA GP
QUALIFYING
Spielberg
RACE
4.326km71Laps=307.146km
(DELAY)
ROUND 7 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.370km78Laps=262.860km
(DELAY)
ROUND 8 CANADA GP
QUALIFYING
Montreal
RACE
4.361km70Laps=305.270km
(DELAY)
ROUND 9 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.146km60Laps=308.743km
(DELAY)

Silverstone
RACE
5.141km60Laps=308.356km
(DELAY)
ROUND 11 FRANCE GP
QUALIFYING
#10 T. SATO
#11 J. VILLENEUVE
#12 O. PANIS
8th (1'27.869)
(1'53.351)
13th (1'28.657)
12th (1'28.381)
R (0LAP)
R (12LAPS)
R (27LAPS)
R (0LAP)
9th (1'37.536)
15th (1'38.141)
13th (1'38.039)
18th (1'38.390)
POLE
PODIUM
POSITION
(FIRST 3)
#2 R. BARRICHELLO
1'25.843
#1 M. SCHUMACHER
#6 J. MONTOYA
#4 K. RAIKKONEN
#1 M. SCHUMACHER
1'35.266
#5 R. SCHUMACHER
13th
9th
8th
R (9LAPS)
(3LAPS)
(2LAPS)
(1LAP)
14th (1'14.748)
19th (1'15.296)
15th (1'14.760)
17th (1'14.996)
R (6LAPS)
9th
10th
R (25LAPS)
(2LAPS)
(DNF)
15th (1'24.253)
14th (1'24.050)
10th (1'23.116)
12th (1'23.821)
R (19LAPS)
R (5LAPS)
7th
R (44LAPS)
(1LAP)
12th (1'18.291)
18th (1'19.002)1
15th (1'18.847)
13th (1'18.472)
R (5LAPS)
R (10LAPS)
7th
R (43LAPS)
(1LAP)
15th (1'09.901)
18th (1'10.058)
17th (1'10.051)
9th (1'09.561)
5th
R (26LAPS)
10th
R (22LAPS)
(49.965)
(1LAP)
11th (1'18.342)
16th (1'19.461)
14th (1'19.252)
18th (1'19.569)
5th
R (22LAPS )
R (44LAPS)
R (51LAPS)
(1LAP)
#5 R. SCHUMACHER
6th (1'14.132)
15th (1'14.940)
9th (1'14.564)
11th (1'14.713)
#1 M. SCHUMACHER
5th
10th
R (8LAPS)
8th
(42.812)
(1LAP)
(1LAP)
18th (1'32.591)
14th (1'31.999)
19th (1'32.968)
12th (1'31.906)
R (26LAPS)
16th
12th
9th
(2LAPS)
(1LAP)
(1LAP)
17th (1'21.636)
14th (1'21.337)
9th (1'21.130)
13th (1'21.274)
7th
R (50LAPS)
4th
5th
(1LAP)
(1LAP)
(1LAP)
DNQ
14th (1'13.542)
13th (1'13.506)
11th (1'13.457)
#6 J. MONTOYA
#1 M. SCHUMACHER
#6 J. MONTOYA
1'13.114
#1 M. SCHUMACHER
#5 R. SCHUMACHER
#3 D. COULTHARD
#1 M. SCHUMACHER
1'21.091
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#5 R. SCHUMACHER
#1 M. SCHUMACHER
1'16.364
#1 M. SCHUMACHER
#6 J. MONTOYA
#3 D. COULTHARD
#2 R. BARRICHELLO
1'08.082
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#6 J. MONTOYA
#6 J. MONTOYA
1'16.676
#6 J. MONTOYA
1'12.836
#3 D. COULTHARD
#1 M. SCHUMACHER
#3 D. COULTHARD
#2 R. BARRICHELLO
#6 J. MONTOYA
1'29.906
#2 R. BARRICHELLO
#1 M. SCHUMACHER
#4 K. RAIKKONEN
#6 J. MONTOYA
1'18.998
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#6 J. MONTOYA
#6 J. MONTOYA
1'11.985
#1 M. SCHUMACHER
Magny Cours
RACE
R (23LAPS)
R (35LAPS)
R (29LAPS)
4.251km72Laps=305.886km
(DELAY)
ROUND 12 GERMANY GP
QUALIFYING
6th (1'15.690)
12th (1'16.072)
11th (1'16.070)
7th (1'15.851)
Hockenheim
RACE
R (59LAPS)
8th
R (27LAPS)
R (39LAPS)
4.574km67Laps=306.458km
(DELAY)
(1LAP)
#5 R. SCHUMACHER
ROUND 13 HUNGARY GP
QUALIFYING
5th (1'14.880)
14th (1'15.804)
13th (1'15.583)
12th (1'15.556)
#2 R. BARRICHELLO
Budapest
RACE
6th
10th
R (20LAPS)
12th
3.975km77Laps=306.069km
(DELAY)
(68.804)
(1LAP)
(1LAP)
ROUND 14 BELGIUM GP
QUALIFYING
14th (1'46.508)
16th (1'46.875)
12th (1'46.403)
15th (1'46.553)
Spa Francorchamps
RACE
R (38LAPS)
11th
8th
12th
6.963km44Laps=306.355km
(DELAY)
(1LAP)
(79.855)
(DNF)
12th (1'22.515)
18th (1'23.166)
9th (1'22.126)
16th (1'22.645)
8th
12th
9th
6th
(70.891)
(1LAP)
(81.068)
(68.491)
9th (1'11.902)
15th (1'12.647)
7th (1'11.738)
12th (1'12.161)
7th
11th
6th
12th
(1LAP)
(1LAP)
(58.211)
(1LAP)
8th (1'33.276)
7th (1'33.090)
9th (1'33.349)
16th (1'34.192)
R (37LAPS)
5th
R (27LAPS)
R (8LAPS)
(82.694)
ROUND 15 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.719km
(DELAY)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
ROUND 17 JAPAN GP
QUALIFYING
Suzuka
RACE
5.821km53Laps=308.317km
(DELAY)
#3 D. COULTHARD
#1 M. SCHUMACHER
1'14.389
#2 R. BARRICHELLO
1'13.333
#1 M. SCHUMACHER
#6 J. MONTOYA
#1 M. SCHUMACHER
#5 R. SCHUMACHER
#1 M. SCHUMACHER
1'43.726
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#6 J. MONTOYA
#6 J. MONTOYA
1'20.264
#2 R. BARRICHELLO
#1 M. SCHUMACHER
#16 E. IRVINE
#1 M. SCHUMACHER
1'10.790
#2 R. BARRICHELLO
#1 M. SCHUMACHER
#3 D. COULTHARD
#1 M. SCHUMACHER
1'31.317
1 Raised from 19th to 18th position after a fuel infringement by E. Irvine resulted in his disqualification.
334
#4 K. RAIKKONEN
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#4 K. RAIKKONEN

P.
TEAM (CAR NO.)
TOTAL
AUS
MAL
BRA
SMR
ESP
AUT
MON
CAN
EUR
GBR
FRA
GER
HUN
BEL
ITA
USA
JPN
221
10
10
16
10
16
14
16
16
10
13
16
16
16
16
16
92
16
10
65
10
10
Mild Seven Renault F1 Team
23
Sauber Petronas
11
DHL Jordan Honda
Jaguar Racing
4
2
1
1
1
5
1
2
1
KL Minardi Asiatech
10
Panasonic Toyota Racing
11
Orange Arrows
4
1
1
1

P.
No.
TOTAL
AUS
MAL
BRA
SMR
ESP
AUT
MON
CAN
EUR
GBR
FRA
GER
HUN
BEL
ITA
USA
JPN
DRIVER (NAT)
M. SCHUMACHER (GER)
144
10
10
10
10
10
10
10
10
10
10
10
77
10
10
10
10
J. MONTOYA (COL)
50
R. SCHUMACHER (GER)
42
D. COULTHARD (GBR)
41
K. RAIKKONEN (FIN)
24
15
J. BUTTON (GBR)
14
14
J. TRULLI (ITA)
16
E. IRVINE (GBR)
10
N. HEIDFELD (GER)
6
6
6
10
6
6
10
3
6
3
2
3
11
G. FISICHELLA (ITA)
12
11
J. VILLENEUVE (CAN)
13
F. MASSA (BRA)
14
12
O. PANIS (FRA)
15
10
T. SATO (JPN)
16
23
M. WEBBER (AUS)
17
24
M. SALO (FIN)
18
20
H. FRENTZEN (GER)
19
25
A. McNISH (GBR)
20
22
A. YOONG (MAS)
21
17
P. DE LA ROSA (ESP)
22
21
E. BERNOLDI (BRA)
23
22
A. DAVIDSON (GBR)
3
1
1
2
1
1
3
1
3
2
1
2
1
2
1
1
335

#16 J. VILLENEUVE
#17 J. BUTTON
6th (1'28.420)
8th (1'28.682)
9th
10th
(65.536)
(65.974)
12th (1'38.289)
9th (1'38.073)
R (0LAP)
7th
(1LAP)
13th (1'14.668)
11th (1'14.504)
POLE
PODIUM
POSITION
(FIRST 3)
#5 D. COULTHARD
QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
ROUND 3 BRAZIL GP
QUALIFYING
Interlagos
RACE1
6th
R (32LAPS)
4.309km54Laps=232.656km
(DELAY)
(17.910)
ROUND 4 SAN MARINO GP
QUALIFYING
7th (1'23.160)
9th (1'23.381)
Imola
RACE
R (19LAPS)
8th
4.993km62Laps=305.609km
(DELAY)
(1LAP)
#2 R. BARRICHELLO
ROUND 5 SPAIN GP
QUALIFYING
11th (1'19.563)
5th (1'18.704)
#1 M. SCHUMACHER
Barcelona
RACE
R (12LAPS)
9th
4.730km65Laps=307.324km
(DELAY)
(2LAPS)
12th (1'10.618)
7th (1'09.935)
12th
4th
(1LAP)
(42.243)
11th (1'16.755)
- 2
R (63LAPS)
#1 M. SCHUMACHER
14th (1'17.347)
17th (1'18.205)
#1 M. SCHUMACHER
R (14LAPS)
R (51LAPS)
17th (1'34.596)
12th (1'33.395)
ROUND 6 AUSTRIA GP
QUALIFYING
Spielberg
RACE
4.326km69Laps=298.494km
(DELAY)
ROUND 7 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.340km78Laps=260.520km
(DELAY)
ROUND 8 CANADA GP
QUALIFYING
Montreal
RACE
4.361km70Laps=305.270km
(DELAY)
ROUND 9 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.148km60Laps=308.863km
(DELAY)
ROUND 10 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)

Silverstone
RACE
5.141km60Laps=308.355km
(DELAY)
ROUND 12 GERMANY GP
QUALIFYING
Hockenheim
RACE
4.574km67Laps=306.458km
(DELAY)
ROUND 13 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.663km
(DELAY)
ROUND 14 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.720km
(DELAY)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
#8 F. ALONSO
1'37.044
#6 K. RAIKKONEN
#2 R. BARRICHELLO
#8 F. ALONSO
#2 R. BARRICHELLO
1'13.807
#11 G. FISICHELLA1
#6 K. RAIKKONEN
#8 F. ALONSO
#1 M. SCHUMACHER
1'22.327
#1 M. SCHUMACHER
1'17.762
#1 M. SCHUMACHER
#6 K. RAIKKONEN
#8 F. ALONSO
#2 R. BARRICHELLO
#1 M. SCHUMACHER
1'09.150
#1 M. SCHUMACHER
#6 K. RAIKKONEN
#2 R. BARRICHELLO
#4 R. SCHUMACHER
1'15.259
#4 R. SCHUMACHER
1'15.529
#3.J. MONTOYA
#6 K. RAIKKONEN
#4 R. SCHUMACHER
#3. J. MONTOYA
7th
(1LAP)
12th (1'16.990)
14th (1'17.077)
9th
R (21LAPS)
(1LAP)
20th (
#3 J. MONTOYA
#6 K. RAIKKONEN
R (51LAPS)
9th (1'22.591)
#1 M. SCHUMACHER
1'27.173
#6 K. RAIKKONEN
1'31.523
#4 R. SCHUMACHER
#3 J. MONTOYA
#2 R. BARRICHELLO
#4 R. SCHUMACHER
1'15.019
#4 R. SCHUMACHER
#3. J. MONTOYA
#1 M. SCHUMACHER
#2 R. BARRICHELLO
1'21.209
#2 R. BARRICHELLO
10th
8th
(63.569)
(45.478)
#3 J. MONTOYA
13th (1'17.090)
17th (1'18.085)
9th
8th
(2LAPS)
(1LAP)
16th (1'24.100)
14th (1'23.847)
R (14LAPS)
10th
(1LAPS)
10th (1'22.717)
7th (1'22.301)
6th
R (24LAPS)
(1LAP)
#2 R. BARRICHELLO
12th (1'13.050)
11th (1'12.695)
#1 M. SCHUMACHER
R (63LAPS)
R (41LAPS)
#6 K. RAIKKONEN
#3 J. MONTOYA
1'15.167
#3 J. MONTOYA
#5 D. COULTHARD
#7 J. TRULLI
#8 F. ALONSO
1'21.688
#8 F. ALONSO
#6 K. RAIKKONEN
#3 J. MONTOYA
#1 M. SCHUMACHER
1'20.963
#6 K. RAIKKONEN
1'11.670
#1 M. SCHUMACHER
#3 J. MONTOYA
#6 K. RAIKKONEN
#10 H. FRENTZEN
#16 T. SATO
ROUND 16 JAPAN GP
QUALIFYING
Suzuka
RACE
5.807km53Laps=307.573km
(DELAY)
For 2003, QF2 times are recorded
13th (1'33.924)
9th (1'33.474)
6th
4th
(51.692)
(33.106)
#2 R. BARRICHELLO
1'31.713
#2 R. BARRICHELLO
#6 K. RAIKKONEN
#5 D. COULTHARD
1 The red flag was displayed during lap 56, and the race was ended. Because more than 75% of the race had been completed, times at
lap 54, two laps before the red flag was displayed, were used as the official times.
2 J. Button was involved in an accident during the Saturday morning free practice session, and was ordered not to participate in the race
by a doctor.
336

P.
TEAM (CAR NO.)
TOTAL
AUS
MAL
158
11
144
142
16
88
26
Sauber Petronas
19
Jaguar Racing
18
16
Jordan Ford
13
10
European Minardi Cosworth
BRA
RSM
ESP
AUT
MON
CAN
EUR
FRA
GBR
GER
HUN
ITA
USA
JPN
16
16
16
14
10
15
16
10
11
15
14
18
18
10
11
12
10
13
12
12
10
10
10
12
11
12
10
14
4
8
10
3
1
3
2

P.
No.
TOTAL
AUS
MAL
M. SCHUMACHER (GER)
DRIVER (NAT)
93
K. RAIKKONEN (FIN)
91
10
J. MONTOYA (COL)
82
65
R. SCHUMACHER (GER)
58
5
6
F. ALONSO (ESP)
55
D. COULTHARD (GBR)
51
10
J. TRULLI (ITA)
33
17
J. BUTTON (GBR)
17
10
14
M. WEBBER (AUS)
17
10
H. FRENTZEN (GER)
13
12
11
G. FISICHELLA (ITA)
12
13
21
C. DA MATTA (BRA)
10
14
N. HEIDFELD (GER)
15
20
O. PANIS (FRA)
16
16
J. VILLENEUVE (CAN)
17
M. GENE (ESP)
18
16
T.SATO (JPN)
19
12
R. FIRMAN (IRL)
20
15
J. WILSON (GBR)
21
15
A. PIZZONIA (BRA)
22
19
J. VERSTAPPEN (NED)
23
18
N. KIESA (DEN)
24
12
Z. BAUMGARTNER (HUN)
RSM
ESP
AUT
MON
CAN
EUR
FRA
GBR
GER
HUN
ITA
USA
10
10
10
10
10
10
8
2
10
10
10
10
11
BRA
5
2
2
2
10
10
5
5
10
1
6
5
2
6
2
3
1
1
3
4
5
10
1
JPN
4
1
4
3
4
3
1
1
337

QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
ROUND 3 BAHRAIN GP
QUALIFYING
Bahrain
RACE
5.417km57Laps=308.523km
(DELAY)
#9 J. BUTTON
#10 T. SATO
4th (1'24.998)
7th (1'25.851)
6th
9th
(70.598)
(1LAP)
6th (1'34.221)
20th ( - )
3th
15th
(11.568)
(DNF)
6th (1'30.856)
5th (1'30.827)
POLE
PODIUM
POSITION
(FIRST 3)
#1 M. SCHUMACHER
1'24.408
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#8 F. ALONSO
#1 M. SCHUMACHER
1'33.074
#1 M. SCHUMACHER
#3 J. MONTOYA
#9 J. BUTTON
#1 M. SCHUMACHER
1'30.139
#1 M. SCHUMACHER
3th
5th
(26.687)
(52.460)
PP (1'19.753)
7th (1'20.913)
2th
16th
(9.702)
(DNF)
14th (1'17.575)
3th (1'15.809)
8th
5th
(1LAP)
(42.327)
3th (1'14.396)1
8th (1'14.827)1
2th
R (2LAPS)
(0.497)
#2 R. BARRICHELLO
5th (1'29.245)
2th (1'28.986)
#1 M. SCHUMACHER
3th
R (47LAPS)
(22.533)
2th (1'12.341)
17th (1'17.004)
3th2
R (48LAPS)
(20.409)
4th (1'10.820)
3th (1'10.601)
Indianapolis
RACE
R (26LAPS)
3th
4.192km73Laps=306.016km
(DELAY)
(22.036)
4th (1'13.995)
7th (1'14.240)
5th
R (15LAPS)
(32.484)
3th (1'18.580)
9th (1'19.688)3
4th
11th

Imola
RACE
4.933km62Laps=305.609km
(DELAY)
ROUND 5 SPAIN GP
QUALIFYING
Barcelona
RACE
4.627km66Laps=305.256km
(DELAY)
ROUND 6 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.340km77Laps=257.180km
(DELAY)
ROUND 7 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.148km60Laps=308.863km
(DELAY)
ROUND 8 CANADA GP
QUALIFYING
Montreal
RACE
4.361km70Laps=305.270km
(DELAY)
ROUND 10 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)

Silverstone
RACE
5.141km60Laps=308.355km
(DELAY)
ROUND 12 GERMANY GP
QUALIFYING
Hockenheim
RACE
4.574km66Laps=301.884km
(DELAY)
ROUND 13 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.663km
(DELAY)
ROUND 14 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
6.976km44Laps=306.927km
(DELAY)
ROUND 15 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.720km
(DELAY)
ROUND 16 CHINA GP
QUALIFYING
Shanghai
RACE
5.451km56Laps=305.066km
(DELAY)
ROUND 17 JAPAN GP
QUALIFYING
Suzuka
RACE
5.807km53Laps=307.573km
(DELAY)
ROUND 18 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
#2 R. BARRICHELLO
#9 J. BUTTON
#9 J. BUTTON
1'19.753
#1 M. SCHUMACHER
#9 J. BUTTON
#3 J. MONTOYA
#1 M. SCHUMACHER
1'15.022
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#7 J. TRULLI
#7 J. TRULLI
1'13.985
#1 M. SCHUMACHER
1'28.351
#7 J. TRULLI
#9 J. BUTTON
#2 R. BARRICHELLO
#9 J. BUTTON
#4 R. SCHUMACHER
1'12.275
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#9 J. BUTTON2
#2 R. BARRICHELLO
1'10.223
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#10 T. SATO
#8. F. ALONSO
1'13.698
#1 M. SCHUMACHER
#8. F. ALONSO
#2 R. BARRICHELLO
#6 K.RAIKKONEN
1'18.233
#1 M. SCHUMACHER
#6 K.RAIKKONEN
#2 R. BARRICHELLO
(10.683)
(33.736)
3th (1'13.674)4
9th (1'14.287)4
2th
8th
(8.388)
(46.842)
4th (1'19.700)
3th (1'19.693)
5th
6th
(67.439)
(1LAP)
12th (2'00.237)
15th (2'01.813)
R (29LAPS)
R (0LAP)
#2 R. BARRICHELLO
6th (1'20.786)
5th (1'20.715)
#2 R. BARRICHELLO
3th
4th
(10.197)
(15.370)
3th (1'34.295)
9th (1'34.993)
2th
6th
(1.035)
(54.791)
5th (1'35.157)
4th (1'34.897)
3th
4th
(19.662)
(31.781)
5th (1'11'092)
6th (1'11.120)
R (3LAPS)
6th
(50.248)
#1 M. SCHUMACHER
1'13.306
#1 M. SCHUMACHER
#9 J. BUTTON
#8 F. ALONSO
#1 M. SCHUMACHER
1'19.146
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#8 F. ALONSO
#7 J. TRULLI
1'56.232
#2 R. BARRICHELLO
1'20.089
#6 K.RAIKKONEN
#1 M. SCHUMACHER
#1 M. SCHUMACHER
#9 J. BUTTON
#2 R. BARRICHELLO
1'34.012
#2 R. BARRICHELLO
#9 J. BUTTON
#6 K.RAIKKONEN
#1 M. SCHUMACHER
1'33.542
#1 M. SCHUMACHER
#4 R. SCHUMACHER
#9 J. BUTTON
#2 R. BARRICHELLO
1'10.646
#3. J. MONTOYA
#6 K. RAIKKONEN
#2 R. BARRICHELLO
1 As a result of an engine replacement penalty, R. Schumacher, who qualified in 2nd position, was demoted 10 grid positions, and J.
Button and T. Sato took the 2nd and 7th positions on the grid, respectively.
2 J. Button was raised from 4th to 3rd position as a result of R. Schumachers disqualification due to a brake duct size infringement.
3 Due to an engine replacement, F. Alonso, who qualified in 6th position, was demoted 10 grid positions; as a result T. Sato moved from 9th
to 8th position on the grid.
4 J. Button qualified 3rd, but was demoted 10 grid positions due to an engine replacement. He started the race in 13th position on the grid,
while T. Sato was raised from 9th to 8th position on the grid.
338

P.
TEAM (CAR NO.)
CAN
USA
FRA
GBR
TOTAL AUS MAL BRN SMR

262
18
15
18
13
18
18
18
18
16
16
10
18
119
10
105
11
10
13
88
69
Sauber Petronas
34
ESP
MON EUR
Jaguar Racing
10
1
4
Jordan Ford
10
Minardi Cosworth
GER HUN
10
1
3
BEL
ITA
CHN
JPN
14
18
10
10
11
11
11
BRA
10
14
12
CHN
JPN
BRA
10
3
1

P.
No.
DRIVER (NAT)
M. SCHUMACHER (GER)

148
10
10
10
10
10
114
J. BUTTON (GBR)
85
F. ALONSO (ESP)
59
J. MONTOYA (COL)
58
J. TRULLI (ITA)
46
K. RAIKKONEN (FIN)
45
10
T. SATO (JPN)
34
R. SCHUMACHER (GER)
24
10
D. COULTHARD (GBR)
24
11
11
G. FISICHELLA (ITA)
22
12
12
F. MASSA (BRA)
12
13
14
M. WEBBER (AUS)
ESP
MON EUR
CAN
USA
FRA
GBR
10
10
10
10
10
10
6
5
1
4
2
4
4
GER HUN
10
BEL
ITA
10
10
10
8
5
10
10
5
3
2
14
17
O. PANIS (FRA)
15
A. PIZZONIA (BRA)
16
15
C. KLIEN (AUT)
17
16
C. DA MATTA (BRA)
18
18
N. HEIDFELD (GER)
19
19
T. GLOCK (GER)
20
21
Z. BAUMGARTNER (HUN)
21
J. VILLENEUVE (CAN)
22
17
R. ZONTA (BRA)
23
M. GENE (ESP)
24
19
G. PANTANO (ITA)
25
20
G. BRUNI (ITA)
5
2
1
2
2
3
1
2
1
339
2
1
6
6
1
1

#3 J. BUTTON
QUALIFYING
Melbourne
RACE
5.303km57Laps=302.271km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
#4 T. SATO
8th (3'12.128)1
- 1
11th
14th
(DNF)
(DNF)
POLE
PODIUM
POSITION
(FIRST 3)
#6 G. FISICHELLA
3'01.4601
#6 G. FISICHELLA
#2 R. BARRICHELLO
#5 F. ALONSO
#4 A. DAVIDSON
9th (3'09.832)1
15th (3'11.890)1
R (2LAPS)
R (2LAPS)
#5 F. ALONSO
3'07.6721
#5 F. ALONSO
#16 J. TRULLI
#8 N. HEIDFELD
#4 T. SATO
ROUND 3 BAHRAIN GP
QUALIFYING
Bahrain
RACE
5.412km57Laps=308.238km
(DELAY)

Imola
RACE
4.933km62Laps=305.609km
(DELAY)
ROUND 5 SPAIN GP
QUALIFYING
Barcelona
RACE
4.627km66Laps=305.256km
(DELAY)
ROUND 6 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.340km78Laps=260.520km
(DELAY)
ROUND 7 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.148km59Laps=303.715km
(DELAY)
11th (3'04.348)1
13th (3'05.563)1
R (46LAPS)
R (27LAPS)
3th (2'44.105)1
6th (2'44.658)1
D 2
D 2
- 3
- 3
#5 F. ALONSO
#5 F. ALONSO
3'01.9021
#16 J. TRULLI
#9 K. RAIKKONEN
#9 K. RAIKKONEN
2'42.8801
#5 F. ALONSO
#1 M. SCHUMACHER
#10 A. WURZ
#9 K. RAIKKONEN
2'31.4211
#9 K. RAIKKONEN
#5 F. ALONSO
#16 J. TRULLI
- 3
- 3
#9 K. RAIKKONEN
2'30.3231
#9 K. RAIKKONEN
#8 N. HEIDFELD
#7 M. WEBBER
4
13th (1'32.594)
16th (1'32.926)
10th
12th
(95.786)
(1LAPS)
#8 N. HEIDFELD
1'30.081
#5 F. ALONSO
#8 N. HEIDFELD
#2 R. BARRICHELLO
ROUND 8 CANADA GP
QUALIFYING
PP (1'15.217)
6th (1'15.729)
Montreal
RACE
R (46LAPS)
R (40LAPS)
4.361km70Laps=305.270km
(DELAY)
#2 R. BARRICHELLO
3th (1'11.277)
8th (1'11.497)
#1 M. SCHUMACHER
R (0LAPS)
R (0LAPS)
8th ( 1'15.051)
5th (1'14.655)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
ROUND 10 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)

Silverstone
RACE
5.141km60Laps=308.355km
(DELAY)
ROUND 12 GERMANY GP
QUALIFYING
Hockenheim
RACE
4.574km67Laps=306.458km
(DELAY)
ROUND 13 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.663km
(DELAY)
ROUND 14 TURKEY GP
QUALIFYING
Istanbul
RACE
5.338km58Laps=309.396km
(DELAY)
ROUND 15 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.720km
(DELAY)
ROUND 16 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
6.976km44Laps=306.927km
(DELAY)
ROUND 17 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
ROUND 18 JAPAN GP
QUALIFYING
Suzuka
RACE
5.807km53Laps=307.573km
(DELAY)
ROUND 19 CHINA GP
QUALIFYING
Shanghai
RACE
5.451km56Laps=305.066km
(DELAY)
4th
11th
(1LAP)
(1LAP)
3th (1'20.207)5
8th (1'21.114)5
5th
16th
(40.264)
(2LAPS)
2th (1'14.759)
8th (1'15.501)
3th
12th
(24.422)
(1LAP)
8th (1'21.302)
10th (1'21.787)
5th
8th
(58.832)
(1LAP)
13th (1'30.063)
- 6
5th
9th
(39.304)
(109.987)
4th (1'21.369)
5th (1'21.477)
8th
16th
(63.635)
(1LAP)
9th (1'47.978)
11th (1'48.353)
3th
R (13LAPS)
(32.077)
4th (1'12.696)
7th
10th
(1LAP)
(1LAP)
2th (1'46.141)
5th (1'46.841)
5th
- 7
(29.507)
4th (1'34.801)
17th (1'37.083)
8th
R (34LAPS)
(41.249)
1
2
3
4
5
#3 J. BUTTON
1'15.217
#16 J. TRULLI
1'10.625
#9 K. RAIKKONEN
#1 M. SCHUMACHER
#2 R. BARRICHELLO
#18 T. MONTEIRO
#5 F. ALONSO
1'14.412
#5 F. ALONSO
#9 K. RAIKKONEN
#1 M. SCHUMACHER
#5 F. ALONSO
1'19.905
#10 J. MONTOYA
#5 F. ALONSO
#9 K. RAIKKONEN
#9 K. RAIKKONEN
1'14.320
#5 F. ALONSO
#10 J. MONTOYA
#3 J. BUTTON
#1 M. SCHUMACHER
1'19.882
#9 K. RAIKKONEN
#1 M. SCHUMACHER
#17 R. SCHUMACHER
#9 K. RAIKKONEN
1'26.797
#9 K. RAIKKONEN
#5 F. ALONSO
#10 J. MONTOYA
#9 K. RAIKKONEN
1'20.878
#10 J. MONTOYA
#5 F. ALONSO
#6 G. FISICHELLA
#10 J. MONTOYA
1'46.391
#9 K. RAIKKONEN
#5 F. ALONSO
#3 J. BUTTON
#5 F. ALONSO
1'11.988
#10 J. MONTOYA
#9 K. RAIKKONEN
#5 F. ALONSO
#17 R. SCHUMACHER
1'46.106
#9 K. RAIKKONEN
#6 G. FISICHELLA
#5 F. ALONSO
#5 F. ALONSO
1'34.080
#5 F. ALONSO
#9 K. RAIKKONEN
#17 R. SCHUMACHER
From the opening round to Round 6 (Monaco), the aggregate of QF1 and QF2 times was used as the qualifying time.
Disqualified due to vehicle weight infringement.
Suspended from racing due to vehicle weight infringement at San Marino GP.
From Round 7, rule changes specified a single Saturday qualifying run.
K. Rikknen, who was qualified in 2nd position, was demoted 10 grid positions due to an engine replacement penalty, raising J. Button
from 3rd to 2nd grid position and T. Sato from 8th to 7th grid position.
6 Qualifying times were deleted as a result of impeding another vehicle.
7 Eliminated from results due to contact with another vehicle.
340

P.
TEAM (CAR NO.)

191
16
10
10
10
12
13
Team McLaren Mercedes
182
10
12
14
10
100
10
14
88
66
38
ESP MON EUR CAN USA FRA GBR GER HUN TUR
12
13
11
14
18
13
13
15
16
10
11
CHN
14
10
14
15
16
15
10
18
10
5
5
JPN
13
Red Bull Racing
34
Sauber Petronas
20
Jordan Grand Prix
12
11
10
Minardi F1 Team
ITA
BEL BRA
4
3

P.
No. DRIVER (NAT)
F. ALONSO (ESP)
133
K. RAIKKONEN (FIN)
112
M. SCHUMACHER (GER)
62
10
J. MONTOYA (COL)
60
G. FISICHELLA (ITA)
58
10
17
R. SCHUMACHER (GER)
45
16
J. TRULLI (ITA)
43
38
J. BUTTON (GBR)
37
10
M. WEBBER (AUS)
36
11
N. HEIDFELD (GER)
28
12
14
D. COULTHARD (GBR)
24
12
F. MASSA (BRA)
11
14
11
J. VILLENEUVE (CAN)
15
16
17
15/37 C. KLIEN (AUT)

18
T. MONTEIRO (POR)
10/35 A. WURZ (AUT)
10
10
6
2
13
10
ESP MON EUR CAN USA FRA GBR GER HUN TUR
8
10
10
2
2
4
4
5
4
10
4
10
3
1
3
6
3
1
6
8
2
2
T. SATO (JPN)
24
25
15/37 V. LIUZZI (ITA)

20
R. DOORNBOS (NED)
10
10
10
10
10
2
4
3
3
3
4
1
4
3
2
1
341
10
BEL BRA
23
A. PIZZONIA (BRA)
10
C. ALBERS (NED)
CHN
21
22
ITA
19
JPN
P. FRIESACHER (AUT)
10
N. KARTHIKEYAN (IND)
20
19
10/35 P. DE LA ROSA (ESP)
10
4
6
18
20
21
10
2
3
10

#11 R. BARRICHELLO #12 J. BUTTON
ROUND 1 BAHRAIN GP
QUALIFYING
Bahrain
RACE
5.412km57Laps=308.238km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
QUALIFYING
Melbourne
RACE
5.303km57Laps=302.271km
(DELAY)
RACE
4.959km62Laps=307.221km
(DELAY)
#23 Y. IDE
6th (1'32.579)
3th (1'31.549)
20th (1'37.411)
21th (1'40.270)
15th
4th
18th
R (35LAPS)
(1LAP)
(19.992)
(4LAPS)
12th (1'34.683)
2th (1'33.986)
21th (1'39.011)
22th (1'40.720)
10th
3th
14th
R (33LAPS)
(1LAP)
(9.631)
(3LAPS)
17th (1'29.943)
PP (1'25.229)
21th (1'32.279)
22th (1'36.164)
7th
10th
12th
13th
(51.904)
DNF
(2LAPS)
(3LAPS)
3th (1'23.242)
2th (1'22.988)
21th (1'27.609)
22th (1'29.282)
10th
7th
R (44LAPS)
R (23LAPS)
(77.851)
(39.635)

Imola
Super Aguri F1
#22 T. SATO
POLE
QF FASTEST
PODIUM
POSITION
LAP TIME
(FIRST 3)
#1 F. ALONSO
#5 M. SCHUMACHER #5 M. SCHUMACHER
#5 M. SCHUMACHER
1'31.431
1'31.431
#3 K. RAIKKONEN
#2 G. FISICHELLA
1'33.840
#12 J. BUTTON
1'33.527
#2 G. FISICHELLA
#1 F. ALONSO
#12 J. BUTTON
#12 J. BUTTON
1'25.229
#12 J. BUTTON
1'25.229
#1 F. ALONSO
#3 K. RAIKKONEN
#7 R. SCHUMACHER
#5 M. SCHUMACHER
#1 F. ALONSO
1'22.795
1'22.579
#4 J. MONTOYA
#23 F. MONTAGNY
ROUND 5 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.148km60Laps=308.863km
(DELAY)
ROUND 6 SPAIN GP
QUALIFYING
Barcelona
RACE
4.627km66Laps=305.256km
(DELAY)
ROUND 7 MONACO GP
QUALIFYING
Monte Carlo
RACE
3.340km78Laps=260.520km
(DELAY)
QUALIFYING
Silverstone
RACE
5.141km60Laps=308.355km
(DELAY)
ROUND 9 CANADA GP
4th (1'30.754)
6th (1'30.940)
21th (1'35.239)
22th (1'46.505)
5th
R (28LAPS)
R (45LAPS)
R (29LAPS)
(72.586)
5th (1'15.885)
8th (1'16.008)
20th (1'18.920)
21th (1'20.763)
7th
6th
17th
R (10LAPS)
(1LAP)
(58.347)
(4LAPS)
- 1
13th (1'14.982)2 19th (1'17.276)2 20th (1'17.502)2
4th
11th
R (46LAPS)
16th
(53.337)
(1LAP)
(3LAPS)
6th (1'20.943)
19th (1'23.247)
20th (1'26.158)
21th (1'26.316)
10th
R (8LAPS)
17th
18th
(1LAP)
(3LAPS)
(3LAPS)
QUALIFYING
9th (1'16.912)
8th (1'16.608)
21th (1'19.088)
22th (1'19.152)
Montreal
RACE
R (11LAPS)
9th
15th
R (2LAPS)
4.361km70Laps=305.270km
(DELAY)
(1LAP)
(DNF)
4th (1'12.109)
7th (1'12.523)
18th (1'13.496)
21th (1'16.036)
R (6LAPS)
R (0LAP)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
ROUND 11 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)
6th
R (3LAPS)
(36.516)
14th (1'17.027)
19th (1'17.495)
22th (1'18.845)
21th (1'18.637)
R (18LAPS)
R (61LAPS)
R (0LAP)
16th
(3LAPS)
#1 F. ALONSO
1'29.819
#1 F. ALONSO
1'29.819
#5 M. SCHUMACHER
#1 F. ALONSO
#6 F. MASSA
#1 F. ALONSO
1'14.648
#5 M. SCHUMACHER #1 F. ALONSO
1'14.637
#5 M. SCHUMACHER
#2 G. FISICHELLA
#1 F. ALONSO
1'13.9622
#3 K. RAIKKONEN #1 F. ALONSO
1'13.532
#4 J. MONTOYA
#14 D. COULTHARD
#1 F. ALONSO
1'20.253
#1 F. ALONSO
1'20.253
#1 F. ALONSO
#5 M. SCHUMACHER
#3 K. RAIKKONEN
#1 F. ALONSO
1'14.942
#1 F. ALONSO
1'14.726
#1 F. ALONSO
#5 M. SCHUMACHER
#3 K. RAIKKONEN
#5 M. SCHUMACHER
1'10.832
1'10.636
#6 F. MASSA
#2 G. FISICHELLA
#5 M. SCHUMACHER
1'15.493
1'15.111
#1 F. ALONSO
#6 F. MASSA
#23 S. YAMAMOTO
ROUND 12 GERMANY GP
QUALIFYING
Hockenheim
RACE
4.574km67Laps=306.458km
(DELAY)
ROUND 13 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.663km
(DELAY)
ROUND 14 TURKEY GP
QUALIFYING
Istanbul
RACE
5.338km58Laps=309.396km
(DELAY)
ROUND 15 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.720km
(DELAY)
ROUND 16 CHINA GP
QUALIFYING
Shanghai
RACE
5.451km56Laps=305.066km
(DELAY)
ROUND 17 JAPAN GP
QUALIFYING
Suzuka
RACE
5.807km53Laps=307.573km
(DELAY)
ROUND 18 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
6th (1'14.934)
4th (1'14.862)
19th (1'17.185)
21th (1'20.444)
R (18LAPS)
4th
R (38LAPS)
R (1LAP)
(18.898)
3th (1'20.085)
4th (1'20.092)3
19th (1'22.967)
22th (1'24.016)
4th
WIN
13th
R (0LAP)
(45.205)
(5LAPS)
14th (1'28.257)
7th (1'27.790)
22th (1'30.850)
21th (1'30.607)
8th
4th
R (41LAPS)
R (23LAPS)
(60.034)
(12.334)
8th (1'22.787)4
5th (1'22.011)4
21th (1'24.289)
22th (1'26.001)
6th
5th
16th
R (18LAPS)
(42.409)
(32.685)
(2LAPS)
3th (1'45.503)
4th (1'45.503)
6th
4th
- 6
16th7
(79.131)
(72.056)
(4LAPS)
8th (1'31.478)
7th (1'30.992)
20th (1'33.666)
1'13.778
#6 F. MASSA
#3 K. RAIKKONEN
#3 K. RAIKKONEN
1'19.599
#6 F. MASSA
1'19.504
#12 J. BUTTON
#4 P. DE LA ROSA
#16 N. HEIDFELD
20th (1'50.326)5 21th (1'55.560)5
12th
4th
15th
17th
(1LAP)
(34.101)
(1LAP)
(3LAPS)
5th (1'11.619)
14th (1'11.742)
20th (1'13.269)
21th (1'13.357)
7th
3th
10th
16th
(40.294)
(19.394)
(1LAP)
(2LAPS)
1
2
3
4
#3 K. RAIKKONEN
1'14.070
#6 F. MASSA
1'26.907
#5 M. SCHUMACHER #6 F. MASSA
1'25.850
#1 F. ALONSO
#5 M. SCHUMACHER
#3 K. RAIKKONEN
1'21.484
#6 F. MASSA
1'21.225
#5 M. SCHUMACHER
#3 K. RAIKKONEN
#17 R. KUBICA
#1 F. ALONSO
1'44.360
#1 F. ALONSO
1'43.951
#5 M. SCHUMACHER
#1 F. ALONSO
#2 G. FISICHELLA
#6 F. MASSA
1'29.599
#5 M. SCHUMACHER #1 F. ALONSO
1'28.954
#6 F. MASSA
#2 G. FISICHELLA
#6 F. MASSA
1'10.680
#5 M. SCHUMACHER #6 F. MASSA
1'10.313
#1 F. ALONSO
#12 J. BUTTON
Qualifying times deleted for a judgment of intentionally impeding another vehicle.

M. Schumachers qualifying times were deleted for intentionally impeding another vehicle in Q3, raising J. Button one position.
Due to an engine replacement penalty, J. Button was demoted 10 grid positions, from 4th to 14th position on the grid.
F. Alonso was qualified in 5th position, but his best three times in Q3 were deleted for impeding F. Massa. This raised R. Barrichello
from 9th to 8th position and J. Button from 6th to 5th position.
5 R. C. Albers qualifying times were deleted for failing to respond to a request for vehicle scrutinizing. As a result, T. Sato was raised
from 21st to 20th position, and S. Yamamoto from 22nd to 21st position.
6 Eliminated from race results for ignoring a blue flag.
7 As a result of 6, S. Yamamoto was raised from 17th to 16th position.
342

P.
TEAM (CAR NO.)
TOTAL BRN MAL AUS

206
10
18
201
Team McLaren Mercedes
110
10
86
BMW Sauber F1 Team
36
35
Red Bull Racing
16
Williams F1 Team
11
Scuderia Toro Rosso
10
Spyker M F1 Team
11
Super Aguri Formula 1
SMR
EUR
ESP
USA
FRA
11
16
13
15
15
10
11
15
16
13
12
12
18
16
18
10
14
MON GBR CAN
6
3
ITA
CHN
JPN
11
14
16
11
16
10
10
15
15
GER HUN TUR
6
2
BRA
5
1

P.
No. DRIVER (NAT)
TOTAL BRN MAL AUS
F. ALONSO (ESP)
134
10
M.SCHUMACHER (GER)
121
F. MASSA (BRA)
80
G. FISICHELLA (ITA)
72
10
K. RAIKKONEN (FIN)
65
6
5
12
J. BUTTON (GBR)
56
11
30
J. MONTOYA (COL)
26
16
N. HEIDFELD (GER)
23
10
R. SCHUMACHER (GER)
20
11
P. DE LA ROSA (ESP)
19
12
J. TRULLI (ITA)
15
13
14
D. COULTHARD (GBR)
14
14
M. WEBBER (AUS)
15
17
J. VILLENEUVE (CAN)
16
17/38 R. KUBICA (POR)
EUR
ESP
USA
FRA
10
10
10
10
10
10
10
10
10
10
5
1
15
C. KLIEN (AUT)
19
20
V. LIUZZI (ITA)
20
21
S. SPEED (USA)
21
18
T. MONTEIRO (POL)
22
19
C. ALBERS (NED)
23
22
T. SATO (JPN)
CHN
JPN
10
10
10
8
10
Y. IDE (JPN)
343
8
10
6
2
1
2
4
2
1
3
BRA
5
1
1
15/37 R. DOORNBOS (NED)
23/41 F. MONTAGNY
ITA
18
27
5
5
6
2
23
GER HUN TUR
3
2
23/41 S. YAMAMOTO (JPN)
N. ROSBERG (GER)
26
MON GBR CAN
3
4
10
25
SMR
17
24
10

QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
5.543km56Laps=310.408km
(DELAY)
ROUND 3 BAHRAIN GP
QUALIFYING
Bahrain
RACE
5.412km57Laps=308.238km
(DELAY)
ROUND 4 SPAIN GP
QUALIFYING
Barcelona
RACE
4.655km65Laps=302.449km
(DELAY)
#8 R. BARRICHELLO
#12 T. SATO
#14 A. DAVIDSON
14th (1'27.264)
17th (1'27.679)
10th (1'28.871)
11th (1'26.909)
15th
11th
12th
16th
(1LAP)
(1LAP)
(1LAP)
(2LAPS)
15th (1'36.088)
19th (1'36.827)
14th (1'35.945)
18th (1'36.816)
12th
11th
13th
16th
(1LAP)
(1LAP)
(1LAP)
(1LAP)
16th (1'33.731)
15th (1'33.624)
17th (1'33.984)
13th (1'33.082)
R (0LAP)
13th
R (34LAPS)
16th
(1LAP)
(DNF)
14th (1'22.120)
12th (1'22.097)
13th (1'22.115)
15th (DNF)
12th
10th
8th
11th
(1LAP)
(1LAP)
(1LAP)
(1LAP)
9th (1'17.498)1
21th (1'18.554)
17th (1'18.250)
11th
10th
17th
18th
(1LAP)
(1LAP)
(2LAPS)
(2LAPS)
15th (1'17.541)
13th (1'17.116)
11th (1'16.743)
17th (1'17.542)
R (0LAP)
12th
6th
11th
ROUND 5 MONACO GP
QUALIFYING 10th (1'17.939)1
Monte Carlo
RACE
3.340km78Laps=260.520km
(DELAY)
ROUND 6 CANADA GP
QUALIFYING
Montreal
RACE
4.361km70Laps=305.270km
(DELAY)

Indianapolis
RACE
4.192km73Laps=306.016km
(DELAY)
ROUND 8 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)

Silverstone
RACE
5.141km59Laps=303.214km
(DELAY)
ROUND 10 EUROPE GP
QUALIFYING
Nurburgring
RACE
5.148km60Laps=308.863km
(DELAY)
ROUND 11 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.663km
(DELAY)
ROUND 12 TURKEY GP
QUALIFYING
Istanbul
RACE
5.338km58Laps=309.396km
(DELAY)
ROUND 13 ITALY GP
QUALIFYING
Monza
RACE
5.793km53Laps=306.720km
(DELAY)
ROUND 14 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
7.004km44Laps=308.053km
(DELAY)
ROUND 15 JAPAN GP
QUALIFYING
Fuji
RACE
4.563km67Laps=305.416km
(DELAY)
ROUND 16 CHINA GP
QUALIFYING
Shanghai
RACE
5.451km56Laps=305.066km
(DELAY)
ROUND 17 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
Super Aguri F1 Team
#7 J. BUTTON
30.439
16.698
24.318
13th (1'12.998)
15th (1'13.201)
18th (1'13.477)
16th (1'13.259)
12th
R (0LAP)
R (13LAPS)
11th
(1LAP)
(1LAP)
12th (1'15.584)
13th (1'15.761)
19th (1'16.244)2
20th (1'16.366)
8th
11th
16th
R (1LAP)
58.885
(1LAP)
(2LAPS)
18th (1'21.335)
14th (1'20.364)
21th (1'22.045)
19th (1'21.448)
10th
9th
14th
R (35LAPS)
(1LAP)
(1LAP)
(2LAP)
17th (1'32.983)
14th (1'32.221)
16th (1'32.838)
15th (1'32.451)
R (2LAPS)
11th
R (19LAPS)
12th
(1LAP)
(1LAP)
17th (1'21.737)
18th (1'21.877)
19th (1'22.143)
15th (1'21.127)
R (35LAPS)
18th
15th
R (41LAPS)
(2LAPS)
(1LAP)
15th (1'28.220)
14th (1'28.188)
19th (1'28.953)
11th (1'28.002)
13th
17th
18th
14th
(1LAP)
(1LAP)
(1LAP)
(1LAP)
10th (1'25.165)
12th (1'23.176)
17th (1'23.749)
14th (1'23.274)
8th
10th
16th
14th
POLE
QF FASTEST
PODIUM
POSITION
LAP TIME
(FIRST 3)
#6 K. RAIKKONEN
1'26.072
#1 F. ALONSO
1'25.326
#6 K. RAIKKONEN
#1 F. ALONSO
#2 L. HAMILTON
#1 F. ALONSO
#5 F. MASSA
1'35.043
#1 F. ALONSO
1'34.057
#2 L. HAMILTON
#6 K. RAIKKONEN
#5 F. MASSA
#5 F. MASSA
1'32.652
#5 F. MASSA
1'31.359
#2 L. HAMILTON
#6 K. RAIKKONEN
#5 F. MASSA
#5 F. MASSA
1'21.421
#5 F. MASSA
1'20.597
#2 L. HAMILTON
#1 F. ALONSO
#1 F. ALONSO
#1 F. ALONSO
1'15.726
#1 F. ALONSO
1'15.431
#2 L. HAMILTON
#5 F. MASSA
#2 L. HAMILTON
#2 L. HAMILTON
1'15.707
#2 L. HAMILTON
1'15.486
#9 N. HEIDFELD
#17 A. WURZ
#2 L. HAMILTON
#2 L. HAMILTON
1'12.331
#1 F. ALONSO
1'11.926
#1 F. ALONSO
#5 F. MASSA
#6 K. RAIKKONEN
#5 F. MASSA
1'15.034
#2 L. HAMILTON
1'14.795
#5 F. MASSA
#2 L. HAMILTON
#6 K. RAIKKONEN
#2 L. HAMILTON
1'19.997
#1 F. ALONSO
1'19.152
#1 F. ALONSO
#2 L. HAMILTON
#1 F. ALONSO
#6 K. RAIKKONEN
1'31.450
#5 F. MASSA
1'30.912
#5 F. MASSA
#15 M. WEBBER
#1 F. ALONSO3
1'19.674
#2 L. HAMILTON
#2 L. HAMILTON
1'19.301
#6 K. RAIKKONEN
#9 N. HEIDFELD
#5 F. MASSA
#5 F. MASSA
1'27.329
#1 F. ALONSO
1'26.841
#6 K. RAIKKONEN
#1 F. ALONSO
#1 F. ALONSO
#1 F. ALONSO
1'21.997
#1 F. ALONSO
1'21.356
#2 L. HAMILTON
72.168
76.958
(1LAP)
(1LAP)
#6 K. RAIKKONEN
14th (1'46.955)
18th (1'47.954)
19th (1'47.980)
21th (1'48.199)
#6 K. RAIKKONEN
#6 K. RAIKKONEN
1'45.994
#6 K. RAIKKONEN
#5 F. MASSA
1'45.070
#1 F. ALONSO
R (36LAPS)
13th
15th
16th
(1LAP)
(1LAP)
(1LAP)
7th (1'26.913)
17th (1'27.323)
21th (1'28.792)
19th (1'27.564)
11th
10th
15th
R (54LAPS)
(DNF)
88.342
(DNF)
#6 K. RAIKKONEN
10th (1'39.285)
17th (1'37.251)
20th (1'38.218)
15th (1'37.247)
#6 K. RAIKKONEN
5th
15th
14th
R (11LAPS)
68.666
(1LAP)
(1LAP)
16th (1'13.469)
11th (1'12.932)
18th (1'14.098)
20th (1'14.596)
R (20LAPS)
R (40LAPS)
12th
14th
(2LAPS)
(3LAPS)
#2 L. HAMILTON
#2 L. HAMILTON
1'25.368
#2 L. HAMILTON
1'35.908
#2 L. HAMILTON
1'24.753
#4 H. KOVALAINEN
#6 K. RAIKKONEN
#1 F. ALONSO
1'35.381
#5 F. MASSA
#6 K. RAIKKONEN
#5 F. MASSA
1'11.931
#5 F. MASSA
1'11.931
#5 F. MASSA
#1 F. ALONSO
1 D. Coulthard was not allowed to start in Q3 and was penalized two positions for impeding another vehicle in Q2. As a result, J. Button
was raised from 11th to 10th position, and R. Barrichello from 10th to 9th position.
2 T. Sato was penalized 10 grid positions for overtaking under a yellow flag during the previous US Grand Prix, and started the race at
the back of the grid.
3 F. Alonso was demoted five grid positions as a result of impeding another vehicle in the qualifying session.
344

P.
TEAM (CAR NO.)
TOTAL
AUS
MAL
BRN
ESP
USA
FRA
GBR
EUR
HUN
TUR
ITA
BEL
JPN
CHN
BRA
204
13
10
16
10
MON CAN
7
11
18
14
18
18
16
18
BMW Sauber F1 Team
101
10
ING Renault F1 Team
51
12
AT&T Williams
33
Red Bull Racing
24
13
Scuderia Toro Rosso
Super Aguri F1 Team
10
Etihad Aldar Spyker F1 Team
11
Vodafone McLaren Mercedes
4
1
10
5
5
1
8
1
1
3
1

P.
No.
TOTAL
AUS
MAL
BRN
K. RAIKKONEN (FIN)
DRIVER (NAT)
110
10
L. HAMILTON (GBR)
109
F. ALONSO (ESP)
109
10
F. MASSA (BRA)
94
N. HEIDFELD (GER)
61
10
R. KUBICA (POL)
39
USA
FRA
GBR
HUN
TUR
ITA
BEL
JPN
CHN
BRA
10
10
10
10
10
10
10
10
10
10
10
10
10
10
H. KOVALAINEN (FIN)
30
G. FISICHELLA (ITA)
21
16
N. ROSBERG (GER)
20
10
14
D. COULTHARD (GBR)
14
11
17
A. WURZ (AUT)
13
12
15
M. WEBBER (AUS)
10
13
12
J. TRULLI (ITA)
14
19
S. VETTEL (GER)
15
J. BUTTON (GBR)
16
11
R. SCHUMACHER (GER)
17
22
T. SATO (JPN)
18
18
V. LIUZZI (ITA)
19
20
A. SUTIL (GER)
20
21
19
S. SPEED (USA)
22
17/38
K. NAKAJIMA (JPN)
23
23
A. DAVIDSON (GBR)
24
21
S. YAMAMOTO (JPN)
25
21
C. ALBERS (NED)
3
1
3
ESP
MON CAN
3
5
2
1
8
5
EUR
2
2
8
4
5
5
5
1
1
1
3
3
1
345
3
4
5
2
4
2
3
4
10
2
8

#7 J. BUTTON #8 R. BARRICHELLO
QUALIFYING
Melbourne
RACE
5.303km58Laps=307.574km
(DELAY)
ROUND 2 MALAYSIA GP
QUALIFYING
Kuala Lumpur
RACE
Super Aguri F1 Team

#12 T. SATO
#14 A. DAVIDSON
13th (1'26.259)
11th (1'26.173)
20th (1'28.208)
22th (1'29.059)
R (0LAP)
- 1
R (32LAPS)
R (0LAP)
11th (1'35.208)
14th (1'35.622)
20th (1'37.087)
22th (1'37.481)
10th
13th
16th
15th
5.543km56Laps=310.408km
(DELAY)
86.214
(1LAP)
(2LAPS)
(1LAP)
ROUND 3 BAHRAIN GP
QUALIFYING
9th (1'35.057)
12th (1'32.508)
22th (1'35.725)
21th (1'34.140)
Bahrain
RACE
R (19LAPS)
11th
17th
16th
5.412km57Laps=308.238km
(DELAY)
77.862
(1LAP)
(1LAP)
13th (1'21.211)
11th (1'21.049)
22th (1'23.496)
21th (1'23.318)
6th
R (34LAPS)
13th
R (8LAPS)
53.010
(1LAP)
13th (1'27.298)
12th (1'27.219)
11th
14th
ROUND 4 SPAIN GP
QUALIFYING
Barcelona
RACE
4.655km66Laps=307.104km
(DELAY)
ROUND 5 TURKEY GP
QUALIFYING
Istanbul
RACE
5.338km58Laps=309.396km
(DELAY)
ROUND 6 MONACO GP
QUALIFYING 12th (1'16.101)2
Monte Carlo
RACE
3.340km76Laps=253.840km
(DELAY)
ROUND 7 CANADA GP
QUALIFYING
Montreal
RACE
4.361km70Laps=305.270km
(DELAY)
ROUND 8 FRANCE GP
QUALIFYING
Magny Cours
RACE
4.411km70Laps=308.586km
(DELAY)
QUALIFYING
Silverstone
RACE
5.141km60Laps=308.355km
(DELAY)
ROUND 10 GERMANY GP
QUALIFYING
Hockenheim
RACE
4.574km67Laps=306.458km
(DELAY)
ROUND 11 HUNGARY GP
QUALIFYING
Budapest
RACE
4.381km70Laps=306.630km
(DELAY)
ROUND 12 EUROPE GP
QUALIFYING
Valencia
RACE
5.419km57Laps=308.883km
(DELAY)
ROUND 13 BELGIUM GP
QUALIFYING
Spa Francorchamps
RACE
7.004km44Laps=308.052km
(DELAY)
ROUND 14 ITALY GP
QUALIFYING
Monza
RACE
RA
5.793km53Laps=306.720km
(DELAY)
ROUND 15 SINGAPORE GP
QUALIFYING
Singapore
RACE
5.067km61Laps=308.950km
(DELAY)
ROUND 16 JAPAN GP
QUALIFYING
Fuji
RACE
4.563km67Laps=305.416km
(DELAY)
ROUND 17 CHINA GP
QUALIFYING
Shanghai
RACE
5.451km56Laps=305.066km
(DELAY)
ROUND 18 BRAZIL GP
QUALIFYING
Interlagos
RACE
4.309km71Laps=305.909km
(DELAY)
(1LAP)
POLE
QF FASTEST
PODIUM
POSITION
LAP TIME
(FIRST 3)
#22 L. HAMILTON
1'26.714
#2 F. MASSA
1'35.748
#4 R. KUBICA
1'33.096
#22 L. HAMILTON
#22 L. HAMILTON
#3 N. HEIDFELD
1'25.187
#7 N. ROSBERG
#1 K. RAIKKONEN
#1 K. RAIKKONEN
#4 R. KUBICA
1'34.188
#23 H. KOVALAINEN
#2 F. MASSA
1'31.188
#1 K. RAIKKONEN
1'21.813
#2 F. MASSA
1'20.584
(1LAP)
(28.408)
19th (1'23.565)
9th (1'20.848)
11th
7th
(67.540)
(53.597)
17th (1'16.306)
18th (1'16.330)
R (16LAPS)
14th
(1LAP)
17th (1'21.631)
16th (1'21.512)
R (38LAPS)
3th
(82.273)
14th (1'15.701)
18th (1'16.246)
17th
R (50LAPS)
#1 K. RAIKKONEN
#2 F. MASSA
#22 L. HAMILTON
#2 F. MASSA
1'25.994
(1LAP)
6th
#1 K. RAIKKONEN
#4 R. KUBICA
#2 F. MASSA
1'27.617
#2 F. MASSA
#22 L. HAMILTON
#1 K. RAIKKONEN
15th (1'16.537)2
11th
#2 F. MASSA
#2 F. MASSA
1'15.787
#2 F. MASSA
1'15.110
#22 L. HAMILTON
#4 R. KUBICA
#2 F. MASSA
#22 L. HAMILTON
1'17.886
#1 K. RAIKKONEN
1'16.449
#4 R. KUBICA
#22 L. HAMILTON
#3 N. HEIDFELD
1'16.909
#9 D. COULTHARD
#2 F. MASSA
1'15.024
#2 F. MASSA
#1 K. RAIKKONEN
#11 J. TRULLI
#22 L. HAMILTON
#23 H. KOVALAINEN #22 L. HAMILTON
#3 N. HEIDFELD
1'21.049
1'19.537
#17 R. BARRICHELLO
#22 L. HAMILTON
1'15.666
#22 L. HAMILTON
#22 L. HAMILTON
#6 N. PIQUET
1'14.603
#2 F. MASSA
(1LAP)
12th (1'20.332)
18th (1'21.332)3
12th
16th
(1LAP)
(2LAPS)
16th (1'38.880)
19th (1'39.811)4
13th
16th
(1LAP)
(1LAP)
#4 R. KUBICA
17th (1'48.211)
16th (1'48.153)
#2 F. MASSA
15th
R (19LAPS)
(1LAP)
19th (1'37.006)
16th (1'36.510)
15th
17th
(73.370)
(1LAP)
12th (1'45.133)
18th (1'46.583)
9th
R (14LAPS)
(19.885)
18th (1'19.100)
17th (1'18.882)
14th
13th
#22 L. HAMILTON
1'20.899
#2 F. MASSA
1'19.068
#23 H. KOVALAINEN
#12 T. GLOCK
#1 K. RAIKKONEN
#2 F. MASSA
1'38.989
#22 L. HAMILTON
1'47.338
#15 S. VETTEL
1'37.555
#2 F. MASSA
1'44.801
#15 S. VETTEL
1'37.842
#2 F. MASSA
#22 L. HAMILTON
#23 H. KOVALAINEN
#3 N. HEIDFELD
1'46.037
#22 L. HAMILTON5
#15 S. VETTEL
#23 H. KOVALAINEN
#23 H. KOVALAINEN
1'35.214
#4 R. KUBICA
#2 F. MASSA
1'44.014
#5 F. ALONSO
#7 N. ROSBERG
#22 L. HAMILTON
#22 L. HAMILTON
1'18.404
#2 F. MASSA
1'17.287
#5 F. ALONSO
#4 R. KUBICA
(1LAP)
(1LAP)
#1 K. RAIKKONEN
18th (1'37.053)
14th (1'36.079)6
#22 L. HAMILTON
16th
11th
(1LAP)
(85.061)
17th (1'12.810)
15th (1'13.139)
13th
15th
(1LAP)
(1LAP)
#22 L. HAMILTON
1'36.303
#2 F. MASSA
1'12.368
#22 L. HAMILTON
#2 F. MASSA
1'34.947
#1 K. RAIKKONEN
#2 F. MASSA
#23 H. KOVALAINEN
#5 F. ALONSO
1'11.768
#1 K. RAIKKONEN
1 Eliminated from the race for ignoring a red signal at the pit lane exit.
2 D. Coulthard was demoted five grid positions as a result of a gearbox change. As a result, J. Button was raised from 12th to 11th position,
and R. Barrichello from 15th to 14th position on the grid.
3 S. Bourdais was demoted five grid positions for impeding another vehicle in the qualifying session. As a result, R. Barrichello was raised
from 18th to 17th position on the grid.
4 Because a change was made to the vehicle in parc ferme, R. Barrichello was required to make a pit start in the race.
5 25 seconds were added to L. Hamiltons racing time for cutting a chicane, demoting him from 1st to 3rd position.
6 M. Webber was demoted 10 grid positions for an engine change, raising R. Barrichello from 14th to 13th position on the grid.
346

TOTAL AUS MAL BRN
ESP
TUR MON CAN
P.
TEAM (CAR NO.)
172
10
18
18
16
Vodafone McLaren Mercedes
151
14
10
11
BMW Sauber F1 Team
135
11
11
ING Renault F1 Team
80
56
Scuderia Toro Rosso
39
Red Bull Racing
29
AT&T Williams
26
14
10
Force India F1 Team
11
Super Aguri F1 Team
GBR
BEL
ITA
18
10
10
14
14
14
13
10
18
11
10
10
15
2
9
FRA
GER HUN EUR
8
10
ESP
TUR MON CAN
SIN
JPN
8
10
CHN BRA
14
16
10

TOTAL AUS MAL BRN
P.
No.
DRIVER (NAT)
22
L. HAMILTON (GBR)
98
F. MASSA (BRA)
97
K. RAIKKONEN (FIN)
75
R. KUBICA (POL)
75
F. ALONSO (ESP)
61
N. HEIDFELD (GER)
60
23
H. KOVALAINEN (FIN)
53
15
S. VETTEL (GER)
35
11
J. TRULLI (ITA)
31
10
12
T. GLOCK (GER)
25
11
10
M. WEBBER (AUS)
21
12
N. PIQUET (BRA)
19
13
N. ROSBERG (GER)
17
14
17
11
15
K. NAKAJIMA (JPN)
16
D. COULTHARD (GBR)
17
14
S. BOURDAIS (FRA)
18
16
J. BUTTON (GBR)
19
21
G. FISICHELLA (ITA)
20
20
A. SUTIL (GER)
21
18
T. SATO (JPN)
22
19
A. DAVIDSON (GBR)
10
1
10
10
10
10
10
10
10
8
1
1
3
10
1
2
10
10
10
10
10
5
3
347
10
10
5
1
2
1
5
8
6
1
4
2
1
2
2
10
6
2
CHN BRA
1
2
JPN
1
8
1
3
SIN
3
2
ITA
1
6
BEL
5
2
3
8
GER HUN EUR

6
4
1
GBR
10
4
5
FRA
List of Companies Participating in Hondas Third-Era


Honda commenced its third era Formula One activities in 2000, competing in Formula One Grand Prix in
collaboration with BAR. From 2006, we fielded our own team, and we were able to taste our long-anticipated first
victory. We owe our ability to continue racing for a nine-year period during our third Formula One era to the generous
assistance provided by the companies listed here, for which we offer them our heartfelt thanks.
We wish also to express our gratitude to all the staff members of these companies for their many tireless efforts
on our behalf, to everyone at Honda and other companies who provided advice and assistance, and, finally, to everyone
who offered us their warm support. We thank you.
A.L.M.T. Corp.
Daido Castings Co.,Ltd.
Hewlett-Packard Japan, Ltd.
AEROSPACE METAL
COMPOSITES Ltd
DAIDO METAL CO.,LTD
HIDAKA SEIKI CO.,LTD.
Daido Steel Co., Ltd.
HINODE-SEISAKUSHO CO,LTD
Daiei Barrel Co.,Ltd.
HIRAI CO.,LTD.
DALLARA AUTOMOBILI
Hitachi Advanced Digital, Inc
DASSAULT SYSTEMES
Hitachi, Ltd.
DEL WEST EUROPE S.A.
Hitachi Metals,Ltd.
DOME CO.,LTD.
Hitachi Zosen Corporation
DOWA HOLDINGS CO., LTD.
Hokuriku Light Metals Industry
dSPACE Japan K.K.
HONDA ENGINEERING CO.,LTD
Ebara Corporation
Honda Foundry Co.,Ltd.
EMZLAB Inc.
Honda Kaihatsu Co.,Ltd.
ENAX,INC.
HONDA PERFORMANCE
DEVELOPMENT INC
AJACOM CO.,LTD.
Akashi Communication CO.,LTD
ALCOA
ALTIMA Corp.
AN motorsport LTD
Arai Seisakusho Co.,Ltd.
Arisawa Manufacturing Co., Ltd.
ARTFAST
ASANO Co.,Ltd.
ASIA NETSUSYORI (in Japanese)
ASPECT Inc.
ATIA, Inc.
ENKEI CORPORATION
Honda Trading Corporation
automax co.,ltd
ERAMET INTERNATIONAL
TOKYO BRANCH
AutoTechnicJapan co.,LTD
FLOW SOLUTIONS LTD.
IBM Japan, Ltd.
AVL Japan K.K.
Ichikawa.co.,ltd.
BLANC AERO Technologies
FUJI FURUKAWA
ENGINEERING &
CONSTRUCTION CO.,Ltd
Bodycote
FUJI KEISOKU SYSTEM Co. Ltd.
IHI Corporation
Bridgestone Corporation
Fuji Manufacturing Co.,Ltd.
Illinois Tool Works Inc.
Busak+Shamban K.K.
FujiStaff, Inc.
IMV CORPORATION
Carpenter Specialty Alloys
FUJITSU LIMITED
INCS INC.
CEROBEAR GmbH
FUJIWARA CO.,LTD.
INOMATA KOGYO (in Japanese)
CHIRIKA.Co.,ltd.
FUNAMIGUMI (in Japanese)
IPEC co.ltd
Chuo Spring Co.,Ltd.
FURUNO ELECTRIC CO., LTD
Ishihara Precision,Inc.
CITIZEN MICRO CO.,LTD.
GEO TECHNOLOGY S.A
CMK CORPORATION
ISHIKAWA DIE & MOLD

CO.,LTD
GIKENSEIKI CO.,LTD
College Master Hands Inc.
Global Active Technology Co.,Ltd.
COMATSU
GS Yuasa Corporation
CorporationYamazaki Setsubi.
Cosworth Electronics Ltd.
HAGA KIGATA KANAGATA

SEISAKUSHO CO.,LTD.
Cranfield University
HEIWA SANGYO CO.,LTD.
IWAHARA Sangyou Co.Ltd.
CYAN Co.,Ltd.
HEPHAIST SEIKO CO.,LTD.
JEFTEC CO.,LTD.
350
Hoshino Corporation
ICS Corporation
IHI INSPECTION &

INSTRUMENTATION CO.,LTD
Ishimoto Architectural &
Engineering Firm, Inc.
ISHIZU EG co.,ltd
JFE Steel Corporation

JTEKT Corporation
Kakamigahara Aero Equipment
Co.,Ltd
MITSUBISHI RESEARCH
INSTITUTE, INC.
MITSUBISHI STEEL MFG.
Co.,Ltd.
OSHIDA SEIKI CO.,LTD.

Pankl Racing Systems AG
PARKER NETSUSHORI KOGYO
CO.,LTD
Kandenko Company, Limited
MITSUMORI SEISAKUSYO (in

Japanese)
PEARL GIKEN CO.,LTD.
KAWAGUCHI YUKI Co.,Ltd
MIYAMA CORPORATION
PHIARO CORPORATION
KDDI CORPORATION
Miyamoto Industry Co.,Ltd.
PLANSEE Japan Ltd.
Keihin Corporation
MIYASAKA RUBBER.,CO.LTD.
PLANSEE SE
Kenko Factory CO.,LTD
PSG
Kinzoku Giken Co,.Ltd.
MORIKAWA INDUSTRIES
CORPORATION
Kistler Japan Co., Ltd.
Morimura Bros.,Inc.
RDS CO.,LTD.
Kitahara Shoji Co.Ltd
MORIYA CORPORATION.
RESS CONSULTING GMBH
Kobe Material Testing Laboratory

Co.,Ltd.
M-TEC Co.Ltd
Retrac Composite Ltd
MTS Systems Corporation.
ROHM CO., LTD.
KOBE STEEL,. LTD
MW RACING GMBH
ROKI Co., Ltd.
Kobelco Research Institute, Inc.
NAGANO KEIKI CO.LTD.
Saint-Gobain K.K.
KOGANEI SEIKI CO.,LTD.
Nagase chemtex corporation
SAITAMA SHATAI CO.,LTD.
KOUKEN CO.,LTD
NDK,Incorporated
Kowa ss CO.,LTD.
NEC TOHOKU.LTD.
SAKAMOTO KYORITSU SEIKI

Co.,Ltd
ktel.co
NGK SPARK PLUG CO.LTD.
KURISAKI GEAR MFG.CO.,LTD NHK SPRING CO.,LTD.

KUWANA.,co.ltd.
Nichiei Co.Ltd.
KYB Corporation
NIHON DENKEI CO.,LTD.
KYOWA ELECTRONIC
INSTRUMENTS CO., LTD.
nippon ITF inc.
KYOWA INDUSTRIAL CO.,LTD.

LMS INTERNATIONAL N.V.
M.T.S
MagCanica,Inc.
Magnesium Elektron Laboratory
MAGNETI MARELLI S.P.A.
MAHLE (Vandervell)
QUALICA Inc.
SANKYO DENNKI SYOKAI (in

Japanese)
Sankyou International Corp
Sansho Shoji Co., Ltd
SATO RASHI CO.,INC
Nihon Mekki Industry Co., Ltd.
SEKI DENNSETSU KOGYO (in

Japanese)
Nihon Unisys,Ltd.
SEKIDAI KOGYO CO.,LTD
Nippon Oil Corporation
SHIBAURA ELECTRONICS
CO.,LTD.
NIPPON SEIKI CO.,LTD

NISHIKAWA KEISOKU .,CO.LTD.
SHINKO ENGINEERING CO.,

LTD.
NISSIN KOGYO Co.,Ltd.
Shin-Etsu Chemical Co.,Ltd
NISSIN MANUFACTURING
CO.,LTD.
SHINRYO CORPORATION
MAHLE ENGINE SYSTEMS UK Nissin Travel Service Co., Ltd.

LTD.
NITTAN VALVE CO.,LTD.
Mrkisches Werk Japan K.K.
NITTO DENKO CORPORATION
MARUBENI INFORMATION
Nittobo Acoustic Engineering Co.,
SYSTEMS CO., LTD.
Ltd.
MARUI-KEIKI CO.,LTD.
NOK CORPORATION
MEIDENSHA CORPORATION
NSK Ltd.
MEIRA Corporation
NTN CORPORATION
METALORE, INC.
OHMIYANISSIN CO.,LTD
MICHELIN
OHNISHI NETSUGAKU Co.,Ltd.
MITEC CO.,LTD.
ONO SOKKI CO.,LTD.
MITSUBISHI PLASTICS,INC
351
Shinto Industrial Co.,Ltd.

Shizuki Electric Co.,Inc.
Shonan Design Co., Ltd.
SHOWA CORPORATION
Showa Denko K.K.
Showa Electric Laborator Co.,Ltd.
Showa Seisakusyo Co.,Ltd.
SINFONIA TECHNOLOGY CO.,
LTD.
SINTOKOGIO,LTD. SINTO
BLASTEC CO.
Sony Manufacturing Systems
Corporation
SOURIAU Japan K.K.

SOYO GOMU Co.,Ltd.
Tokyo Institute of Technology

Graduate School of Science and
Engineering
STANLEY ELECTRIC CO.,LTD.
Tokyo Metal-Pack Co.,Ltd.
Sugihara Software & Electron

Industry
Tokyo R&D Co.,LTD.
SUM Electro Mechanics Co.,LTD.

Sumico Lubricant Co.,Ltd.
Sumitomo 3M Limited
Tokyo Sokki Kenkyujo Co.,Ltd.

TOKYO.FERROMITE.INDUSTRY.CO.LTD
TONETS CORPORATION
Sumitomo Corporation
TOSEI ELECTROBEAM
COMPANY LIMITED
Sumitomo Electric Industries, Ltd
TOSHIBA CORPORATION
SUMITOMO LIGHT METAL

INDUSTRIES, LTD
TOYAMA PREFECTURAL
UNIVERSITY Department of
Intelligent Systems Design
Engineering
Sumitomo Metal Technology Inc.

Sumitomo Wiring Systems, Ltd.
Super Resin, Inc.
SURFCOAT CO.,LTD
Suzuki Kinzoku Kougyou Co.,Ltd.
SUZUKI PRECION Co,Ltd
TOYO Corporation
TOYO DENSO CO.,LTD.
TOYO iTEC Co., Ltd.
TOYO KOKU DENSHI Co.,Ltd.
T.RAD Co.,Ltd.
Trelleborg Sealing Solutions Japan

K.K.
Taikisha Ltd.
Tsukasa Sokken Co., Ltd.
TAISEI CORPORATION
UBE Scientific Analysis .Inc
TAKASAGO LTD.
UCHIDA Co.,LTD
Takasago Thermal Engineering Co.,

Ltd.
UCHINO SEISAKUSHO Co.,LTD.
TAKATA CORPORATION
TAMAGAWA SEIKI CO.,LTD.
TANAKA SEIMITSU KOGYO
Co.,Ltd
TDK Corporation
Techno-Core Corporation
TECHNOPRO Engineering, Inc.
TEIKOKU PISTON RING Co.,Ltd
Tigers Polymer Corporation
TIMET
TNO-Automotive Japan K.K.
TOAMEC Inc.
TODA CORPORATION
Tohoku Steel Co., Ltd.
TOKAI SEIKI CO,Ltd.
TOKIWA CO.,LTD.
UESAWA WORKS,INC.
Vector & Scalar Products Limited
(VSPL)
Wave Front Co.,Ltd
Westmoreland Mechanical Testing
& Research, Ltd
Watanabe Architects & Engineers
co., Ltd
YACHIYO-SEIKI Ltd.
YAGISHITA GIKEN CO.,LTD.
Yamada Manufacturing Co., Ltd.
Yamaguchi University Graduate
School of Science and Engineering
Yamaguchi University Faculty of
Agriculture
YAMASAN CO., LTD
YAMASHITA RUBBER CO.,LTD.
YAMATO GOKIN
TOKOROZAWA ALLOY
FOUNDRY CO.,LTD.
Yokohama National University

Graduate School of Engineering/
Faculty of Engineering
TOKYO GASKET CO.,LTD.
Yuei Co., Ltd.
352
Comments from Project Members of Hondas Third-Era

Comments from Project Members of Honda's Third-Era Formula One Activities
I entered the company (MSD at the time) in 99 and

could witness almost all of Honda Formula One Third
era activities. It is good experience and memory for me
that I participated in challenging duties within the limited
time and shared emotions with many people in the ten
years.
Daisuke Aoki
Test in Barcelona from November 17-19, 2008.
For me, it was the first running test as a data
engineer.
It was also the last running test for the Honda
Formula One Third era activities.
To be honest, I wanted to continue participating in
the work since Ive become long-awaited data engineer.
Kenji Akimoto
Ive been taking charge of general management of
technologies as ALPL of KERS. I feel proud that we
could complete a competitive unit from the first year of
the regulation owing to strenuous efforts of all project
members and local staffs. I really enjoyed the job. Id
like to express my deep gratitude to the management
members that awarded me such a chance, project
members who pushed themselves to the max to achieve
target performance and schedule and comrades who
supported us in life in the UK and Germany. Thank you
very much.
Hirofumi Atarashi
HRD Machining Div. members are local staffs and
have been conducting training on site until they become
able to perform the same level machining as HGT. HRD
is now able to manufacture original manufacturing and
I realize that we have contributed to making HRD
independent. Associates on the spot/actual parts are the
starting point to win sympathy of people who are
different in cultural background and the way of thinking.
I intend to make use of the experience in teaching junior
fellows.
Toshikazu Arai
The Honda Formula One Third era activities
developed me since when I entered the company. I met
excellent people and gained wonderful experience in an
environment which seemed to have embodied the
challenging spirit. I dont think I was happy; I couldnt
be happier. Let me say thank you and looking forward
to come back some day to the stage of Formula One.
Yoshimi Ishii
My dream is to win the championship with an F1
racing car I constructed. I entered the company in 2008
and was assigned to the MS in October of the year but,
immediately after that, Honda decided to withdraw from
Formula One. I felt frustrated losing the opportunity to
directly participate in the Honda Formula One Third era
activities. Id like to realize my dream in the fourth era
activities never forgetting the vexation.
Yasuharu Ichikohara
I participated in the development of F1 engine valve

drive system materials as well as DLC and Ti materials
for three years and the state-of-the-art technologies
including the surface form control technology. I cannot
forget the emotion even now I experienced at the victory
of Button in Hungary GP in 06. I believe that Honda
will come back to the stage of Formula One in the near
future.
Naoaki Ito
I entered the company in April 2008 and assigned to
the MS in October to participate in the development of
F1 engine but, to my regret, Honda decided to withdraw
from Formula One just when I was highly motivated
becoming an F1 engineer which is my dream.
Nevertheless, Im proud of the experience of
participating in Honda Formula One Third era activities
if only for a short period of time.
Yuki Ito
I devoted myself to development making positive
progress every day. I cannot forget, however, the days I
passed in vexation unable to achieve results. I had to
make efforts more and yet more. I believe that Honda
will come back to the stage of Formula One mainly
supported by junior staffs if we make use of the
frustration as leverage for the next development.
Nobuyuki Imai
Our objective is simple and clear: to win the race.
In retrospect, I found myself always pursuing higher
power. We held conviction to carry through all by our
own unaided efforts and the spirit of engine developer
never to give up whether beaten or have broken up.
However, the Honda Formula One Third era race
events were not as simple as CART races.
Id like to spring back stronger from the vexation of
finishing without driving to victory.
Yoshiyuki Ugomori
It was the moment that my dream from the days of
my earliest recollection was realized when I became a
member of the Honda Formula One Third era activities.
Feeding on the experience and the memory, I intend to
vigorously push forward to realize a new dream. Lastly,
Id like to say a word to me in the past, Yes, you did
it! One of your BIG DREAM from early COME TRUE
!!!
Tsubasa Uchida
The experience of the challenging work in the
development of F1 racing car I desired to participate in
from the beginning when I entered Honda has become
a precious knowledge of my life. I thank everyone who
supported me in the chance and would like to engage
in everything with the spirit of Never give up! also
in future.
Mamoru Uraki
354
Dream in my boyhood days; ambitions; engine

designer; chance; joy; BAR HONDA; energetic work;
expectation; excitement; eagerness; accumulation;
hardships; ALL HONDA; race track experience; world;
culture; victory; delight; setback; challenging; struggle;
chagrin; growth; bro; thanks for all.
Hajime Endo
I participated in Honda Formula One Third era
project for about six years in the area of aerodynamics
development. I acquired precious experiences beyond
count and it was the greatest moment in my life when
Honda won a victory with FOX EAR of my personal
contrivance installed on the racing car.
Tsuyoshi Okubo
I participated in the development of F1 engine for
eight years from when I entered the company. I
remember well that I found it very tough physically at
the beginning unable to catch up with the pace of work.
The experience in Formula One is very precious having
not only built up the base of work for me but also taught
me importance of collaborating with comrades.
Yasuhiro Okura
We introduced high-strength material Mg casting/
thin-wall Al vacuum casting and all-out high-quality
casting utilizing X-ray CT scanning into actual race cars
for ENG weight-saving. Id like to continue to
vigorously practice those matters I experienced in the
high-technology manufacturing in the future work.
Yuuji Ooshima
I committed for three years from 2000 to the job of
supporting in those F1 staffs that are sent overseas on
business.
I was given a job in the race building in the middle
of the years. I remember that I became glad and sad by
turns in front of TV feeling Formula One closer to me
and realizing that Im a member of the team. Though I
only had a little might, Im proud that we fought together
and felt sympathy.
Ryosuke Otsuka
With a simple objective to win the race, I could
participate in the development of the unit giving priority
to the performance.
As the result, we could get through to application of
the heat spreader having excellent heat conductivity to
racing cars.
Masatoshi Okumura
About five years and three months from 2003 when
I entered the company and assignment until withdrawal
from Formula One have passed in just about not time
like a dream. I met and cooperated with various people
through development, Test, Race events acquiring a lot
of knowledge, techniques and experiences. Id like to
thank fans who gave me a lot of deep emotion.
Toru Ozawa
We achieved zero trouble in all race events

overcoming difficulties in introducing Hondas first
original GBX into actual racing by organizational
strength and technological competency and speed. We
are proud of pursuit of high-precision finishing
technology which is underlying the seamless shift and
the luster of light-weight final gear. Id like to leverage
in the next job the sense of fulfillment/accomplishment
I got in Formula One activities.
Shin Odagiri
Not to let defective products we manufactured
outflow, that was the duty of us participating in F1
activities and was worth doing. Not to mention the
processing precision, we were to solve problems one by
one with respect to the surface scratches and dents to
achieve high precision assurance. While it was difficult
to achieve both quality and schedule, I intend to make
use of the experience for supplying manufactured
products in the future.
Tokio Kasai
I participated in the development of electrical
components from 2001 to 2008.
I believe that the experience of being engaged upon
the work toward clear and inexcusable objective to
complete the development and let the racing car get in
the grid by the first racing event in March next year hold
clues also for developing mass-produced cars at present.
Toshikazu Karube
While it was a work of the greatest difficulty to
complete development within such a short period of time
as a year and a half, we succeeded to put the worlds
first HEV F1 racing car on the racing track as the result
of all persons involved in KERS including team
members and HRF1 engineers who became singleminded in the efforts. Id like to express my respect in
particular to Mr. Dunkan Elliott of HRF1 and Miss Ohno
who took charge of translation into English. Thank you
very much everyone.
Masato Kita
I acquired precious experiences actually feeling the
severity of racing and meeting excellent technicians and
persons. Im determined to catch up with competitors
and overtake them in ten years dreaming a day when
Honda makes a strong impression on people all over the
world and win the championship with overwhelming
technology.
Hiroshi Kimura
While it was a short period of time, the experience
of learning many things among capable technicians is my
wealth. In addition, Im proud of having participated in
F1 race events with excellent comrades.
Tatsuya Kyomitsu
355
Personally, I regret that the part I was developing

would be introduced into F1 racing cars if Hondas
withdrawal was decided a year later. Including this, I feel
afresh that it was a racing.
That is, race against time, fight for performance and
competition against ourselves.
Im very much grateful that I could have precious
experience on the stage which was my childhood
ambition.
Yoeru Kono
I participated in F1 activities for about 10 years in
the G/Box and ENG areas. I could have a precious
experience to directly feel Honda DNA through my skin.
Id like to express my gratitude to comrades with whom
I worked together and business partners who supported
us despite the peculiar field of racing. I have my heart
on becoming No. 1 in the Formula One Fourth era
activities.
Daiki Konagaya
I watched F1 race event on TV in 2000 for the first
time to love and yarn for Formula One and entered
HONDA. The ten years of Honda Formula One Third
era activities is precious for me encouraging me with
stuff of dreams and is the period of time when my dream
to be engaged in Formula One was realized. But I have
another dream yet. That is to become a Honda Formula
One Fourth era activities engineer!
Megumi Kobayashi
In the several years during which I worked near the
F1 activities staffs, I didnt like Formula One but
actually felt the biggest pleasure in working to make
arrangements for those staffs who are striving for the
objective in F1 with the utmost energy. I was much
encouraged by them.
Thank you very much.
Yuichi Koyanagi
I have been engaged in the development of fuel
injection system for about three years from 2005 when
I entered the company. Though it is regrettable that
Honda made the decision to withdraw from Formula One
when we were working toward the 09 season, I
remember well that I was deeply moved when we won
a precious victory in Hungary.
Yohei Sakai
While it was such a short period time as two months
from when I was assigned to the MS Materials, I gained
very dense experience. What I felt in then is that it is
important to interact with the staffs of other areas not
shutting ourselves up in the work area we are
commissioned. Im determined to apply the experience
in the future work.
Katsutoshi Sakurai
The sudden withdrawal from Formula One was really

shocking for me.
I applied for the recruitment of Formula One activity
staffs and transferred from Sayama Plant in April 2006
to participate in the parts management job for about three
years.
I learned many things while making many fluffs
through F1 activities. Thank you for all your trouble. The
experience of participating in the F1 activities, very
pinnacle of the motor sport, is lifelong treasure of me.
Satoshi Sasaki
While it was only a year and three months from when
I was transferred to come to the MS as a KERS staff
until Hondas withdrawal from Formula One, I learned
many things from bench test to circuit running test was
being supported by many people. I intend to devote
myself so that I can make use of the experience for the
next work.
Kinji Sato
Weve been pursuing wear resistant material for and
technology to reduce diameter of the valve guide for the
valve drive system and introduced them in actual race
events assessing productivity of them aiming at
reliability assurance of ultra-high speed ENG. I intend
to make use of the approach I acquired through F1
activities in various fields even in difficult manufacturing
processes. Id like to express my deep gratitude for that
I could participate in Honda Formula One Third era
activities.
Masakatsu Satoh
I wanted a little more time and its regrettable that
we couldnt achieve results. I intend to devote myself
to everyday operations taking motive power from with
the regret for the fourth era activities. While it was such
a short period of time as a year and a half, Im proud
of having participated in the F1 activities desiring to by
myself and Id like to express my gratitude for the
chance given to me.
Makoto Sawada
We hit a losing streak and withdrew from F1 in the
end losing an opportunity for recovering lost ground. We
were disgraced which will last me a lifetime. Im now
on the stage of development for mass-production. We are
not allowed to play second fiddle in the technical field.
I intend to see if there is any way in which I can be of
help for becoming peerless.
Yosuke Sawada
356
I watched Formula One for the first time when I was

the first grade of an elementary school when Maclaren
Honda with Senna was flowering. I was a high school
student when Honda made a comeback to Formula One
in 2000. And I was a college student when Takuma Sato
was taking an active part and Honda won a victory for
the first time in the Honda Formula One Third era
activities. I entered Honda in 2008. It was two months
after I was assigned to the F1 engine design team of the
Honda R & D that Honda decided to withdraw from
Formula One. I feel regret but Im looking forward to
the fourth era activities.
Yuki Shigeta
I was transferred to the MS in January 2008 and I
only worked just about a year for the F1 activities. I feel
I did nothing but to learn the work in the F1 activities.
Nevertheless, I expect that the experience in promoting
development giving priority to speed and anticipating the
future and how to have selections at an early stage will
become my wealth for the future.
Kentaro Shimada
I enjoyed very much participating in the F1 activities
which is my childhood dream. While it is regrettable to
withdraw from Formula One without winning the
championship, the experience of the vertiginous busyness
in tests and the feeling of tension on the site of racing
event and in particular that I was on the spot when
Honda won a victory in Hungary are unforgettable
memories.
Takashi Shimada
It was on August 30, 2007 that I was entrusted with
MS_BL which may best be described by the expression
thunderbolt from a clear sky. I received a telephone
call communicating me a business order that I should be
Operating Officer from September 1. That was from the
day after the next day. A year and a half since then had
elapsed before I knew it. In Barcelona where we got
points for the first time, I felt so excited as if my heart
was leaping into a mouth. And that was only for the
sixth place. I wondered how I should be if we were
competing for victory. While we couldnt achieve good
results the year, our competitive power was surely
improved inviting Ross Brawn. Its very regrettable that
we no longer have a chance to deliver the product next
year. Id like to finish taking operational responsibility
in the MS expecting everybody to take an active part to
get your own way that much, I believe you will
naturally.
Katsuhiko Suzuki
I applied for the project recruiting staffs who wish

to participate in the F1 activities and was assigned as a
member of the Racing Development Div. in October
2001. From then, I participated in the LIC Racing Car
Components Sec. for seven years. While the section was
taking charge of works doubly behind the scenes, we
insatiably pursued how to respond to real-time to
information and materials flow that changes every hour.
Regrettably, Honda decided to withdraw from the F1
activities before startup of the fastest physical
distribution system, I learned through the system
development and physical distribution promotion works
that victory can only be achieved on confidential
relationship and collaboration between people. Im
looking forward to the fourth era activities.
Ken Takahashi
The suddenly started and suddenly ended ten years
have elapsed in just about no time.
I was made re-realize the importance of improving
individual technologies at the same time the profundity
of driving a car at a high speed and racing where people
compete through the experiences that we couldnt win
only with individual good technologies.
It is regrettable that I have to leave F1 activities
without knowing everything of difficulties and
fascinating aspects of Formula One.
Hiromasa Tanaka
Participating as an engine designer, I leaned the
difficulty of making design being conscious of the
Number one in the world. I also had very precious
experience of coming in touch with new attitude of the
different idea/viewpoint unique to the EU.
Katsuyoshi Chubachi
I competed for F1 World Championship in one place
after another as an engine technology staff of the Super
Aguri F1 Team. Im proud of getting a place by an all
Japan team (team with nationality of Japan, Japanese
team representative, Japanese driver, Japan-made tires
and Japan-made engine). Thank you very much for your
cheering.
Yoshito Tsukamoto
Id like to thank our business partners who supported
our development activities and staffs of other divisions
and sections who always willingly aided us in
consultation.
We could commit to development to the full thanks
to the intensively allocated resources. We couldnt be
happier as engineers.
Masayuki Tsuchiya
I got stimulated competing on the world stage. I

understood well at which point my ability is. I still have
a lot of things I want to do though I have to leave
without fulfilling my life ambition. I feel contented that
I could work in such an environment that makes me have
such feeling. Id like to gain more victories next time.
Shiroh Takasaki
357
I was very happy that I could work on the childhood

dream stage. While the withdrawal of this time is
regrettable, Im proud of the experience of participating
in Formula One and Im determined to do what I can
do to the full extent of my power. Id like to thank those
fans who have supported us until today. See you in the
fourth era activities.
Hiroki Tsuboi
I participated in the development of KERS. While it
was such a short period of time as no more than a year
and a half, Id like to express my gratitude to all those
who supported me for enabling me to have very dense
experience. Id like to make the experience I gained
more valuable making best use of it for activities in the
future. Thank you very much.
Masaya Deguchi
I played a part to the best of my ability for F1
technologies as a race car electrical component engineer.
I came to participate in supporting SAF1 in 2006 to
witness the F1 technologies and felt much of the Three
Reality Principle.
I intend to make efforts to reflect the technologies
and the spirit I cultivated in the F1 activities in everyday
life in the near future.
Akiko Tokimoto
I took charge of the ENG casting dies and could
make significant contribution to improvement in
efficiency/improvement in initial quality of casting by
unification of design/manufacturing 3D data by
CATIAV5 and complete application of casting
coagulation CAE. Im proud of this and Id like to make
its use for the job in the future.
Yousuke Tokiwa
Taking charge of casting core for F1 engine, I could
introduce new core technologies such as the closed block
between narrow shafts introduced in EUR GP in 2002
and the complex core cylinder head introduced later. In
the Honda Formula One Third era activities, I could
experience challenge spirit of never give-up and ideas
becoming reality/actual product.
Masanori Totsuka
While I only participated in KERS development for
17 months, it was a happy experience to have
participated in the development until it is almost
completed for introduction into race event in 09
performing development with unprecedented density
feeling passions of members who wish to avoid being
defeated in the hybrid technology aiming at number one
in the world.
Yasuhiro Toyoda
I could participate in the aerodynamics development

for Formula One I desired to from when I entered the
company and finally realized to reside in HRF. It was
very precious time in my life.
If, some day, Honda challenges Formula One again,
I want to participate in it again.
Yusuke Nakai
I was given a chance to take charge of engine design
and gear box design on the supreme stage of Formula
One to compete the number one in the world which is
my childhood dream. Id like to thank those who
supported me for letting me have the very precious
experience. Thank you very much.
Yasuo Nakagawa
I participated in the development of F1 car gear box
for about seven years from 2002.
To say nothing of durability and preciseness of F1
car components and high-level control program, I learned
various matters including development speed and
processes until introduction to race events.
Ill never forget the excitement we got when we
overcame difficulties.
Hiroshige Nagata
My name is Teruoki Naganuma. I took charge of the
development of gears. While we couldnt achieve results
in the Honda Formula One Third era activities, Im glad
that the results we achieved are shared by way of this
paper. Personally, I was very happy that I could
participate in Formula One which had been my adoration
since my childhood. Id like to express my heartfelt
gratitude to the comrades of the F1 team.
Teruoki Naganuma
In 2002, when I was a J.V Race engineer, Jacques
Joseph Charles Villeneuve mocked that HONDA ENG
is not a horse but a donkey in his comment on the ENG
at that time after running as I remember.
While we tried new items such as Rotary Throttle at
the time to introduce to the race car, we couldnt achieve
good results and we of the Honda R & D and actual
running members had very tough time.
Spring back from the vexation, we succeeded to
introduce a new ENG in the middle of the season which
made Jacques not use the word Donkey, and the ENG
was developed into one without problem both in the
Power and DR. While we couldnt achieve results in race
events, we of the ENG producers take pride in that even
drivers recognized the result of taking on challenges.
Satoshi Nakamura
358
I participated in manufacturing the exhaust pipe

referred to as a craftwork in the Honda Formula One
Third era activities. We felt pleasure and accomplishment
supporting the development by solving number of
challenges and realizing technologies including creation
of alternative plan, joggling/bulkheading/down-sized
construction by might and main to improve ENG
performance aiming at pulsation effect. I intend to make
use of the developed EXPI and cultivated K/H in the
future.
Takumi Nakayama
In the F1 activities, I experienced on site various
works which were very beneficiary to me. While I
experienced both hard times and enjoyments, those
works proved worthwhile and enabled me to build selfconfidence. I believe that the experience on site will be
of use for work in the future.
Katsura Nariai
I could take on various challenges participating in F1
activities for five years from September 03.
In the night in Suzuka in 04 after calm down from
the excitement of the race event, there had happened a
sudden HONDA call from the main stand. We got
stimulation/support from passion/expectations of fans we
of all members directly felt going out to the race track.
Yuji Niihara
Finishing the development of CART ENG in 2002,
I joined together with the F1 development team. I
committed to the development of cladding type valve
seat and completed the work supported by the team, the
Prototype Sec., EG and subcontractors. I felt special
accomplishment when the part was put into practical use
taking pains.
Hiroaki Nishida
Honestly speaking, visionary prototype V8 engine
stands out in my memory. The engine, put on 04 chassis
by force ahead of regulation modification in 06, was the
first compliant ENG in the world as with Hondas
KERS. While Im going to be transferred to the GT Div.,
I intend to continue to pursue becoming the First to
arrive.
Ken Nishimori
I pride myself on participating in the F1 activities,
if indirectly, as a facility infrastructure exploratory
member from the inauguration of the HRD in 99 until
the HRD & F integration plan in 09 and it gave me a
precious experience. In addition, I feel Im lucky as a
facility management personnel as the facility I introduced
is producing results still now. I intend to respond to this
by handing down the precious experience to the posterity
when Honda makes comeback to the stage of Formula
One.
Masamichi Hagiwara
Participating in the F1 racing car development job

for five years from the time I entered the company, Im
grateful to the Honda Formula One Third era activities
that gave me various precious experiences which cannot
be obtained other than through racing including the
unique feeling of tension, importance to carry through
to the end there and pleasure of watching the race car
running with the part I developed installed.
Junichi Hashimoto
I participated in the Prototype Group from 03. Im
proud of being able to do parts delivery with confidence
assuring the quality. I could hang in there with utmost
effort despite the tight schedule owing to the comrades
who were doing their best on the circuit and many
people who supported us. Id like to extend my thanks
to Jenson Button who won a victory at which I became
teary-eyed as well as to the Honda Formula One Third
era activities.
Yoshihisa Hayakawa
Im grateful to Honda Formula One racing activities
which subjected me to various and precious experiences.
Takuji Hiroma
Im grateful to a month I passed in the MS meeting
capable associates whom I should have in my sight and
precious experiences I had in the course of my work. Im
looking forward to a day when HONDA makes
comeback to Formula One. I am going to make every
single effort to better myself everyday so that I will be
able to run about with an MSD seal affixed to my PC.
Gakuyo Fujimoto
There are things which can only be understood by
being on the spot and cannot be understood only by
watching from outside. Whether it is a good thing or bad
thing, everything will enrich your body and soul in your
later life. Id like to make full use of what I gained in
the future life. HONDA RACING SPIRITS passes for
at any time and place. I wish good luck to my comrades
and I will do my best myself.
Ryuichi Furukawa
While we couldnt achieve results worth the name of
Honda, I was happy being able to work as an engineer
in Formula One, my personal dream. Id like to express
my heartfelt thanks to all in and out of the company who
supported me. I want to realize subsequence of the
dream some day with new experiences I will have in the
future.
Yasutaka Masumitsu
359
I took charge of F1 engine valve drive system in

2003, 2005 and 2006 committing to racing engine since
when I entered the company. Particularly in 2005, Honda
introduced a new valve drive system for the first time
in the past 13 years to largely reduce weight of the ENG
HEAD as well as to enhance the power and the ENG
was appreciated as the mightiest engine in the world of
Formula One in the last year of V10 engines and in
which I take pride.
Hirotake Matsubara
Im very pleased to have participated in the Formula
One project for six years from autumn 2002 and have
passed the time with good comrades. The network I
constructed with many people of the indirect divisions,
Body Manufacturing Div., Materials Handling Div.,
Electrical Equipment Div. and Prototype Div. I started
from device development for sensing high-speed world
and I got various experiences including examining
buildings for the job. I was also moved by dedicated
efforts on the part of engineers working behind the scene
to feel the sense of pleasure and accomplishments. I
believe that if Honda makes come back to run at full
speed on F1 track, we can further make great strides
making use of the experiences gained in the Honda
Formula One Third era activities.
Keiji Miura
Hollow connecting rod which was developed as a
result that the words box shape is best suited for lightweight and high-strength inspired the manufacturer
spirit. We introduced the component in 03 season race
events after spending about two years. Then, we also
committed to hollow crank. We had almost completed
the development of the component when junction was
banned and it was abandoned before introduction to
actual race cars. The pick off by the regulations is the
proof that our technology was recognized by the world.
Ken Mizogawa
In the ten years when I participated in the Honda
Formula One Third era activities, I experienced many
joys and difficulties of taking on a challenge and have
greatly grown up as an engineer.
Id like to thank all people who supported Honda up
to now.
Powering dreams one lap at a time
Sadami Minato
In the busy days where racing events dont wait for
us, we passed hearing BGM of high-pitched sound from
the bench. When I got positively turbulent in the circuit
PIT, I got back to myself glued to Formula One in my
boyhood (The Power of Dreams). Im very proud of
having been able to participate in the F1 activities as an
employee of Honda. Id like to win a victory to make it
the wealth of my life.
Takahiro Minezaki
I was very sorry when I heard the news of

withdrawal from the F1 activities. I was so delighted
when Honda won the first victory in the Honda Formula
One Third era activities in the 13th race event Hungary
GP in 2006 that I shouted loud in front of TV of which
I have a distinct memory.
Hiroyuki Miyake
In the Honda Formula One Third era activities, we
supported enormous progress of ENG by element
development and manufacturing at fierce speed in
response to various changes to regulations. The
experience, among others, of having contributed to the
development of apparitional 09 NEW ENG with the
maximum speed and quality and passion is a very
precious experience which can be made use of for the
future.
Kazufusa Miyajima
Technology development in the F1 activities was
really a competition in a state-of-the-art area of the
world. I will feed back the precious technologies I
studied in the F1 hybrid material development to massproduced cars commencing with INSIGHT in the future.
Takehiro Miyoshi
It must be seldom that we become so vexed or happy
that we feel like crying. In the F1 activities that Honda
joined in competition advocating development of young
engineers, I could experience both. Im grateful to
Honda for giving me such a chance.
Shigeki Morie
I learned importance of personal initiative and
responsibility, ability to take action taking initiative and
giving output on a timely basis from senior associates
who have high motivation never saying No granted
that they may say Yes, but.
Since my objective and intention toward technical
development must be invariable, I intend to make use
of them for the future.
Akihiro Yanase
The Honda Formula One Third era activities were
actually global competition for material engineers. We
rivaled global competitors to keep on evolving
component specs making full use of advanced material
technologies. I learned how it is important to continue
to take on a challenge in race events.
Id like to repay the favor by handing down the
Honda DNA I have taken over from Mr. Matsubara and
senior associates to junior fellows.
Hiroshi Yamada
360
It was such a short period of time as a little less than

a year from January to December 2008 that I took charge
of making arrangements for overseas business trip for F1
activity staffs.
The period of a year when I felt Formula One
familiar to me though I was ignorant to Formula One
in the past was all fresh to me and I have become
interested in Formula One. Thank you very much.
Mari Yamada
I fully understood what work is, what a car is and
what racing is. What I learned here will flow in my
future life like vein. Hopefully, the water will blast out
again into the field of Formula One. Im grateful to the
precious days.
Ryozo Yamamoto
I applied for the project recruiting personnel who
wish to participate in the F1 activities and was assigned
to the Racing Assembly Div. transferred from the
Service Div. of the Head Office. I found there that the
racing ENG had remarkable quality and the race site
harshness was out of ordinary. Instead, the sense of
fulfillment and the contentment gained by carrying
through those difficulties have invaluable allurement.
The experience and the memory at the time that those
who have passion gained working hard together will
never fade away.
Makoto Yoshino
[Glossary]
ENG
GBX , G/BX
TH
HEAD
EXPI
K/H
J.V
MSD, MS, MS_BL
HRF
Facility control expert
LIC
EG
Engine
Gearbox
Throttle
Cylinder Head
Exhaust Pipe
Know-How
Jacques Villeneuve
Motor Sports Division
HONDA Racing F1 Team
home base
Facility Control Section expert
Parts Control Division
Honda Engineering Co., Ltd.
361
Honda R&D Co., Ltd.

Honda R&D Technical Paper Committee
Automobile R&D Center
Editorial Committee for F1 Special (The 3rd Era Activities)
Afterword
Almost a year has passed since Honda suddenly announced its withdrawal from Formula One, and we
commenced planning this special issue of the Technical Review. An older colleague once told me that being
guaranteed a position in the company and being able to participate in race development was great good
fortune for an engineer; for me, being involved in race development and being able to bring together the
outcomes of those projects as an editor has been an even more precious experience.
Today, the gulf between race technologies and mass-production technologies is pointed out to be great
in comparison to Hondas first and second Formula One eras. While their applications and effects may not
immediately be apparent, the technologies that we produce by reaching for the ultimate will be useful and
valuable in some way to the future, and I will be happy if their recording here in this form can show this
to be the case.
On the final note, I would like to take this space to remember our friends and colleagues who, to our
sadness, passed away during the time that they participated in Hondas Formula One project: Tomohiro
Kumagai, who was involved in the development of both hardware and software for vehicle bodies and gear
boxes; Tadasu Takahashi, who was involved in engine research based on track running and carried out
research on gear boxes; Kenji Nakashima, who was involved in engine research; and Masanori Wakasa,
who was involved in the development of control systems. This volume is dedicated to their memory.
(Hiromasa Tanaka)
Honda R&D Technical Review
2009 Honda Motor Co., Ltd. (All rights reserved)
Published by:
Honda Motor Co., Ltd.
2-1-1, Minami-Aoyama, Minato-ku,
Tokyo 107-8556, Japan
Edited by:
Honda R&D Co., Ltd.
Honda R&D Technical Paper Committee
Printed:
Dec. 1, 2009
(F1, FORMULA ONE, FORMULA 1, FIA FORMULA ONE WORLD CHAMPIONSHIP, GRAND PRIX and related logos are trademarks of
Formula One Licensing BV, a Formula One group company. All rights reserved.)

F1-SP2e All

Uploaded by

Copyright:

Available Formats

F1-SP2e All

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

F1-SP2e All

Uploaded by

Copyright:

Available Formats

Special Issue Hondas Third Formula One Era

Director and Advisor

Honda R&D Technical Review 2009

F1 Special (The Third Era Activities)

Descriptions of Engine Technologies

Descriptions of Gearbox Technologies

Development of Lightweight and Compact Differential for Formula One Car

Descriptions of Chassis Technologies

Descriptions of Vehicle Control Technologies

Descriptions of Electronic Equipment Technologies

Descriptions of Material Technologies

Honda R&D Technical Review 2009

F1 Special (The Third Era Activities)

Development of Metal Matrix Composite Piston

Descriptions of Race Management Technologies

List of Companies Participating in Hondas Third-Era Formula One Activities

Comments from Project Members of Hondas Third-Era Formula One Activities

< Prior Reports >

Looking back on Hondas Third-Era

In attempting to define the scope of the challenge that

moving ahead with the development of each element. A

* Automobile R&D Center

Honda R&D Technical Review 2009

activities, to also engage with this issue in the area of

F1 Special (The Third Era Activities)

EX camshafts geared at the front and rear ends, were

Looking back on Hondas Third-Era Formula One Activities

above, we developed individual element technologies

Transition of Regulation and Technology in

(1) First era (1964-1968): Raced under the Honda name as

Teams entered in the Formula One World

Transition of Regulation and Technology in Formula One

Formula One regulations, and will then go on to consider

a slight deviation from the narrow scope they define

2. Characteristics of Formula One

2.2. Example of a Conflicting Interpretation

Honda R&D Technical Review 2009

between the FIA and the teams regarding the level to

F1 Special (The Third Era Activities)

has no intention of arguing with the I.C.A. ruling.

3. Changes in Formula One Regulations

Transition of Regulation and Technology in Formula One

Type used by general public

505 kg (without driver)

3.75 bar maximum

605 kg for qualifying

600 kg (with driver)

605 kg (with driver)

800 mm above reference plane

19000 rpm 18000 rpm

1 engine for 1 engine for 1 engine for 2 race events 8 engines

Automatic gear changing

Refueling during race

1998 1999 2000 2001

Between 100 mm and 300 mm Between 150 mm and 350 mm and

Between 50 mm and 250 mm

40 GPa/(g/cm3) limit for all parts

Designated for engine parts

Honda R&D Technical Review 2009