A53-Knowledge tit-bits on Transformer

Electrical Steels-Part-1
 Published on November 30, 2021

P Ramachandran

Specialist, Power Transformers

66 articles

This is in continuation to the previous two articles on the subject.

https://www.linkedin.com/pulse/a51-electrical-steels-transformers-books-p-ramachandran

https://www.linkedin.com/pulse/52-transformer-electrical-steels-key-technical-iec-p-
ramachandran

1. Steel Grades

Transformer CRGO steel available today can be divided in to following types.

1) Based on thickness – 0.18 - 0.20- 0.23 - 0.27 - 0.30 - 0.35 mm

Thinner the lamination, less is the specific core loss due to the reduction in eddy loss
component. But lamination factor will be less, reducing the effective core area. With reduced
eddy loss, hysteresis loss component goes up due to misalignment of the secondary
recrystalised grain orientation relative to Goss orientation ie (100) [100]. Hence for a stacked
core it is doubtful to have grades with thickness less than 0.18 mm as it will be
counterproductive. Eddy loss is also reduced by the increased resistivity of steel through the
higher % of silicon content. But fabrication difficulties and coarse crystalline grains limit the
silicon content in CRGO to around 2.9 -3.2%.
Hysteresis loss component is reduced by better grain orientation. The third component of core
loss viz anomalous loss component from rotational flux at corner joint is reduced by refining
the domains of crystals formed on the laminations.

2) Based on the manufacturing process and specific loss (W/kg at 1.7 T) –

The following types of CRGO are available today based on the manufacturing process:

(a) Conventional grain oriented type

(b) High Permeability type (Hi-B)

Permeability is the ratio of magnetic flux density to magnetizing force. If it is high,

magnetizing the steel is easy ie for any induction level, magnetizing force will be lower ie
magnetizing current (also magnetizing kVA) will be lower.

(c) Domain refined high permeability type

Iron loss (eddy loss component) is reduced by providing linear regions of low permeability
on steel surface, almost right angles to the rolling direction. Domains are refined by the
demagnetizing effect of the magnetic poles generated. These paths are created by two ways-
one by mechanically scratching the surface of lamination or by irradiating the steel surface
with a narrowly focused, high temperature, plasma flame (laser) for a short period or other by
mechanically scratching the laminations. Advantage of this later method developed by
Nippon steel is that these steels can be annealed. Laser irradiated steels cannot be annealed
as their magnetic properties will be adversely affected by annealing.

Initially CRGO grade designation adopted by AISI (American Iron and Steel Institute) was
based on steel thickness.

M6 -0.35mm; M5-0.30 mm; M4 -0.27 mm M3-0.23 mm and M2-0.18 mm. Each thickness of
steel was specified with a certain specific core loss limit at 1.5 T for 50 & 60 Hz. When, I
joined industry 55 years back, the only steel grades available in India for transformers were
M6 & M5. In 1970 when high permeability steels came in to the market, the above grading
was not enough and for each thickness, grades such as M0H, M1H, M3, M4 were adopted
based on specific losses at 1.7 T. (H designated high permeability grade) Each mill started to
use their own designation, like H0, H1 (Armco) RGH grades (Kawasaki) and ZH -high
permeability; ZDKH - domain refined by laser irradiation; ZDMH- domain refined by
mechanical scratching (Nippon Steel)

Grade designation as per the latest IEC standard for CRGO is as below: Table 1& 2 of IEC
60404-8-7 “CRGO steel strip and sheets”

Example -M120-30P5

M- Electrical steel

120- Specific loss W/kg at 1.7 T x 100

30 -Thickness in mm x100
P - High permeability Grade (Hi-B) or S- Conventional grade

5 - Frequency in Hz x1/10

Each thickness of lamination has 2-5 grades of steel based on specific loss level, under S & P
categories.

2. Measurement of magnetic properties

One of the earliest hurdles faced by transformer engineers was how to measure the magnetic
properties (specific losses W/kg and specific excitation VA/kg) of soft iron samples taken
from a lot. This was essential to correctly assess the expected losses during the design stage.
After lot of trials, a German engineer I. Epstein came up with a novel idea in 1900 to assess
the properties of electrical steels. Build a small magnetic square (25 Cm Epstein frame as of
today) using 24 numbers 30x 280-320 mm rectangular sample steel laminations (after stress
relieving by annealing) with interleaving (double lapping) at four corners. Turns are wound
around the sides of the frame to excite the core and to find out the losses and excitation kVA
at different flux densities. (I.Epstein, Elektrotechnische Zeitschrift 21, Page: 303, 1900).
After several refinements, this test is codified in IEC 60404-2 ed3.1-1996+AMD 2008 /
ASTM 343- A343 M-14(2019) / IS 649 “Measurement of magnetic properties of electrical
steel and strips by Epstein Frame.”

Specific loss of CRGO laminations is guaranteed, based on the Epstein test value and not
based on actual watt loss measurement in transformer. When Japanese developed laser
irradiated (domain refined, high permeability) CRGO during late 1970’s they faced a
problem. This type of steel will get affected by annealing, essential requirement for Epstein
test samples. At mill, Epstein samples will be made and laser scribed after annealing and then
measure Epstein W/ kg for furnishing the guaranteed value. They developed single sheet
tester (SST Method) to measure the W/kg of such steels. IEC 60404-3 ed3.0- 1992+AMD
2009 / ASTM A 804-A804M-04(2021)/ IS 3024 Measurement of magnetic properties of
electrical steel and strips by Single sheet tester. Sample size for single sheet tester-500x500
mm.

But the W/ kg from single sheet tester will be different from W/kg from Epstein test, the basis
for the guaranteed values. So, the need for a correction factor arose while comparing test
values as per these two methods. Last 30 years this was a problem and different correction
factors were used (SST-82 & SST-92 factors based on the standards on single sheet test
issued in 1982 & 1992). SST-92 test values are generally higher than Epstein test values by
2-10 %. This problem was sorted out by IEC in 2017 by issuing the IEC Technical Report TR
62981-2017 “Comparison of magnetic measurements on CRGO –single sheet test vs Epstein
Frame Test-Technical Report”, where the single sheet loss correction factors for different
grades of CRGO laminations are given to arrive at guaranteed Epstein test value from single
sheet test values.

3. Grounding of Core

Transformer core is built with very thin (0.18-0.35 mm thick) CRGO laminations with
inorganic insulation coating on both sides. Hence each lamination is electrically floating
during service. Usually only a single point earthing is provided from the core lamination
stack to tank using a copper strip or cable. Then how individually insulated laminations are
getting grounded? The capacitance between the individual laminations is very high due to the
large surface area between laminations and extremely small insulation gap between metal of
the laminations. Capacitive impedance between laminations becomes small, due to the high
value of capacitance between individual laminations plus high frequency components in
surge. Thus, during voltage surge conditions, these capacitances will get bridged and
laminations get effectively grounded. Whenever cooling ducts or insulation boards (Nomex)
are inserted in the lamination stack, a bridging copper piece or metal screw will be provided
to shunt the gap as capacitance there will be low.

How the individually insulated lamination stack will behave when a high voltage AC is
applied on it? This will happen when the innermost winding gets buckled due to high radial
forces from the short circuit current in the winding. During such a failure mode, bare LV
winding conductor (due to rupture of conductor insulation) may touch the earthed core and
heavy current will flow through the core, puncturing the insulation coating on laminations.
This current, spread through several paths from the point of fault to core grounding lead, will
melt core laminations, creating small holes in laminations all through these current paths.
A54-Knowledge tit-bits on Transformer
Electrical Steels-Part 2
 Published on December 3, 2021

P Ramachandran

Specialist, Power Transformers

64 articles

This is in continuation to Part-1 of this article where we found some knowledge nuggets on
grade designation for CRGO silicon steel; measurement of magnetic properties; grounding of
laminations in a core stack. In this part we will consider aspects such as mixing of different
grades of CRGO steel; heat dissipation from lamination stack; composition of electrical steel
and silicon content; insulation coating on laminations; burr height on edges and significance
of magnetic polarisation in CRGO steel.

A53-Knowledge tit-bits on Transformer Electrical Steels-Part-1

https://www.linkedin.com/pulse/a53-knowledge-tit-bits-transformer-electrical-p-
ramachandran

1. Mixing laminations of different grades

It is not a good engineering practice to mix steel laminations of different grades as parallel
path in a core. Steel with higher permeability will carry more flux and results in higher flux
density (and more losses) in it than the nominal flux density value considered. In such a case
of mixing, wider sheets shall be of better grade than smaller width sheets due to the higher
magnetic sensitivity of better grades to mechanical stresses from handling.

For wound core applications, core loss will be lower if the inner wraps are wound with higher
permeability steel and outer wraps with lower permeability steel. The core loss will be higher
if the mixing is reversed.

2. Heat dissipation from core laminations.

 When the transformer is energized, heat will be generated continuously in the core due to
eddy and hysteresis losses in the core. When transformer is loaded, the leakage flux from
winding will enter in to core, creating local flux saturation in core limb or yoke (depending
on direction of power flow) by adding leakage flux to main flux. Typical iron loss in core in
kW is about 0.30 % (10kVA), 0.07 % (20MVA), 0.04 % (100 MVA), 0.012 % (1500 MVA
bank) of the transformer rating in kVA. This heat generation will not be uniform in all parts
of the core. Loss per unit volume will be more at corner joints of core (150-175 %) due to the
flux movement at cross direction to the lamination rolling direction. Loss in W/kg in the 45-
90-degree cross direction to the lamination rolling direction can be 4 times the loss value in
the rolling direction. This increase in cross directional specific core loss will be more for
better grades of steel.

In large transformers or auto-transformers with very low co-ratio, high levels of leakage flux
from windings can enter the core surface and create hot spots and circulating currents on the
core surface. Manufacturers take care of such situations through special means- slitting the
front packets of core or copper shielding.

In core type transformers, maximum core temperature will be in the yoke, above the middle
limb. The maximum permissible temperature inside core is 130 degree C during normal
cyclic loading, as per Table 2 of IEC 60076-7 Loading Guide. Core lamination insulation
coating can withstand much higher temperatures, but above temperature limit is stipulated to
avoid gas generation (hydrogen and methane) from the breakdown of thin oil film between
laminations (occurs at about 140 degree C).

Almost the entire heat generated in core is dissipated through the lamination edges and very
little in the perpendicular direction as the insulation coatings on each lamination prevent the
heat transfer. Hence designers always provide cooling ducts on pressboard components that
directly touch the core edges (eg. lower yoke) to take away the heat through oil flow. Where
ever solid insulation touches the core, surface temperature of core is limited to the maximum
winding hot spot.

3. Silicon content in CRGO steel – Material Composition of Transformer steel

During the early days, transformer cores were made from thin sheets of pure wrought iron.
Soft iron made from Swedish sources was popular due to their better magnetic properties.

Then around 1900, it was found that adding 3-6 % of silicon or aluminium improved the
magnetic characteristics of steel. It increased the electrical resistivity, thereby reducing eddy
losses in steel. Adding 0.03- 0.07 % of carbon improved the texture of hot rolled steel sheets
by promoting formation of an austenite phase.

When cold rolled steel was made in early 1940’s, with grain refinement in the rolling
direction, it was discovered that to get loss reduction, grain formation in direction other than
the Goss direction has to be stopped. Armco (USA) found out that adding Manganese
sulphide (MnS) plus aluminium nitride (AlN) inhibited such grain growth in CRGO during
the rolling process. Alternate inhibitors were developed in later years eg MnSe + Sb
(Kawasaki, Japan) and MnS+ BN (Alleghney, USA). Generally, for conventional CRGO,
MnS inhibitors are added and for high permeability material, MnS + AIN is used.
Modern day CRGO grades contain 2.9 -3.2 % of Silicon. Higher % of silicon is not preferred
due to difficulties in rolling as steel becomes more brittle with higher silicon content.

Japanese have developed special silicon steel with silicon content of 6.5 %. The chemical
vapor deposition of the silicon allows higher silicon content without the problems of
brittleness during the rolling operation. Apart from reducing eddy loss, high silicon content
also brings down magnetostriction, the cause for transformer noise.

4. Insulation on laminations

During the first three decades of transformer engineering, 0.5 mm thick soft iron sheets were
used in transformer cores with paper insulation between laminations. Later laquers or
varnishes were used for insulation. By 1950s an inorganic insulation coating comprised of
magnesium silicate was developed. This is formed on the lamination surface by the reaction
of annealing separator with the steel surface during high temperature annealing. This coating
is sometimes called as ‘mill glass’ or Forsterite glass film coating, even though the coating is
not strictly glass. This EC-2 coating (designation as per IEC 60404-1-1) is sufficient for
wound cores but an additional magnesium phosphate or silicate-based coating is required for
stacked core laminations of power transformers. This is termed as EC-5-G coating over EC-2
(C5 coating over C2 as per ASTM 976). This can withstand stress relief annealing treatment
at a temperature up to 845-degree C. The coating thickness should be sufficient for the
induced voltage per turn but it should not be so thick as to reduce the space factor of core. At
a particular flux density, adequacy of surface insulation for steel is a function of lamination
width. Wider laminations require a better insulation coating. Modern insulation coatings
create a stretch tension on lamination surface that reduces sensitivity of steel to mechanical
stresses, reduces iron losses and magnetostriction.

Adherence of insulation coating to the base metal is checked by a bend test (ASTM A 720).
A strip is bent at least 90 degrees around a 20-22 mm diameter mandrel and insulation
looseness on concave surface shall be less than 15 %.
The insulation coating resistance expressed in Ω. mm2 represents the electrical resistance
offered to the passage of current through the coating. The measured insulation coating
resistance before or after the possible application of a stress relief heat treatment shall be not
less than 500 Ω. mm2 (Clause 7.3.5 Insulation coating Resistance - IEC 60404-8-7-2017
CRGO Strips and sheets; Test Method IEC 60404-11 & ASTM A 717 &737)

5. Permissible Burr height in lamination edges

The determination of the burr height applies only to slit coils delivered in the width in which
they will finally be used. The measured burr height shall not exceed 0,025 mm. (25 micro
meters) (Clause 7.2.7 Burr height - IEC 604-8-7 -2017 “CRGO strips and sheets”) A high
burr height will cause shorting of adjacent laminations, resulting in circulating currents and
extra eddy losses.

Nearly half a century back, slitting cutters were of high carbon steel and the burr heights on
the slitted coils were excessively high. My company then used to grind the edges of slit
laminations to remove burrs and anneal each lamination in a continuous annealing oven to
relieve the stress. Today silicon carbide or diamond cutters are used for slitting and typical
burr height values for good laminations are of the order- 90 % of laminations with less than
10 micro-meters and 100 % less than 20 micro- meters. Measurement is made at lamination
edges, at every 50mm on a sample (one metre long), using a hand-held micro
meter. Measurement method is as per IEC60404-9

6. Magnetic Polarization

In National Standards for CRGO (Cold Rolled Grain Oriented Silicon Steel) like IEC 60404-
8-7-2017, there is a requirement of “Minimum magnetic polarization for Peak magnetic field
strength H=800 A/m “

The word polarization means magnetization ie flux content (similar to current density in
electric circuit) at a particular magnetic field strength (similar to voltage). So, when standard
asks for a minimum value of polarization at a particular field strength, it indirectly gives the
magnetic conductivity (permeability) of the electrical steel. So normal CRGO steel has a
minimum magnetic polarization of 1.78- 1.8 T and high permeability CRGO steel has 1.85 -
1.88 T at a peak magnetic field strength of 800 A/m. Looking it another way, it means you
require less magnetizing force with high permeability cores to achieve the same flux density.

B= µ0 (permeability of vacuum) x µr (relative permeability) x H and polarization J = B - µ0

(ie magnetic constant or the permeability of vacuum) x H
A59-Knowledge tit-bits on Transformer
Electrical Steels-Part 3
 Published on January 29, 2022

P Ramachandran

Specialist, Power Transformers

66 articles

In previous parts of this article, we covered topics such as CRGO silicon steel grade
designation; measurement of magnetic properties; grounding of laminations in a core stack;
mixing of different grades of CRGO steel; heat dissipation from lamination stack;
composition of electrical steel; insulation coating on laminations; burr height on lamination
edges and magnetic polarization in CRGO steel. In this part, let us discuss about iron loss
building factor; magnetic ageing of core; steel grades in transformer specifications;
magnetostriction and noise.

A 53-Knowledge tit-bits on Transformer Electrical Steels-Part-1

https://www.linkedin.com/pulse/a53-knowledge-tit-bits-transformer-electrical-p-
ramachandran

A 54-Knowledge tit-bits on Transformer Electrical Steels-Part 2

https://www.linkedin.com/pulse/a54-knowledge-tit-bits-transformer-electrical-2-p-
ramachandran

1) Core loss Building Factor

It is the ratio between W/kg from transformer test bed (measured no load loss / net weight of
silicon steel used) to the W/kg from actual Epstein test value or the value given in test report
from CRGO maker. Building factor will be different in limb, yoke and at joints of the
core. At joints, it will be several times higher than in the limbs. But average value as
estimated above is sufficient for quality control, validating the designs or improving core
design and construction.

Core loss building factor = Total measured core loss (W) / {specific iron loss as per Epstein
test for the working flux density and frequency (W/kg) x Core weight (kg)}

Core loss building factor will be always more than one and affected by both magnetic and
manufacturing factors.

Magnetic Factors- Depending on the core construction, non-uniform flux distribution and
distorted flux waves; Grade of steel (superior the grade, more is building factor); holes in
laminations; type of mitred joint at corners (step-lap joints reduce building factor); core
dimension (a higher limb height results in lower BF; Flux density, core construction (wound
core have lower BF than stacked core; a 3 phase 5 limbed core has higher BF than a 3 phase 3
limbed core)

Manufacturing Factors– mechanical stress from slitting (wider the lamination, less is building
factor), cutting and handling of the laminations; building of the core; burr height at edges,
insulation damage at slit/cut edges; Very high or very low core clamping pressure increases
the building factor.

2) Magnetic Ageing of CRGO

Magnetic “ageing” was a troublesome phenomenon in transformers till 1900 when change
over from soft iron sheets to silicon steel started in core. Hysteresis loss in iron sheets used to
increase by two to three times within a short time after putting the transformer in to service.
This used to cause overheating of transformers and “core fires” in dry type transformers was
common. This was overcome with the introduction of silicon steels.

Today’s transformer cores have stable magnetic characteristics and iron loss will not increase
with service age, thanks to better silicon steel and superior insulation coatings developed.
Some transformer manufacturers may claim that no-load losses will increase with service, but
this is not true or acceptable.

To check the stability of magnetic characteristics, a magnetic ageing test is used. When
samples for Epstein test are aged at 150 degrees C heating for 216 hours (9 days) the increase
in iron loss shall be less than 3 % when tested in 25 cm Epstein frame. Equivalent accelerated
ageing tests are 225 degrees C for 24 hours or 200 degrees C for 100 hours.
3) Should I specify CRGO grade while buying transformer?

From a user’s angle, there is no need to specify any specific grade of CRGO in specifications.
Of course, he can indicate or prefer any specific make at the time of order or design
review. Designer selects the grade based on the working flux density selected, marginal
savings in W/kg versus marginal increase in price for better grade. User is benefited only by
lower losses and not by using a better grade of core steel.

When Hi B grades were introduced in 1970s, many users were carried over by these new
grades and started specifying Hi B grades while keeping the maximum flux density as 1.55 T
in their specifications. At such low flux densities W/kg reduction with better grades is
negligible but extra cost is incurred without any compensating reduction in total iron loss.

4) Magnetostriction

Magnetostriction is one of the magnetic properties that accompany ferromagnetism. It causes

reversible deformations of a material body due to the magnetization arising from an applied
magnetic field. Magnetostriction of electrical steels is recognized as the main cause for the
acoustic noise generated from transformers. Standardization of magnetostriction
measurement methods is necessary to find means to mitigate it and there by transformer
noise. Transformer noise is considered as an environmental pollution, especially in urban
areas.

When a core is magnetized by alternating magnetic field, core laminations increase in length
in the direction of magnetization and contract at right angles there to. This change is called
magnetostriction. This dimensional change can be positive or negative. Magnetostriction is a
magneto-mechanical phenomenon which accompanies the change of the volume fraction of
magnetic domains which have a certain magnetic orientation with respect to the direction of
the applied magnetic field. This phenomenon is intrinsically sensitive to stress. The stress
sensitivity is dependent on material conditions such as grain orientation, residual stress, and
coating tension.

The magnetostriction of electrical steel is increased by compressive stresses in the

magnetizing direction rather than tensile stresses. Its magnitude depends on type and grade of
electrical steel, the induction level, silicon content in steel, direction of magnetization with
respect to direction of steel rolling. A silicon content of 6.5 % reduces magnetostriction to
almost zero but such high silicon content affects workability. In normal CRGO with silicon
content of less than 3 %, magnetostriction depends on the quality of steel. A low loss steel
has small and negative magnetostriction, while high loss grade steels have larger and mainly
positive magnetostriction. A high degree of preferred grain orientation, flatness, and
uniformity of magnetic properties in steel give lower magnetostriction.

Magnetostriction measurement on lamination samples can be done by any of these

standards- IEC TR 62581 ed1.0-2010 - by Single sheet and Epstein test specimens using
optical sensors and accelerometers.; IEC 60404-17:2021 - Measurement of magnetostriction
by means of a single sheet tester and an optical sensor

5) Excitation Current Ratio

 Some users ask for the ratio of transformer excitation currents at 100 % and 110 % of rated
voltage. Earlier days this factor was 2-3 due to the low flux densities used (due to limitations
of CRGO available then) and 90 degrees over lapped joints at the corners. Today this factor is
4-10 due to the higher flux densities employed (compared to the saturation flux density),
step-lap mitred joints at corners for power transformers and single-phase wound cores for
distribution transformers. This factor will be less with Hi B and domain refined core grades
compared to conventional steels. Some users consider this factor as a means of indirectly
checking the margin of working flux density to the saturation flux density of the core steel
used.

6) Units & Conversions

Field Strength:

1Oersted = 80 A/m

Flux Density:

1Tesla (Wb/m2) =10,000 gauss (lines/cm2)

(2022-01-29)