Design and Testing of a Diode-Based Electric Field Probe Prototype

Zlatko Živković, Damir Senić, Antonio Šarolić, Ante Vučić

FESB – Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture
Ruđera Boškovića 32, 21000 Split, Croatia
E-mail: [email protected], [email protected], [email protected]

examples of amplitude modulated (AM) signals with

Abstract: The paper presents the design and the response
measurements of the simple, one-axis diode-based electric field different modulation frequencies and pulse modulated (ASK
probe, consisting of short dipole, diode detector, low-pass filter, and GSM) signals with different durations and duty cycles
transmission line and monitoring instrument. Frequency were applied, since it has been noticed that these signals
response measurement results were compared to the simulation cause significant measurement errors [7], [8].
results, where simulation was achieved by a method of
combining the numerical FDTD simulation with dipole/diode
circuit simulation. The measured results showed good 2. ELECTRIC FIELD PROBE DESIGN
agreement with simulation results in the operational range.
Additionally, the measurements of the probe response to
The diode based electric field probe generally consists of
amplitude modulated and pulse modulated signals were
performed. As expected, the response to the modulated signals five main components [2]: short dipole printed on dielectric
showed considerable deviation from the sinusoidal CW substrate, nonlinear detector (zero-bias Schottky diode)
response. Measurement error for such signals was calculated connected between dipole arms, low-pass filter, resistive
and presented. transmission line and monitoring instrumentation (Fig. 1).
The incident RF electric field induces the RF oscillating
voltage Uoc on the short dipole. The DC component of the
1. INTRODUCTION diode detector voltage (Ud) is then proportional to the voltage
Uoc induced between the dipole arms. The resistive
The electric field probes, based on the diode-loaded short transmission line (which also acts as a low-pass filter)
dipoles, are widely used in various aspects of EMC transmits the DC signal component from the diode to the
measurements, including the radiation hazard survey [1]-[4]. monitoring instrument. Its high resistance per unit length
These probes represent wideband, non-selective ensures reduced incident field reception and scattering by the
instrumentation and, thus, are mainly used for quick, line. The additional low-pass filter prevents RF voltages to
wideband measurements in free-space or even in human reach the diode detector from the wrong side.
tissues [3], [4]. The modern diode-based probes should
exhibit very wide operational frequency range (from few
hundreds kHz to few GHz), flat response (up to ±3 dB) in the
defined operational range, wide amplitude dynamic range and
isotropic response (for three-axes probes). Furthermore, the
probe has to be electrically small since it must not perturb the
field distribution.
Figure 1 – Electric field probe – schematic view
The fabrication of simple, one-axis electric field probe is
presented in this work. The frequency response and probe
linearity for sinusoidal continuous wave (CW) signal were The fabricated electric field probe is presented in Fig. 2.
measured in Gigahertz Transverse Electromagnetic (GTEM) The dipole with half-length l = 2 cm was printed on
cell. To make sure that the measured CW values were electrically thin ( h = 1.6 mm ) FR4 dielectric substrate with
correct, the probe was also modeled by combining numerical relative permittivity ε r = 4.6 . The zero-bias Schottky diode
SEMCAD X simulation software [6] and NI Multisim (BAT62-03W) was soldered between the dipole arms. The
dipole/diode system equivalent circuit. As it has been showed low-pass RC filter consisted of two resistors (1 MΩ each) in
that the measurement error of the diode based electric field combination with 10 pF capacitor which ensured the cut-off
probes significantly depends on strength and modulation of frequency of fc = 8 kHz . Each lead of the parallel resistive
the applied signal [7]-[10], the additional error measurements
for large modulated signals were performed. Several transmission line was realized as a large resistance of 20 MΩ,
This research was supported by the Ministry of Science, Education and which yielded the overall transmission line resistance of
Sports of the Republic of Croatia (Project No. 023-0000000-3273 and 40 MΩ.
No. 036-0361630-1631).
 where Prec is the received power, Aef is effective area, S is
incident wave power density, λ is desired wavelength, E is
the electric field strength set to 20 V/m, and D is directivity
obtained from SEMCAD X simulations for different
frequency values. Calculated Uoc, along with RA and CA
obtained from SEMCAD X, were used in NI Multisim
simulation for each different frequency. The output DC
voltage was captured at each frequency and later compared to
Figure 2 – Fabricated electric field probe
the measured one.

3. SIMULATION AND MEASUREMENTS OF

ELECTRIC FIELD PROBE CHARACTERISTICS

For the simulation purposes the NI Multisim model

combined with SEMCAD X simulation results was used.
The simple developed SEMCAD X model is presented in Figure 4 – Simplified NI Multisim model of electric field
Fig. 3. The simulations were used to obtain directivity and probe
the input impedance of printed dipole antenna. Since the
transmission line, low-pass filter and the diode should not All measurements were performed in TESEQ 750 GTEM
considerably change the directivity and input impedance of cell having the maximum septum height of 75 cm. The
the dipole, and for the sake of simplicity, they were omitted schematic layout of measurement setup is shown in Fig. 5.
from the simulation. The dimensions and electrical The input CW signal was generated by Rohde & Schwarz
parameters of simulated dipole antenna on dielectric substrate SM300 signal generator and amplified by RF power
were identical to the real, fabricated model. amplifiers AR 150W1000 (80 – 1000 MHz) and Ophir 5140
(0.7 – 3 GHz). The produced electric field probe (DUT) was
placed inside GTEM cell on styrofoam base, inside the
defined test volume, next to the commercial HI-4455
isotropic probe, which was used to control the electric field in
the vicinity of DUT. To avoid the polarization loss, the probe
was oriented parallel to the incident electric field. The tested
probe was connected to the HP 3490A digital multimeter
Figure 3 – Simulation model of the printed dipole on which was used as DC voltmeter. Due to fact that DC voltage
substrate is measured at the end of the highly resistive transmission
line (40 MΩ), the measuring instrument input resistance has
The dipole/diode equivalent circuit was created using NI to be high enough (>GΩ) to ensure accurate results.
Multisim. All relevant parameters of the designed probe were Several measurements were performed using described
included into equivalent circuit model, presented with Fig. 4, setup. For the probe response analysis in the frequency range
where Uoc presents the rms open-circuit voltage across dipole from 100 MHz to 2.4 GHz, the value of the incident electric
input terminals, RA is dipole resistance, and CA is dipole field was kept constant at 20 V/m. The measured DC voltage
capacitance. The low-pass filter and the transmission line at probe’s terminals showed flat response within ±3 dB up to
were omitted from equivalent circuit since they should not 1 GHz, with the value of measured voltage of approximately
affect the output steady-state DC voltage. Dipole impedance 0.35 V (presented with solid blue line in Fig. 6). This is in a
was obtained from SEMCAD X simulation results and Uoc good agreement with simulation results in the proposed
voltage was then calculated from well-known equations: operational range (red dashed line in Fig. 6).
Prec = Aef ⋅ S , (1) The linearity of the electric field probe was measured at
100 MHz for incident electric field strength from 5 V/m to
185 V/m, as shown in Fig. 7. Generally, the diode-based
U oc2 λ2 E2 electric field probes operate in two different regimes,
= ⋅D⋅ , (2)
4 RA 4 π 120π depending on the incident signal strength. If α·Ud >> 1 [3],
(where α is a diode specific parameter and equals to 38.7 for
λ ⋅ E D ⋅ RA this diode, and Ud presents the DC component of the diode
U oc = ⋅ , (3) detector voltage) the diode operates in large signal regime. In
π 120 the large signal regime the rectified DC voltage is linearly
proportional to the amplitude of the incident field, while for
the small signals it is proportional to the square of the 4. MEASUREMENTS OF ELECTRIC FIELD PROBE
incident field amplitude. For the intermediate signals, the DC RESPONSE TO MODULATED SIGNALS
voltage exhibits neither purely linear nor purely quadratic
proportionality. Due to the applied incident field strengths In the small signal regime, the probe will respond to the
and the resulting induced voltages, the electric field probe true RMS field strength, while in the large signal regime it
operated mainly in the large signal regime. Measured will respond to the peak field strength, even if the measured
linearity response, shown in Fig. 7, was quite linear for signal is of the complex waveform [7]-[10]. Hence, the
incident electric field from 5 V/m to 185 V/m. significant measurement error could be expected for the large
modulated signals, especially for amplitude modulated (AM)
or time-division multiple access (TDMA) signals, such as
pulsed radar signals or GSM signals.
As observed in previous section, fabricated probe operates
in large signal regime for the applied incident field strength
dynamic range. Thereby, significant error in the probe
detection is expected when the probe is exposed to modulated
signals. To verify the former, the probe has been exposed to
various modulated signals; AM, ASK and GSM TDMA, and
measurement error was monitored.
According to [8] and [10], measurement error (Δ) is
expressed as logarithmic ratio of the field strength meter
reading Edisplay and true rms field strength:
Figure 5 – Schematic layout of measurement setup
2
Edisplay Edisplay
Δ = 10 log10 2
= 20 log10 (4)
E rms Erms

To determine the probe measurement error, it is necessary to

compare the displayed field strength Edisplay with the true rms
value of field strength Erms as defined in (4).
According to [7], the measurement setup should ensure
controllable waveform, amplitude and frequency, and
measurement of true rms value of incident electric field.
Controllable waveform is necessary to estimate the probe
response for various waveforms, as the measurement error is
expected to be dependable of waveform parameters. Beside
the waveform, the measurement error is also dependable on
the incident electric field amplitude. This is significantly
Figure 6 – Frequency response emphasized at the upper part of probe’s amplitude range
where the probe's response should be purely linear. Hence, it
is necessary to produce high fields with magnitude of
100 V/m or even higher.
Measurement of true rms value of incident field could be
achieved by measuring the true average power from
amplifier. This power is proportional to the true rms value of
incident electric field. For this purpose the Rohde-Schwarz
NRP Z21 true average power sensor was used.
The measurement error should not depend on carrier
frequency as long as the frequency is inside the probe’s
working range.
Used measurement setup for measuring probe response to
various modulated signals, was similar to setup used for
measuring probe’s characteristics, shown in Fig. 5. Only
Figure 7 – Measured linearity response at 100 MHz
difference is the directional coupler that is connected between error was in all cases considerably lower than for ASK
the amplifier and GTEM cell. Its coupled port was used for signals.
power measurements. The reflected power from GTEM cell As expected, the error increased at higher incident field
was also measured with and without probe inserted, in order amplitudes. However, it varied by only 1 dB from 20 to 100
to check for possible field perturbations caused by probe’s V/m, which again confirmed that the probe was well into the
presence. Modulated signals were monitored with spectrum linear regime already at 20 V/m.
analyzer and oscilloscope to make sure that amplifier did not
cause the signal deformation. No deviations were observed
during this testing procedure.
The initial measurements were performed using CW signal
adjusting its electric field magnitude to the predefined value.
For every level of the predefined incident electric field value,
the true average power was measured and DC voltage at
probes terminals was noted. This presented the reference
level, corresponding to the true rms field strength in (4).
After that, the different types of modulation were
generated with Rohde & Schwarz SM300 signal generator.
Employing modulation, instead of CW signal, generally
causes change in power reading depending on modulation
parameters. By setting the modulated signal power level
(measured by true average power sensor) to the average
power of the CW signal, the rms field strength of the
modulated signal should be the same as the rms field strength Figure 8 – Measured DC voltage at probe’s terminals as a
of the CW signal. However, the measured DC voltage was function of incident true rms field value
generally different and measurement error could be
calculated according to (4). For this calculation, the output
voltage was used instead of measured electric field in (4).
The measurements were performed for three different
modulation types: AM with 10 kHz, 40 kHz, and 80 kHz
modulating frequency and modulation index of 80%; ASK
with 1 kHz repetition frequency and duty cycles of
DC = 1/10 and 1/2, and finally GSM TDMA scheme. The
GSM TDMA scheme was achieved as ASK with 217 Hz
repetition frequency and duty cycle of DC = 1/8.
The carrier frequency was set to 100 MHz for all
modulations – the error should not depend on the carrier
frequency but only to the waveform and modulation
parameters.
Results of measured voltage for different modulations
compared with CW signal is shown in Fig. 8, and probe’s Figure 9 – Electric field probe measurement error
measurement error for three types of modulation with
different parameters is shown in Fig. 9. Significant error in
probe’s display could be observed depending on the type of 5. CONCLUSION
the applied modulation.
The electric field probes, based on diode-loaded short
Error was highest for the ASK signals with small duty
dipole, are often used in EMC measurements including
cycles (ASK with 1/10 duty cycle and GSM with 1/8 duty
radiation hazard surveys. The simple one-axis design of such
cycle) and reached -9dB. The negative sign presents the
electric field probe was presented in this paper. Its frequency
underestimation error. This is especially dangerous since it
response and linearity response for sinusoidal CW signal
can lead to undetected overexposure i.e. radiation hazard.
were measured and presented. The measurement results
Considering the amplitude modulation, the error increased showed good agreement with model developed by a
as the modulation frequency decreased. However, the AM combined numerical FDTD simulation and dipole/diode
circuit simulation.
 In addition, the probe's response to amplitude and pulse
modulated signals was measured in terms of measurement
error. The obtained results showed significant deviation
compared to CW probe response. This leads to considerable
measurement errors when measuring such modulated signals.

REFERENCES

[1] Motohisa Kanda: “Standard Probes for Electromagnetic

Field Measurements”, IEEE Transactions on Antennas and
Propagation, October 1993, Vol. 41, p.p. 1349-1364.
[2] Howard I. Bassen, Glen S. Smith: “Electric Field Probes
– A Rewiev”, IEEE Transactions on Antennas and
Propagation, September 1983, Vol. 31, p.p. 710-718.
[3] Ronold W. P. King, Glenn S. Smith: “Antennas in Matter:
Fundamentals, Theory, and Applications”, The MIT Press,
1981.
[4] Thomas Schmid, Oliver Egger and Niels Kuster:
“Automated E-Field Scanning System for Dosimetric
Assessments”, IEEE Transactions on Microwave Theory and
Techniques, 1996, Vol. 44, p.p. 105-113.
[5] Motohisa Kanda, Lanny D. Driver, “An Isotropic
Electric-Field Probe with Tapered Resistive Dipoles for
Broad-Band Use, 100 kHz to 18 GHz”, IEEE Transactions on
Microwave Theory and Techniques, 1987, Vol. 35, p.p. 124-
130.
[6] SEMCAD X by SPEAG, www.speag.com
[7] Antonio Šarolić, Borivoj Modlic, “Measurement of
Electric Field Probe Response to Modulated Signals Using
Waveguide Setup”, IEEE Antennas and Wireless Propagation
Letters, October 2010, Vol. 9, pp. 1041-1044.
[8] J. Randa and M. Kanda, “Multiple-source, Multiple-
frequency Error of an Electric Field Meter,” IEEE
Transactions on Antennas and Propagation, Jan. 1985, vol.
AP-33, pp. 2–9.
[9] Antonio Šarolić, Borivoj Modlic and Dragan Poljak,
“Electric Field Probe Measurement Error Due to Sine Signal
Calibration”, 2005 International Conference on Software,
Telecommunications and Computer Networks - SoftCOM
2005 (CD ROM), September 2005, Split, Croatia, pp. 1-5.
[10] Antonio Sarolic, Vesna Roje, and Borivoj Modlic,
“Measurement of Electric Field Probe Error for Pulsed
Signals,” in Proc. IEEE Int. Symp. Electromagn. Compat.,
Portland, OR, 2006, vol. 2, pp. 244–248.