Article

Development of a Multi-Channel Ultra-Wideband

Electromagnetic Transient Measurement System
Shaoyin He * , Xiangyu Chen, Bohao Zhang and Liang Song

State Key Laboratory of Electrical Insulation and Power Equipment, School of Electrical Engineering, Xi’an
Jiaotong University, Xi’an 710049, China; [email protected] (X.C.);
[email protected] (B.Z.); [email protected] (L.S.)
* Correspondence: [email protected]

Abstract: In complex electromagnetic environments, such as substations, converter stations

in power systems, and the compartments of aircraft, trains, and automobiles, electromag-
netic immunity testing is crucial. It requires that the electric field sensor has features such as
a large dynamic measurement range (amplitude from hundreds of V/m to tens of kV/m),
a fast response speed (response time in the order of nanoseconds or sub-nanoseconds),
a wide test bandwidth (DC to 1 GHz even above), miniaturization, and robustness to
strong electromagnetic interference. This paper introduces a multi-channel, ultra-wideband
transient electric field measurement system. The system’s analog bandwidth covers the
spectrum from DC and a power frequency of 50 Hz to partial discharge signals, from DC to
1.65 GHz, with a storage depth of 2 GB (expandable). It overcomes issues related to the
instability, insufficient bandwidth, and lack of accuracy of optical fibers in analog signal
transmission by using front-end digital sampling based on field-programmable gate array
(FPGA) technology and transmitting digital signals via optical fibers. This approach is
effectively applicable to measurements in strong electromagnetic environments. Addition-
ally, the system can simultaneously access four channels of signals, with synchronization
timing reaching 300 picoseconds, can be connected to voltage and current sensors simulta-
neously, and the front-end sensor can be flexibly replaced. The performance of the system
is verified by means of a disconnect switch operation and steady state test in an HVDC
converter station. It is effectively applicable in scenarios such as the online monitoring of
Academic Editor: Arcady Zhukov
transient electromagnetic environments in high-voltage power equipment, fault diagnosis,
and the precise localization of radiation sources such as partial discharge or intentional
Received: 27 December 2024
Revised: 7 February 2025
electromagnetic interference (IEMI).
Accepted: 12 February 2025
Published: 14 February 2025 Keywords: transient electric field measurement; ultra-wideband electromagnetic pulse;
Citation: He, S.; Chen, X.; Zhang, B.;
electric field probe; field programmable gate array
Song, L. Development of a
Multi-Channel Ultra-Wideband
Electromagnetic Transient
Measurement System. Sensors 2025, 25, 1. Introduction
1159. https://doi.org/10.3390/
Electromagnetic radiation fields are present during processes such as lightning strikes,
s25041159
high-voltage switch operations, static electricity build-up, and partial discharge caused
Copyright: © 2025 by the authors.
by insulation damage. These electromagnetic radiation fields are characterized by high
Licensee MDPI, Basel, Switzerland.
intensity, rapid variation, and a broad frequency spectrum, and they can couple into
This article is an open access article
distributed under the terms and
electronic systems through antennas, apertures, cables, and other paths [1,2], potentially
conditions of the Creative Commons causing temporary or permanent damage to various electronic devices and systems [3–6].
Attribution (CC BY) license The measurement of transient electromagnetic fields is a crucial aspect of electromagnetic
(https://creativecommons.org/ field protection research. It plays a vital role in studying the propagation characteris-
licenses/by/4.0/).

Sensors 2025, 25, 1159 https://doi.org/10.3390/s25041159

Sensors 2025, 25, 1159 2 of 18

tics of transient electromagnetic fields, developing protection methods, and evaluating

their effects.
Currently, there are two main approaches to measuring a transient electric field—
the first one is based on the photoelectric effect, where the electric field signal is con-
verted into an optical signal for measurement. However, optical electric field sensors
have complex manufacturing processes, are significantly affected by external environmen-
tal factors such as temperature, and have a limited dynamic range, which restricts their
widespread application.
The other approach involves electric field sensors based on the principle of elec-
tromagnetic induction. Such sensors have been applied previously in the field of EMP
measurement, and the related measuring technology is more mature. Baum and others sys-
tematically summarized nuclear EMP measurement technology [7]. As EMP measurement
efforts have progressed, the International Electrotechnical Commission has successively for-
mulated and published a series of international standards related to HEMP measurements,
including radiation interference, conduction interference, immunity, and high-power tran-
sient parameter measurements [8–11]. Subsequently, Montena Ltd. (Rossens, Switzerland)
commercialized the production of gradual conical dipole antennas, with the parameters of
their products adjusted according to different frequency requirements, reaching an upper
cutoff frequency of 10 GHz. In 2007, Koshelev and others from the Russian Institute of
High-Current Electronics designed a two-dimensional active electric dipole antenna for
measuring ultra-wideband pulses, with a working bandwidth of 0.6 to 4.6 GHz [12]. Baum
et al. proposed a differential electric field sensor known as D-dot for EMP electric field
signal measurement [13]. The output voltage signal of the sensor is proportional to the
differential of the electric field, making it suitable for ultrafast pulse signal measurement.
However, this sensor requires integration of the output signal, which is suitable in the
low-SNR and -frequency scenario. Ref. [14] proposes an ultra-wideband time-domain
antenna coupling electric-field energy through a pair of differential dipoles, with operating
frequency range of 10 MHz~1 GHz.
For measuring the transient electromagnetic environment in power system substations,
a nanosecond-level transient electric field measurement system has been developed, with a
bandwidth range of 1 kHz to 460 MHz [15]. Refs. [16–19] explored the frequency character-
istics of the radiated electromagnetic fields generated by substation switching operations
and their coupling effects on secondary equipment. Their results indicate that the ampli-
tude of the transient radiated electromagnetic fields generated by high-voltage switches
ranges from hundreds of V/m to tens of kV/m, with spectral components spanning from
several hundred kHz to several hundred MHz.
With the deepening research in electromagnetic compatibility, electromagnetic in-
terference, and protection technologies within power systems, measurement systems for
transient electromagnetic fields in power systems are required to have the following charac-
teristics: (1) sufficient bandwidth; (2) good sensitivity across different amplitude levels due
to the signal under test potentially varying significantly in amplitude; (3) a compact and
miniaturized size which should not affect the electromagnetic environment being measured;
and (4) effective shielding, grounding, and filtering measures to suppress common-mode
interference [20].
The current transient electric field measurement systems experience issues such as
insufficient bandwidth, limited precision, and unstable amplitude coefficients. In the
electromagnetic environment of power systems, the spectrum of steady-state signals can
be as low as 50 Hz or even lower. However, existing electric field measurements often
focus on improving high-frequency performance, making them ineffective for measuring
low-frequency signals. In power grids, switching operations of power electronics such
Sensors 2025, 25, 1159 3 of 18

as IGBTs can produce spectral components in the range of several kHz or even lower.
The current measurement systems struggle to effectively address both low-frequency and
high-frequency electromagnetic interference signals [21].
To achieve interference-resistant measurement in complex, strong electromagnetic
environments, fiber optic technology is widely used [21,22]. However, after multiple field
applications, it has been found that directly transmitting analog signals through fiber optic
links presents the following drawbacks: (1) The analog bandwidth is limited by the fiber
optic transmitter and receiver, with the low-frequency cutoff generally only reaching as
low as around 1 kHz, and achieving a high-frequency cutoff of 600 MHz or above is very
challenging [15]. (2) The transient electric field signals in the environment under test have a
very wide amplitude range, from tens of V/m to the kV/m level, or even higher. Within this
range, the conversion accuracy of the fiber optic link for signals with smaller amplitudes is
poor and unstable, leading to significant measurement errors and even making it difficult
to determine the effective amplitude.
With the development of high-speed data acquisition technology, particularly advance-
ments in analog-to-digital converter (ADC) and field-programmable gate array (FPGA)
technologies, digital measurement technology has become possible. Digital measurement
technology allows for long-distance digital data transmission, making data easier to store
and process, and can capture transient signals with high precision. In the UK, Mohamed
et al. developed a novel time-based triggering logic implemented in FPGA, which is used
for locating partial discharges in power cables [23]. Russer et al. developed a real-time
operating time-domain electromagnetic interference measurement system based on data
acquisition and digital signal processing technology, applied in the frequency range of
30 MHz to 1 GHz [24,25]. However, it cannot satisfy the measurement requirements for
power systems in the low-frequency band. Cheng et al. designed a pulse signal data
acquisition system based on FPGA technology. This system uses fiber optics to directly
transmit digital signals, solving the problem of signal distortion in long-distance transmis-
sion of analog signals through cables in nuclear explosion pulse beam measurements [26].
Thus, it is evident that digital measurement technology offers certain advantages over
traditional analog measurement techniques in terms of signal measurement, transmission,
and processing.
To the best of the authors’ knowledge, no studies have been reported on multi-channel
measurement systems designed for ultra-wideband electromagnetic transient measure-
ments which achieve time-sequence synchronization across multiple measurement chan-
nels. Such systems are in high demand for a wide range of critical applications, including
locating and detecting transient electromagnetic radiation sources such as partial discharge
or intentional electromagnetic interference (IEMI), state sensing of power equipment, fault
monitoring and diagnostics based on multi-physical transient signals involving electro-
magnetic fields, overvoltage, and overcurrent, etc.
Conventional measurement strategies transmit analog signals via optical fibers, which
face issues such as unstable amplitude coefficients leading to large measurement errors,
weak electromagnetic interference resistance of analog signals, and limited bandwidth
due to the constraints of the optoelectronic conversion module. In contrast, this paper
recommends digitizing the analog signal at the output of the sensor, i.e., at the front end of
the measurement system, and transmitting the digital signal through an optical fiber link.
This approach not only enhances interference resistance but also effectively overcomes
the bandwidth limitation and unstable amplitude coefficient problems imposed by the
optoelectronic devices. The comparison of measuring strategies and the schematic diagram
are shown in Figure 1.
 Sensors 2025, 25, x FOR PEER REVIEW 4
Sensors 2025, 25, 1159 4 of 18

Analog Signal Communication

Optical
Fiber
Sensor Optical Transmitter Optical Receiver Osilloscope

(a) Conventional measurement strategy based on analog data transmission

Sensor Sensor Sensor Sensor

Photoelectric Signal Display and

ADC FPGA
Conversion Module Analysis

Data Cache

High-Speed Digital Acquisition Data Communication

(b) Proposed measurement strategy based digital data transmission

Figure 1.strategy
Figure 1. The measuring The measuring strategy
comparison comparison
between between
conventional conventional
analog analog data
data transmission transmissio
and
digital data transmission.
digital data transmission.

This paper focusesThis on the focuses

paper measurement
on the issues of broadband
measurement issues oftransient electric
broadband field electric
transient
signals and studies electromagnetic pulse measurement technology based on high-speed
signals and studies electromagnetic pulse measurement technology based on high-s
digital samplingdigital
techniques usingtechniques
sampling FPGA. The mainFPGA.
using research
Thecontent includescontent
main research the following
includes the fo
aspects: Section 2 explores the design of a miniaturized transient electric field sensor
ing aspects: Section 2 explores the design of a miniaturized transient electric field se
based on electrically
basedsmall monopole small
on electrically antenna and develops
monopole corresponding
antenna and develops ultra-wideband
corresponding ultra-w
conditioning circuits. Section 3 introduces the local high-speed digital sampling
band conditioning circuits. Section 3 introduces the local high-speed technologydigital sam
based on FPGA, enabling based
technology the digital sampling
on FPGA, of measured
enabling the digitalanalog
samplingsignals. The time-
of measured analog sig
domain calibration of the transient
The time-domain electric of
calibration field
the measurement system
transient electric field is illustrated in
measurement system is
Section 4. Section 5 carries
trated out experimental
in Section 4. Section 5 validation
carries outinexperimental
a power system converter
validation in astation.
power system
Conclusions and discussions
verter station. are presentedand
Conclusions in Section 6. are presented in Section 6.
discussions

2. Electric Field SensorField Sensor

2. Electric
Monopole antennas are widely
Monopole antennasused areinwidely
electromagnetic interference testing,
used in electromagnetic electro-
interference testing, ele
magnetic compatibility assessments, and electromagnetic environment
magnetic compatibility assessments, and electromagnetic environment measurem measurements.
When the electric Whenfieldthecomponent of an
electric field electromagnetic
component wave is present
of an electromagnetic around
wave the an-
is present around th
tenna, it induces a voltage along the axial direction of the antenna, thereby
tenna, it induces a voltage along the axial direction of the antenna, thereby sensin sensing the
electric field in the environment.
electric field in the When the antenna
environment. Whensatisfies the electrically
the antenna satisfiessmall condition small c
the electrically
(where the geometric size isthe
tion (where much smallersize
geometric than the shortest
is much smallerwavelength in thewavelength
than the shortest signal), the in the sig
time for the signal to propagate
the time through
for the signal the antenna
to propagate is much
through shorteristhan
the antenna much the rise time
shorter than the rise
of the leading edge
of theofleading
signal,edge
thusofpreserving
signal, thus the original shape
preserving of theshape
the original pulse.ofBy theusing
pulse. By usin
an electrically small antenna,
electrically smallwhich is largely
antenna, whichindependent of frequency,
is largely independent as a receiving
of frequency, as a receivin
antenna to detect tenna to detect
transient pulsetransient
electric pulse
fields,electric fields, the waveform
the time-domain time-domain waveform
of the incidentof the inc
electric field canelectric field can with
be measured be measured
minimalwith minimalAdditionally,
distortion. distortion. Additionally,
the antenna thehas
antenna ha
the characteristiccharacteristic of omnidirectional
of omnidirectional reception ofreception of electromagnetic
electromagnetic fields in the horiz
fields in the horizontal
plane [22]. plane [22].
The design of an The design
electric of an
field electric fieldprobe
measurement measurement
primarily probe primarily
consists of the consists
monopole of the mono
antenna, the signal conditioning circuit, and a metal shielding case. The monopole antennamonopol
antenna, the signal conditioning circuit, and a metal shielding case. The
tenna converts
converts electromagnetic field electromagnetic
signals in spacefield into signals
electricalin signals.
space into Theelectrical signals. The ou
output signal
signal power of a monopole electrically small antenna
power of a monopole electrically small antenna is limited, making it unsuitable for trans- is limited, making it unsuitab
transmission
mission via coaxial via coaxial
cables. Therefore, cables. Therefore,input
a high-impedance a high-impedance
of an operational inputamplifier
of an operationa
is used to achieve impedance transformation and matching for the antenna. Thefor
plifier is used to achieve impedance transformation and matching the antenna
metal
Sensors 2025, 25, 1159 5 of 18

shielding case encloses the internal circuit board, serving both as a shield and as a mirror
ground plane for the antenna.

2.1. Principle of a Monopole Electrically Small Antenna

An electrically small antenna operates under the condition h ≪ λ/2π, where h
is the antenna height and λ is the operating wavelength. It has advantages such as a
small geometric size, a wide bandwidth, an adjustable and large measurement range,
omnidirectional measurement, and adjustable parameters, meaning that it is commonly
used for measuring transient electric field signals. In this regime, the monopole antenna
primarily functions as a capacitive element, responding to the displacement current induced
by the incident electric field. The induced voltage V across the antenna terminals is
proportional to the effective height he and the local electric field E:

h
V ( t ) = h e · E ( t ), h e ≈ (1)
2
The sensitivity of the rod antenna can be adjusted by varying its height h to accom-
modate different measurement environments. To ensure a wideband frequency response,
the antenna can be modeled as a parallel RC circuit, where Ca represents the antenna
capacitance and Rl is the load impedance of the measurement system. The transfer function
H (s) of the system is given by

V (s) sRl Ca he
H (s) = = (2)
E(s) 1 + sRl Ca

From this, the lower cutoff frequency f L can be expressed as

1
fL = (3)
2π (Cl + Ca ) Rl

To extend the low-frequency response (e.g., down to 50 Hz), a high-impedance buffer

circuit is employed to maximize Rl , ensuring accurate measurements of both steady-state
and transient signals. The higher cutoff frequency f h is limited by the size of the antenna,
which is determined by the condition of h ≪ λ/2π.
Compared to traditional D-Dot sensors, which measure the time derivative of the
electric field and require numerical integration, the monopole antenna directly measures the
electric field strength E(t). This simplifies the signal processing chain while maintaining a
wideband response, making it particularly suitable for capturing sub-nanosecond transients
and low-frequency steady-state components simultaneously.
Practical design considerations include minimizing parasitic capacitance and induc-
tance to extend the high-frequency response and optimizing the antenna geometry to
balance sensitivity and physical size. These principles ensure that the monopole antenna
achieves a flat response across DC to GHz frequencies, enabling precise transient electro-
magnetic measurements in power system applications.

2.2. Simulation of the Electric Field Sensor

To analyze the time-domain response and the amplitude–frequency characteristics in
the frequency domain, a simulation model was constructed in CST Microwave Studio, as
shown in Figure 2. The model features a rod antenna designed as a cylindrical structure
with a height of 10 mm and a radius of 1 mm. Surrounding the rod is a hollow cuboidal
shell with a thickness of 2 mm. A hole with a radius of 2 mm is created on the upper
surface of the hollow shell, and the cylindrical rod is positioned inside. Both the cylindrical
 ground potential of the metal case in the practical use of the sensor, the designed condi-
tioning circuit has an equivalent input impedance consisting of a 5 pF capacitor in parallel
Sensors 2025, 25, x FOR PEER REVIEW
with a 1012 Ω resistor. To account for this, a lumped element is added between the6pole of 19
Sensors 2025, 25, 1159 6 of 18
and the metal case. The sensor features a metallic casing and an enclosed structural design
to shield against external electromagnetic interference, which adopts a compact structural
cylindrical
design rod and
to avoid the hollow
interfering withshell are made
the test results. ofThe
copper, whilesize
sensor’s 4 cm 4 cm
theisremaining 3 cm
simulation
rod and the hollow shell are made of copper, while the remaining simulation region is set
region is width
(length set to vacuum.
height).
to vacuum.
Considering that the ground potential of the conditioning circuit is equal to the
ground potential
Hollow of the metal
Monopole Electrically case Input
Equivalent in the practical use of the sensor, the designed condi-
Impedance
Cuboid Shell Small Rod Antenna of the Conditioning Circuit
tioning circuit has an equivalent input impedance consisting of a 5 pF capacitor in parallel
2mm
with a 1012 Ω resistor. To account for this, a lumped element is added between the pole
10mm
and the metal case. The sensor features 4mma metallic casing and an enclosed structural design

to shield against external electromagnetic C

interference, which adopts a compact structural
L

design to avoid interfering with the test R

results. The sensor’s size is 4 cm 4 cm 3 cm
L

(length width height).

Finiteelement
Figure2.2.Finite
Figure elementsimulation
simulationmodel
modelof
oftransient
transientE-field
E-fieldsensor.
sensor.
Hollow Monopole Electrically Equivalent Input Impedance
Cuboid Shell Small Rod Antenna of the Conditioning Circuit
IfConsidering
a Gaussianthat thewaveform
pulse ground potential
is used of as the
theconditioning
excitation, the circuit is equal to the ground
amplitude–frequency re-
2mm
potential
sponse of the
curve metal
of the case in the
rod-shaped practical
monopole use of is
antenna the sensor,
shown in the designed
Figure 3, with conditioning
a 3 dB band-
10mm
circuitranging
width has an equivalent
from DC toinput
1.966 impedance
GHz.4mm consisting of a 5 pF capacitor in parallel with
a 1012 Ω resistor. To account for this, a lumped element is added between the pole and
C L
the metal case. The sensor features a metallic casing and an enclosed structural design to
R L
shield against external electromagnetic interference, which adopts a compact structural
design
Figure 2.toFinite
avoid interfering
element with model
simulation the test of results.
transientThe sensor.size is 4 cm × 4 cm × 3 cm
sensor’s
E-field
(length × width × height).
If
If a Gaussian
Gaussian pulse
pulsewaveform
waveformisisused usedasasthe the excitation,
excitation, thethe amplitude–frequency
amplitude–frequency re-
response
sponse curvecurveof of
thethe rod-shaped
rod-shaped monopole
monopole antenna
antenna is shown
is shown in Figure
in Figure 3, awith
3, with 3 dBaband-
3 dB
bandwidth
width ranging ranging
fromfrom
DC to DC to 1.966
1.966 GHz.GHz.

Figure 3. Transient E-field sensor amplitude–frequency characteristic curve.

2.3. Design of Conditioning Circuit

To maximize the electric field energy transferring to the monopole antenna, the im-
pedance of antenna was matched to the impedance of the measurement system. The
BUF802 produced by Texas Instruments was utilized to convert the high impedance to 50
ohm. The BUF802 has two operating modes, Buffer (BF) Mode and Composite Loop (CL)
Mode, each suitable for different situations.
Figure
FigureIn 3.
3.
CL Transient
Transient
mode, the E-field
E-field sensor
sensor
BUF802 amplitude–frequency
amplitude–frequency
features a main path andcharacteristic
characteristiccurve.
curve.
an auxiliary path. The input signal
is2.3.
split by a capacitor divider
Design of Conditioning Circuit into high-frequency and low-frequency components. The
2.3. Design of Conditioning Circuit
high-frequency signal passes through the main path, while the low-frequency signal
To maximize the electric field energy transferring to the monopole antenna, the
passesTo maximize
through the electric
the auxiliary fieldThese
path. energytwotransferring
components toare
thethen
monopole antenna,
recombined theout-
at the im-
impedance of antenna was matched to the impedance of the measurement system. The
pedance
put to form of the
antenna was matched
final output to theare
signal. There impedance
transitionof the segments
band measurement system.
at both The
low- and
BUF802 produced by Texas Instruments was utilized to convert the high impedance to
BUF802 produced
high-frequency by Texas
ranges, whereInstruments
the main path wasand
utilized to convert
auxiliary the high
path work impedance
together to 50
to produce
50 ohm. The BUF802 has two operating modes, Buffer (BF) Mode and Composite Loop
ohm.
the The BUF802
output signal. has two operating
Wideband modes, requires
measurement Buffer (BF) Mode
that the and Composite Loop (CL)
amplitude–frequency re-
(CL) Mode, each suitable for different situations.
Mode, of
sponse eachthesuitable
sensor be forconsistent
different situations.
in the low- and high-frequency ranges. If there is a dif-
In CL mode, the BUF802 features a main path and an auxiliary path. The input signal
ference In in
CLgain,
mode, thethe BUF802 features
measurement a main
results will path
showand an auxiliary
amplitude path. The
distortion. Theinput signal
schematic
is split by a capacitor divider into high-frequency and low-frequency components. The
is split by
diagram of athecapacitor divider
conditioning into design
circuit high-frequency
is shownand low-frequency
in Figure components.
4. A frequency The
sweep test
high-frequency signal passes through the main path, while the low-frequency signal passes
high-frequency
was conducted onsignal passes through
the conditioning circuitthe main
using path, while
a network the low-frequency
analyzer, measured by the signal
S21
through the auxiliary path. These two components are then recombined at the output to
passes through the auxiliary path. These two components are then recombined at the out-
form the final output signal. There are transition band segments at both low- and high-
put to form the final output signal. There are transition band segments at both low- and
frequency ranges, where the main path and auxiliary path work together to produce the
high-frequency ranges, where the main path and auxiliary path work together to produce
output signal. Wideband measurement requires that the amplitude–frequency response of
the output signal. Wideband measurement requires that the amplitude–frequency re-
the sensor be consistent in the low- and high-frequency ranges. If there is a difference in
sponse of the sensor be consistent in the low- and high-frequency ranges. If there is a dif-
gain, the measurement results will show amplitude distortion. The schematic diagram of
ference in gain, the measurement results will show amplitude distortion. The schematic
the conditioning circuit design is shown in Figure 4. A frequency sweep test was conducted
diagram of the conditioning circuit design is shown in Figure 4. A frequency sweep test
on the conditioning circuit using a network analyzer, measured by the S21 parameter as
was conducted on the conditioning circuit using a network analyzer, measured by the S21
Sensors 2025, 25, x FOR PEER REVIEW 7 of 19

Sensors 2025, 25, 1159 parameter as shown in Figure 5. According to the frequency response curve, the measure- 7 of 18
ment bandwidth of the conditioning circuit is DC to 1.65 GHz.
parameter as shown in Figure 5. According to the frequency response curve, the measure-
mentshown
bandwidth of the
in Figure conditioning
5. According circuit
to the is DC to
frequency 1.65 GHz.
response curve, the measurement bandwidth
of theAntenna High circuit
conditioning Frequencyis DC to 1.65 GHz.
BUF802
1uF IN
Bias Out
Antenna33pF High Frequency 10MΩBUF802
Aux 1pF 50
1uF IN
80kΩ Bias Out
33pF 10MΩ
+
Aux 1pF 50
2kΩ
20kΩ Out
80kΩ −56pF
100pF
+ 2kΩ
20kΩ Out
−
20kΩ56pF 80kΩ 100pF

Low Frequency
20kΩ 80kΩ
Low Frequency

Figure 4. Schematic diagram of the conditioning circuit.

Figure 4. Schematic diagram of the conditioning circuit.

5. Frequency
Figure 5. Frequencyresponse
responsetest
testofof
the conditioning
the circuit
conditioning by an
circuit byR&S ZVA24
an R&S vector
ZVA24 network
vector analyzer.
network an-
alyzer.
3. High-Speed Digital Acquisition System Based on FPGA
Figure 5. Frequency response test of the conditioning circuit by an R&S ZVA24 vector network an-
3.
alyzer. In this paper, Digital
High-Speed we propose digitizing the
Acquisition analog Based
System signals output
on FPGA by the sensors at the
measurement front end and then transmitting the digital signals via optical fibers. Unlike
In this paper, we propose digitizing the analog signals output by the sensors at the
3. High-Speed
measurementDigitalfront endAcquisition
traditional oscilloscopes, the high-speedSystem Based
digitalon
analog-to-digital
and then transmitting the FPGA
conversion
signals
module in this system
via optical fibers. Unlike
must also handle tasks such as analog-to-digital conversion, optical fiber communication,
In this paper,
traditional we proposethe
oscilloscopes, digitizing the analog
high-speed signals output
analog-to-digital by the module
conversion sensors inat the
this sys-
large data storage, and remote parameter adjustment.
measurement
tem mustfront also end
handleandtasks
then such
transmitting the digital signals
as analog-to-digital via optical
conversion, opticalfibers.
fiberUnlike
communi-
The system components of the high-speed digital acquisition system are shown in
traditional
cation,oscilloscopes, the high-speed
large data storage, and remote analog-to-digital conversion module in this sys-
parameter adjustment.
Figure 6. The main components include the FPGA control unit, the ADC high-speed
tem must The also system
handle components
tasks such asofanalog-to-digital
the high-speed conversion, optical system
digital acquisition fiber communi-
are shown in
digital sampling unit, the data buffering unit, timing generation and control, and the data
cation, large6.data
Figure Thestorage, and remoteinclude
main components parameter
the adjustment.
FPGA control unit, the ADC high-speed dig-
communication interface.
The system components
ital sampling of the
unit, the data high-speed
buffering unit, digital
timing acquisition
generation system are shown
and control, and thein data
Figure 6. Logic
The main
communication
3.1. components
Control of FPGA include the FPGA control unit, the ADC high-speed dig-
interface.
ital sampling unit, the data buffering unit, timing generation and control, and the data
The FPGA chip is primarily responsible for configuring and controlling the clock
communication interface.
chip and ADC to ensure their proper operation. The main control chip used in the broad-
band transient electromagnetic pulse signal acquisition system is the Xilinx Virtex-7 series
XC7VX690T-2FFG1761I. This chip features 600,000 logic cells, 36 GTH pairs, and supports
x8 Gen 2 PCIe interfaces. It also has 850 user I/O interfaces, making it adaptable to various
application interfaces. Additionally, the FPGA includes 3100 DSP slices and supports the
MIG core, enabling access to DDR3 memory, providing high data processing and buffering
performance. Its main functions include the following:
1. The FPGA acquires high-speed sampled digital signals from the ADC (analog-to-
digital converter) and processes these data in real-time, including operations such as
data buffering, filtering, and windowing.
2. It implements the control and management of complex data streams, separating and
buffering data for subsequent storage or transmission.
Sensors 2025, 25, 1159 8 of 18

3. The FPGA stores the data received from the ADC into high-speed memory (such as
DDR3) and transfers the data to other modules or external devices as needed—for
example, transmitting to a host computer via optical fibers.
4. The FPGA is responsible for transmitting the acquired digital signals to external
systems through the Quad Small Form-factor Pluggable (QSFP) interface, achieving
interface control and communication protocols. The external I/O level is 3.3 V, while
the FPGA’s I/O level is 1.8 V. The SN74LVDSI chip is used to achieve compatible level
conversion between them.
5. Precise timing control is crucial in high-speed data acquisition systems. The FPGA
Sensors 2025, 25, x FOR PEER REVIEW 8
generates and manages the clock signals for the entire analog-to-digital conversion
circuit, ensuring synchronization between the ADC and other components.

SN74LVDS1
IO
Temperature
Detection

ADC SPI
RX_IN
AD9208 Application Interface
RX_IN
Optical Fiber
RX_IN ADC SPI
AD9208 Optoelectronic
10 Gigabit Ethernet
RX_IN Conversion Module
SFP+
TL-FC31F
FPGA
XC7VX690T-
CLK_REF Clock Management 2FFG1761I
LMK04832
EXT_REF

Signal +15V
Power Management Power Supply
Interface DDR3_4GB
MT41K512M16x4

FLASH
N25Q256 Level Conversion
SN74LVC8T245

FT232

Figure 6. Functional block diagram of a high-speed digital acquisition system.

Figure 6. Functional block diagram of a high-speed digital acquisition system.
3.2. High-Speed Analog-to-Digital Conversion
3.1. Logic Control of FPGA
The analog-to-digital conversion function is primarily accomplished using high-speed
ADC chips, takingThe intoFPGA chipfactors
account is primarily
such responsible
as bandwidth,for configuring and controlling
sampling frequency, timingthe clock
requirements,and ADC
power to ensure their
consumption, andproper operation. protocols.
communication The main control chip usedADC
Two high-speed in the broadb
transient electromagnetic pulse signal acquisition system is
chips, AD9208 from Analog Devices (ADI), were selected for this purpose. The AD9208 is the Xilinx Virtex-7 s
XC7VX690T-2FFG1761I.
a dual-channel, 14-bit ADC with a maximum This chipsampling
features 600,000 logicMSPS.
rate of 3000 cells, 36 GTH
This pairs,isand supp
device
optimized for a wide input bandwidth, high sampling rates, good linearity, and low power to var
x8 Gen 2 PCIe interfaces. It also has 850 user I/O interfaces, making it adaptable
consumption application interfaces.
in a small package. Additionally,
It includes the
on-chip FPGAand
buffers includes 3100 DSP slices
a sample-and-hold and support
circuit,
designed specifically for low power, small size, and ease of use. The product is used for and bu
MIG core, enabling access to DDR3 memory, providing high data processing
ing performance.
sampling wideband Its main
analog signals. Thisfunctions include
dual-channel the following:
ADC core employs a multi-stage,
differential pipeline
1. The architecture
FPGA acquireswith integrated
high-speed output errordigital
sampled correction logic.
signals fromEach
theADC
ADC (analo
has a wideband buffered
digital input, supporting
converter) variousthese
and processes selectable
data input ranges.including
in real-time, The integrated
operations su
reference voltage source simplifies
data buffering, design.and
filtering, Thewindowing.
data outputs of each ADC are internally
connected to 2.an optional 1/2 decimation
It implements the controlmodule.
and management of complex data streams, separating
Both the analog input and
buffering dataclock signals are differential
for subsequent storage or input signals. The data outputs
transmission.
of each ADC 3. are internally
The FPGAconnected
stores thetodata
tworeceived
digital downconverters
from the ADC into (DDCs). Each DDC
high-speed memory (suc
consists of four cascaded
DDR3)signal processing
and transfers thestages: a numerically
data to controlled
other modules oscillator
or external (NCO)
devices as needed—
and three half-band decimation
example, filters that
transmitting to support 2×, 4×, and
a host computer 8× decimation.
via optical fibers.
4. The FPGA is responsible for transmitting the acquired digital signals to external
tems through the Quad Small Form-factor Pluggable (QSFP) interface, achievin
terface control and communication protocols. The external I/O level is 3.3 V, w
the FPGA’s I/O level is 1.8 V. The SN74LVDSI chip is used to achieve compa
level conversion between them.
5. Precise timing control is crucial in high-speed data acquisition systems. The F
generates and manages the clock signals for the entire analog-to-digital conver
Sensors 2025, 25, 1159 9 of 18

3.3. Data Cache

Since the application involves capturing nanosecond pulse train signals at a high
sampling rate while providing millisecond-level storage duration, the system requires a
large storage capacity. To address concerns about potential memory bandwidth limita-
tions with two 12-bit ADCs operating at 3 GS/s each, our system employs four Micron
MT41K256M16 DDR3 memory chips, each with a 16-bit data width and 4 Gb capacity.
Configured in parallel, they form a 64-bit wide DDR3 memory interface totaling 2 gigabytes
(GB), directly connected to the FPGA for high-speed data acquisition and storage. This
design ensures that our system can store and process data at the full rate generated by the
ADCs, effectively mitigating concerns related to memory bandwidth limitations. FLASH
memory is used to store the nonvolatile FGPA’s configuration.

3.4. Timing Clock Generator

The sampling clock signal frequency of our designed digital acquisition system is fixed
at 3 GSps. We selected clock chips LMK04832 from Texas Instruments (TI) to provide the
reference clock for the entire sampling system. The LMK04832 can provide 14 output clock
signals, with a frequency range up to 3255 MHz, which provides two 3 GHz clock signals to
the two AD9208 chips. Additionally, by using frequency division and working in conjunc-
tion with the FPGA, the clock signals are distributed to the data storage, communication
interface, and other modules.

3.5. Data Communication

The FPGA transmits the acquired digital data at high speed through QSFP, with the
digital signals transmitted via optical fibers. Each channel typically achieves a transmission
rate of 10 Gbps or even higher. The optical signals reach the remote host computer and
are converted back into electrical signals, which are then transmitted through the network
interface to the host computer for display and control. Additionally, the host computer can
send control commands via optical fibers to configure the parameters and trigger, reset,
and control the acquisition system.

4. Calibration for the Electric Field Sensor

4.1. Calibration for Conversion Factor Curve
Typically, the electric field E in space is measured using an electric field sensor, which
outputs a voltage signal V. It is necessary to determine the proportionality factor K = V/E
and evaluate the linearity of K under different electric field intensities. This paper conducts
sensor calibration based on a TEM cell which is also known as a transverse electromagnetic
wave chamber [27,28]. It is constructed with the outer conductor’s middle section, formed
by joining the top and bottom plates, while the inner conductor’s middle section is a
strip-shaped partition. The ends of the partition gradually taper and connect to the inner
conductor of a coaxial cable with a characteristic impedance of 50 Ω. This configuration
forms a main transmission section in the middle and horn-shaped transition sections at
both ends. A square voltage signal is injected into one end, whose rising time is about
800 ps, corresponding to an upper frequency limit of f H = 0.35 tr = 437 MHz. Then, a
uniformly propagating transverse electromagnetic wave is generated within the TEM cell.
Using the TEM cell structure in Figure 7 as an example, the input port is connected to
the pulse source via a coaxial cable, and the output port is connected to an oscilloscope.
Once the conversion factor K is determined, the numerical coefficient between the output
voltage and the input electric field can be achieved. The fitted curve of conversion factor is
shown in Figure 8, which shows good linearity. The slope of the fitted line represents the
conversion factor K, which is approximately 0.87 mV/(V/m).
 Once the conversion factor K is determined, the numerical coefficient between the o
put voltage and the input electric field can be achieved. The fitted curve of convers
Sensors 2025, 25, 1159 factor is shown in Figure 8, which shows good linearity. The slope of the
10 fitted
of 18 line rep
sents the conversion factor K , which is approximately 0.87 mV/(V/m).

(a) Schematic diagram of the calibration system based on a TEM cell

Sensors 2025, 25, x FOR PEER REVIEW (b) Physical photograph 11 of 19

Figureof7.transient
Figure 7. Principle Principleelectric
of transient electriccalibration
field sensor field sensor calibration
based on a TEMbased
cell.on a TEM cell.

Fitted Curve of Conversion Factor

10,000
Data Point
9,000
Fit Line
8,000

7,000
Induced Voltage(mV)

6,000

5,000

4,000

3,000

2,000

1,000

0
0 2,000 4,000 6,000 8,000 10,000
Electric Field Intensity(V/m)

Figure 8.
Figure Fitted curve
8. Fitted curve of
of the
the conversion
conversion factor.
factor.

A D-dot
A D-dot antenna
antenna (SGE3-5G
(SGE3-5G from
from Montena
Montena Inc.)
Inc.) and
and the
the proposed
proposed monopole
monopole antenna
antenna
are selected for calibration in the TEM cell, as shown in Figure 9. In the time-domain
are selected for calibration in the TEM cell, as shown in Figure 9. In the time-domain wave-
waveform,
form, the antenna
the D-dot D-dot antenna
exhibitsexhibits local oscillations
local oscillations in the high-level
in the high-level flat regionflat region
of the of
square
the square wave, indicating poor low-frequency response. This observation
wave, indicating poor low-frequency response. This observation is further confirmed in is further
confirmed in the frequency-domain analysis, where the spectrum of D-dot shows obvious
the frequency-domain analysis, where the spectrum of D-dot shows obvious differences
differences compared with the reference square wave. In comparison, the proposed sensor
compared with the reference square wave. In comparison, the proposed sensor demon-
demonstrates good consistency with the calibrated square wave in both the time-domain
strates good consistency with the calibrated square wave in both the time-domain wave-
waveform and the frequency spectrum, highlighting its superior performance.
form and the frequency spectrum, highlighting its superior performance.

1.2
Calibrate square wave
1 Montena probe
Proposed Sensor
0.8

0.6

0.4
 wave, indicating poor low-frequency response. This observation is further confirmed in
the frequency-domain analysis, where the spectrum of D-dot shows obvious differences
compared with the reference square wave. In comparison, the proposed sensor demon-
Sensors 2025, 25, 1159 strates good consistency with the calibrated square wave in both the time-domain
11 of wave-
18
form and the frequency spectrum, highlighting its superior performance.

1.2
Calibrate square wave
1 Montena probe
Proposed Sensor
0.8

0.6

0.4

0.2

0
-100 0 100 200 300 400 500 600 700 800
Time(ns)
(a) Time domain waveform.
Amplitude

(b) Spectrum in frequency domain.

Sensors 2025, 25, x FOR PEER REVIEWFigure
Figure 9. Measured
9. Measured waveform
waveform ofproposed
of the the proposed sensor
sensor and D-dot
and D-dot antenna
antenna Cell. Cell.12 of 19
in TEM
in TEM

4.2. 4.2.
Calibration for High-Frequency
Calibration Response
for High-Frequency Response
However,the
However, thecalibration
calibrationbybythethe
TEM TEMcellcell is insufficient
is insufficient to evaluate
to evaluate the system’s
the system’s up-
upper frequency limit of 1.6 GHz. Therefore, a monocone antenna
per frequency limit of 1.6 GHz. Therefore, a monocone antenna is employed to test the is employed to test the
system’s time-domain response capabilities at its upper frequency limit.
system’s time-domain response capabilities at its upper frequency limit. The time-domain The time-domain
calibrationof
calibration ofthe
thehigh-frequency
high-frequency bandwidth
bandwidth response
responsecharacteristics
characteristicsofofthe thesensor
sensorbased
based on
the monocone antenna was implemented as described in detail in [29,30].
on the monocone antenna was implemented as described in detail in [29,30]. The config- The configuration
of the calibration
uration experiment
of the calibration is shown
experiment in Figure
is shown 10. The
in Figure 10.reference waveform
The reference waveform used used
in the
experiment is a bipolar pulse with a rise time of Tr = 160 ps and a positive-negative
in the experiment is a bipolar pulse with a rise time of Tr = 160 ps and a positive-negative pulse
delaydelay
pulse of 500ofps,
500which is shown
ps, which in Figure
is shown 11. In 11.
in Figure comparison, the rise
In comparison, time
the riseoftime
the electric
of the
field signal measured by the sensor is 210 ps, which is slightly longer
electric field signal measured by the sensor is 210 ps, which is slightly longer than the than the rise time
riseof
the reference
time pulse. The
of the reference corresponding
pulse. The correspondingupper cutoff
upperfrequency is approximately
cutoff frequency 1.66 GHz,
is approximately
consistent with the frequency sweep curve of the sensor shown in
1.66 GHz, consistent with the frequency sweep curve of the sensor shown in Figure 5. Figure 5.

Laptop

Digital
Monocone antenna
acquisition
E-sensor
device

Coaxial Optical
RF cable fiber

Ground plane

Voltage pulse generator Oscilloscope

Figure10.
Figure Experimentalconfiguration
10.Experimental configurationofoftime
timedomain
domaincalibration
calibrationbased
basedon
onsingle
singlecone
coneantenna.
antenna.
 Voltage pulse generator Oscilloscope
Sensors 2025, 25, 1159 12 of 18
Figure 10. Experimental configuration of time domain calibration based on single cone antenna.

Figure 11.
Figure Calibration measured
11. Calibration measured waveform
waveform in
in time
time domain.
domain.

5. Site
5. Site Testing
Testing in
in HVDC
HVDC Station
Station
Based on the principles described above, the electric field sensor is shown in Figure 12a,
Based on the principles described above, the electric field sensor is shown in Figure
Sensors 2025, 25, x FOR PEER REVIEW
the outline size of the high-speed sampling module is 156 mm × 166 mm × 72 mm, and 13 of 19
the
12a, the outline size of the high-speed sampling module is 156 mm 166 mm 72 mm,
front-end sensor can be flexibly replaced. The final physical picture is shown in Figure 12b.
and the front-end sensor can be flexibly replaced. The final physical picture is shown in
Figure 12b.

(a) Electric field sensor (b) High-speed sampling module

Figure 12.Calibration
Figure12. Calibrationmeasured
measuredwaveform
waveformin
intime
timedomain.
domain.

To
Toverify
verifythe
theperformance
performanceofofthe
theelectric
electricfield
fieldmeasurement
measurementsystem,
system,on-site
on-sitetests
testswere
were
conducted
conducted at the ±
at the 400kV
±400 kVQaidam
QaidamConverter
ConverterStation
StationininQinghai
Qinghai Province,
Province, China.
China. TheThe
ra-
radiated electric field was measured under two conditions: during the operation of
diated electric field was measured under two conditions: during the operation of the GIS the GIS
disconnect
disconnectswitch
switchandandduring
duringstable
stableoperation
operationnearnearthe
theAC ACbusbar.
busbar.
5.1. Disconnect Switch Operation Test
5.1. Disconnect Switch Operation Test
During the power outage maintenance process at the converter station, it is necessary
During the power outage maintenance process at the converter station, it is necessary
to segment and disconnect two 350 kV busbars, involving the operation of multiple GIS
to segment and disconnect two 350 kV busbars, involving the operation of multiple GIS
circuit breakers and disconnect switches. Research has shown that the GIS circuit breaker
circuit breakers and disconnect switches. Research has shown that the GIS circuit breaker
contains an SF6 arc extinguishing chamber, which has a certain ability to suppress over-
contains an SF6 arc extinguishing chamber, which has a certain ability to suppress over-
voltages caused by breaking unloaded lines. Therefore, the probability of overvoltage
voltages caused by breaking unloaded lines. Therefore, the probability of overvoltage
events caused by arc reignition during circuit breaker operation is considered lower than
events caused by arc reignition during circuit breaker operation is considered lower than
that during disconnect switch operation. When the disconnect switch is opened, due
that during disconnect switch operation. When the disconnect switch is opened, due to its
to its poor arc extinguishing performance, the opening operation can cause multiple arc
poor arc extinguishing performance, the opening operation can cause multiple arc reigni-
reignitions in the contact gap, forming very fast transient overvoltages (VFTOs) that are
tions in the contact gap, forming very fast transient overvoltages (VFTOs) that are coupled
coupled to the GIS enclosure and grounded, generating a transient ground potential rise.
to the GIS enclosure and grounded, generating a transient ground potential rise. Simulta-
Simultaneously, transient grounding currents arise, and transient electromagnetic fields
neously, transient grounding currents arise, and transient electromagnetic fields are cou-
are coupled into the surrounding space, as detected by our sensor. The transient electric
pled into the surrounding space, as detected by our sensor. The transient electric field
field pulse signals generated by GIS switch operations can have extremely fast rise times,
pulse signals generated by GIS switch operations can have extremely fast rise times, reach-
reaching the nanosecond or picosecond level, covering a wide frequency spectrum from
ing the nanosecond or picosecond level, covering a wide frequency spectrum from a few
a few kilohertz (kHz) to several gigahertz (GHz). This test scenario is one of the typical
kilohertz (kHz) to several gigahertz (GHz). This test scenario is one of the typical applica-
tional scenarios in electrical equipment fault diagnosis, and it can be used to evaluate the
performance of the proposed sensor as well. The electric field sensor is placed about 1 m
from the disconnect switch, at a height of approximately 0.5 m above the ground. The
voltage and current sensor are connected to the digital acquisition system at the same time.
 Sensors 2025, 25, 1159 13 of 18

applicational scenarios in electrical equipment fault diagnosis, and it can be used to evaluate
the performance of the proposed sensor as well. The electric field sensor is placed about
1 m from the disconnect switch, at a height of approximately 0.5 m above the ground. The
voltage and current sensor are connected to the digital acquisition system at the same time.
All sensors are connected to a high-speed digital acquisition system inside a shielded box
via a high-frequency shielded coaxial cable. So, it is possible to measure the electric field,
Sensors 2025, 25, x FOR PEER REVIEW voltage, and current simultaneously. The experimental setup is shown in Figure 13, 14 whose
of 19

parameters are shown in Table 1.

Disconnector
switc h

Electric field Phase Phase Phase

Sensor C B A

Voltage Cu rrent
Sensor Sen sor

Figure
Figure13. On-site
13.On-site equipment
equipment configurationfor
configuration forthe
thedisconnect
disconnectswitch
switchopening
openingtest.
test.

From
Table Figure
1. The 14, itofcan
indicators be observed
measuring that the time-domain waveforms of the electric
equipment.
field, voltage, and current all exhibit similar pulses with identical occurrence times. The
Sensor Bandwidth Measurement Range
three pulse times are 0 µs, 17 µs, and 1190 µs, which reflect the reignition moments of the
arc. FigureVoltage Sensor
15 shows 220 MHz waveform and corresponding
the local detailed time-domain 27 kV spec-
Current Sensor 16 MHz 30 kA
trum of the pulse at the moment of the opening operation. The electric field appears as a
pulse train with a damped oscillatory wave shape, reaching a maximum amplitude of 5.65
From
kV, with Figure
a pulse rise14, it can
time bens
of 12 observed that time
and a decay the time-domain waveforms
of approximately of dominant
1 µs. The the electric
spectral components of the electric field signal are 15.6 MHz, being the same times.
field, voltage, and current all exhibit similar pulses with identical occurrence The
with the
three pulse
voltage times are
and current, whose0 µs,upper
17 µs,frequency
and 1190components
µs, which reflect the80
are over reignition
MHz. moments of
the arc. Figure 15 shows the local detailed time-domain waveform and corresponding
spectrum of the pulse at the moment of the opening operation. The electric field appears
as a pulse train with a damped oscillatory wave shape, reaching a maximum amplitude
of 5.65 kV, with a pulse rise time of 12 ns and a decay time of approximately 1 µs. The
dominant spectral components of the electric field signal are 15.6 MHz, being the same
with the voltage and current, whose upper frequency components are over 80 MHz.
This test confirms that the sensor possesses sufficient response speed and bandwidth to
effectively measure the nanosecond(a)electromagnetic transient such as electric field, voltage
and current signals simultaneously.

(b)
 trum of the pulse at the moment of the opening operation. The electric field appears as a
pulse train with a damped oscillatory wave shape, reaching a maximum amplitude of 5.65
kV, with a pulse rise time of 12 ns and a decay time of approximately 1 µs. The dominant
Sensors 2025, 25, 1159 14 of 18
spectral components of the electric field signal are 15.6 MHz, being the same with the
voltage and current, whose upper frequency components are over 80 MHz.

(a)

(b)

Sensors 2025, 25, x FOR PEER REVIEW 15

(c)

Figure 14. Measured

Figureswitch field pulse
14. Measured train
switch signal
field intrain
pulse timesignal
domain: (a) electric
in time domain: field; (b) voltage;
(a) electric field; (b) volta
(c) current. current.

(a) (b)

(c) (d)

(e) (f)

Figure
Figure 15. Detail of the15.pulse
Detailsignal
of thegenerated
pulse signal
by generated by switch
switch operation: (a)operation: (a) electric
electric field field in the tim
in the time
domain; (b) electric
main;field
(b) in the frequency
electric domain;
field in the (c) voltage
frequency domain;in the
(c) time domain;
voltage in the(d) voltage
time in the
domain; (d) voltage
frequency domain; (e) current in the time domain; (f) current in the frequency domain.
frequency domain; (e) current in the time domain; (f) current in the frequency domain.

This test confirms that the sensor possesses sufficient response speed and bandw
to effectively measure the nanosecond electromagnetic transient such as electric field
age and current signals simultaneously.
Sensors 2025, 25, x FOR PEER REVIEW 16 of 19
Sensors 2025, 25, 1159 15 of 18

5.2. Power Frequency Steady-State Electric Field Test

To test the low-frequency performance of our designed monopole antenna, we mea-
350 kV
suredACthebus steady-state electric field beneath an AC overhead busbar. To compare the
response of different antennas at the 50 Hz power frequency, the measurement included
Isulatin g
Sensors 2025, 25, x FOR PEER REVIEW
a D-dot
bu sh ingantenna from Montena Mon Inc.,tena 16 of 19
a traditional monopole sensor, and our designed
D-dot
ultra-wideband monopole sensor. sensor The layout ofProposed
the experiment is shown in Figure 16.
Mon opole
Sen sor

350 kV
AC bus

Measuring Traditional
Isulatin g
Mon opole
bu sh ingsensor Mon tena
Sen sor
D-dot
sensor Proposed
Mon opole
Figure 16. On-site equipment configuration for the steady-state
Sen sor
test.

The measured electric field waveform is shown in Figure 17, and the spectrum is
shown in Figure 18. The blue curve represents the measurements performed by the tradi-
tional monopole sensor, which exhibits significant waveform distortion and only reflects
Measuring Traditional
a repeating period of approximately 20 ms. The red and black curves represent the wave-
Mon opole
sensor
Sen sor
forms measured by our designed ultra-wideband monopole sensor, which more accu-
rately reflect the waveformconfiguration
of the power frequency voltage with a repeating frequency of
Figure 16. On-site
Figure 16. On-site equipment
equipment configuration for for the
the steady-state
steady-state test.
test.
50 Hz. This is because the conditioning circuit of the electric field sensor adopts a split-
bandThematching
The measured
measured design for high
electric
electric fieldand
field low frequencies,
waveform
waveform is shown resulting
is shown in Figurein
in Figure 17,a and
17, flatter
and thefrequency
the spectrumre-
spectrum is
is
sponse
shown curve
in in
Figure the18.low-frequency
The blue region.
curve represents the measurements
shown in Figure 18. The blue curve represents the measurements performed by the tradi- performed by the
tionalNotably,
traditional
monopole there
monopole aresensor,
sensor, spikes
whichof opposite
which
exhibits polarity
exhibits
significant superimposed
significant
waveform waveform at specific
distortion andphases
distortion onlyand ofonly
the
reflects
reflects
power a repeating
frequency
a repeating period
periodwaveform,
of ofwhich
approximately
approximately may 20 red
be caused
20 ms. The ms.by The
and red and
high-frequency
black black
curves curves
pulses
represent fromrepresent
thecorona
wave-
the
forms measured by our designed ultra-wideband monopole sensor, which more more
waveforms
discharges on measured
the busbar. by
Fromour designed
the spectrum,ultra-wideband
apart from monopole
the sensor,
fundamental which
frequency of
accu-
accurately
50
rately theyreflect
Hz, reflect allthe the waveform
contain
waveform of the
someofhigh-frequency
the powerpower frequency
components,
frequency voltage
voltage such with
with asa350a repeating frequency
Hz andfrequency
repeating 650 Hz. of
of 50 Hz.
In This
comparison, is because
the the
D-dot conditioning
antenna from circuit
Montena of the
Inc electric
does not
50 Hz. This is because the conditioning circuit of the electric field sensor adopts a split- field sensor
measure an adopts
effective a
split-band
electric field
band matching matching
waveform,
design design
for for high
because
high the and
low low
andcoupled frequencies,
electric
frequencies, field resulting
from
resulting the ainflatter
in power a flatter frequency
frequency
frequency volt-
re-
response
age exceeds curve
the in the
lower low-frequency
frequency
sponse curve in the low-frequency region. limitregion.
of the sensor.
Notably, there are spikes of opposite polarity superimposed at specific phases of the
power frequency waveform, which may be caused by high-frequency pulses from corona
discharges on the busbar. From the spectrum, apart from the fundamental frequency of
50 Hz, they all contain some high-frequency components, such as 350 Hz and 650 Hz.
In comparison, the D-dot antenna from Montena Inc does not measure an effective
electric field waveform, because the coupled electric field from the power frequency volt-
age exceeds the lower frequency limit of the sensor.

Figure 17. The waveform of the power frequency induced electric

electric field
field signal.
signal.

Notably, there are spikes of opposite polarity superimposed at specific phases of the
power frequency waveform, which may be caused by high-frequency pulses from corona
discharges on the busbar. From the spectrum, apart from the fundamental frequency of
50 Hz, they all contain some high-frequency components, such as 350 Hz and 650 Hz.
In comparison, the D-dot antenna from Montena Inc does not measure an effective
electric field waveform, because the coupled electric field from the power frequency voltage
Figure 17. The waveform of the power frequency induced electric field signal.
exceeds the lower frequency limit of the sensor.
Sensors 2025, 25, 1159 16 of 18
Figure 17. The waveform of the power frequency induced electric field signal.

Figure 18. The spectrum of the power-frequency-induced electric field signal.

6. Conclusions
To address the current shortcomings of transient electric field sensors, such as poor
low-frequency performance, narrow bandwidth, insufficient accuracy, large size, unstable
amplitude coefficients, and low system integration, we propose a new technical approach
based on FPGA high-speed digital sampling and optical fiber communication. To design
the sensor with electromagnetic interference resistance, we employed front-end high-
speed analog-to-digital technology to sample the analog signal and used optical fibers to
transmit the digital signal. This approach maximizes the integrity of signal while achieving
electromagnetic interference resistance, and it effectively improves the sensitivity and
coefficient stability of the sensor.
Based on the working principle of the rod-shaped electrically small antenna, we
studied and optimized the design of the conditioning circuit for the induced voltage of the
electric dipole antenna, extending the sensor’s analog bandwidth from the original 1 kHz
to 460 MHz to an ultra-wide bandwidth from DC to over 1.6 GHz.
On-site validation tests were conducted at the ±400 kV Qaidam Converter Station in
Qinghai Province, where we measured the transient electric fields generated by discon-
nect switch operations as well as the steady-state power frequency electric fields under
normal operating conditions. A comparative analysis of the performance of different
antenna sensors showed that our designed sensor outperforms the D-dot and traditional
monopole antenna sensors in low-frequency scenarios. The proposed sensor successfully
captured the coupled electric field waveform at the power frequency (50 Hz), with its
waveform matching the period of the power frequency voltage, in accordance with the
quasi-electrostatic field theory. In comparison, the D-dot antenna failed to detect this low-
frequency signal due to noise interference, while the waveform of the traditional sensor
exhibited significant distortion.
This study demonstrates that the electric field measurement capabilities of the sensor
cover a range from power frequency steady-state signals to VFTO and high-frequency
transient signals such as partial discharges. This measurement system shows great potential
in the fields of condition monitoring, fault detection, and diagnostics, as well as in partial
discharge fault localization in power equipment.

Author Contributions: Conceptualization: S.H.; Methodology: S.H. and X.C.; Software: X.C., B.Z.
and L.S.; Validation: X.C. and L.S.; Formal analysis: S.H. and X.C.; Investigation: S.H. and X.C.;
Data curation: X.C. and L.S.; Writing—original draft preparation: S.H. and X.C.; Writing—review
and editing: S.H., X.C., B.Z. and L.S.; Visualization: S.H. and X.C.; Supervision: S.H.; Project
administration: S.H.; Funding acquisition: S.H. All authors have read and agreed to the published
version of the manuscript.

Funding: This research was funded by the key research and development project of Shaanxi Province
(China), grant number 2023-YBGY-064.
Sensors 2025, 25, 1159 17 of 18

Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.

Data Availability Statement: Data are unavailable due to privacy restrictions.

Conflicts of Interest: The authors declare no conflicts of interest.

Abbreviations
The following abbreviations are used in this manuscript:
FPGA Field-programmable gate array
EMP Electromagnetic pulse
HEMP High-altitude electromagnetic pulse
IGBT Insulated gate bipolar transistor
ADC Analog-to-digital converter
TEM Transverse electromagnetic
GIS Gas insulated switchgear
VFTO Very fast transient overvoltage

References
1. Paul, C.R.; Scully, R.C.; Steffka, M.A. Introduction to Electromagnetic Compatibility; John Wiley & Sons: Hoboken, NJ, USA, 2022.
2. Giri, D.V.; Sabath, F.; Hoad, R. High-Power Electromagnetic Effects on Electronic Systems; Artech House: Norwood, MA, USA, 2020.
3. Nitsch, D.; Camp, M.; Sabath, F.; ter Haseborg, J.L.; Garbe, H. Susceptibility of some electronic equipment to HPEM threats. IEEE
Trans. Electromagn. Compat. 2004, 46, 380–389. [CrossRef]
4. Radasky, W.A.; Baum, C.E.; Wik, M.W. Introduction to the special issue on high-power electromagnetics (HPEM) and intentional
electromagnetic interference (IEMI). IEEE Trans. Electromagn. Compat. 2004, 46, 314–321. [CrossRef]
5. Kim, K.; Iliadis, A.A. Impact of microwave interference on dynamic operation and power dissipation of CMOS inverters. IEEE
Trans. Electromagn. Compat. 2007, 49, 329–338. [CrossRef]
6. Kasmi, C.; Esteves, J.L. IEMI threats for information security: Remote command injection on modern smartphones. IEEE Trans.
Electromagn. Compat. 2015, 57, 1752–1755. [CrossRef]
7. Baum, C.E. An equivalent-charge method for defining geometries of dipole antennas. Sens. Simul. Notes 1969, 72, 90167-8.
8. IEC 61000-4-20:2010; Electromagnetic Compatibility. Part 4-20: Testing and Measurement Techniques—Emission and Immunity
Testing in Transverse Electromagnetic (TEM) Waveguides. IEC Standard: Geneva, Switzerland, 2010.
9. IEC 61000-4-23:2016; Electromagnetic Compatibility. Part 4–23: Testing and Measurement Techniques—Test Methods for
Protective Devices for HEMP and Other Radiated Disturbances. IEC Standard: Geneva, Switzerland, 2001.
10. IEC 61000-4-24:2015; Electromagnetic Compatibility. Part 4–24: Testing and Measurement Techniques—Test Methods for
Protective Devices for HEMP Conducted Disturbance. IEC Standard: Geneva, Switzerland, 2015.
11. IEC 61000-4-25:2001; Electromagnetic Compatibility. Part 4-25: Testing and Measurement Techniques—HEMP Immunity Test
Methods for Equipment and Systems. IEC Standard: Geneva, Switzerland, 2001.
12. Balzovskii, E.V.; Buyanov, Y.I.; Koshelev, V.I. An active antenna for measuring pulsed electric fields. Russ. Phys. J. 2007, 50,
503–508. [CrossRef]
13. Baum, C.E. From the electromagnetic pulse to high-power electromagnetics. Proc. IEEE 1992, 80, 789–817. [CrossRef]
14. Peng, Z.; Liu, Y.; Wu, Q.; Liu, L.; Chen, J. An ultra-wideband time-domain antenna. In Proceedings of the 2022 IEEE MTT-S
International Microwave Workshop Series on Advanced Materials and Processes for RF and THz Applications (IMWS-AMP),
Guangzhou, China, 27–29 November 2022; pp. 1–3.
15. Kong, X.; Xie, Y.Z.; Li, Q.; He, S.Y.; Jin, Y.B. Development of one-dimensional norm detector for nanosecond-level transient electric
field measurement. IEEE Trans. Electromagn. Compat. 2017, 59, 1035–1040. [CrossRef]
16. Wu, H.T.; Jiao, C.Q.; Cui, X.; Liu, X.F.; Ji, J.F. Transient electromagnetic disturbance induced on the ports of intelligent component
of electronic instrument transformer due to switching operations in 500 kV GIS substations. IEEE Access 2017, 5, 5104–5112.
[CrossRef]
17. Wang, L.; Huang, J.; Ji, J.; Pang, F.; Yuan, Y. Measurement of transient electromagnetic coupling and interference caused by
disconnector operation in substation. Measurement 2017, 96, 1–7.
18. Wiggins, C.M.; Thomas, D.E.; Nickel, F.S.; Salas, T.M.; Wright, S.E. Transient electromagnetic interference in substations. IEEE
Trans. Power Deliv. 1994, 9, 1869–1884. [CrossRef]
Sensors 2025, 25, 1159 18 of 18

19. Tatematsu, A.; Moriguchi, S.; Ueda, T. Switching surge analysis of an EHV air-insulated substation using the 3-D FDTD method.
IEEE Trans. Power Deliv. 2018, 33, 2324–2334. [CrossRef]
20. Zhou, B.; Shi, L.; Wang, J.; Chen, H.; Chen, B. Electromagnetic Pulse and Its Engineering Protection; National Defense Industry Press:
Beijing, China, 2019; pp. 1–7.
21. He, S.; Zhang, B.; Ni, X.; Qiu, P.; Song, Y.; Song, L. Design of Nanosecond-level Transient Electric Field Sensor and its Application
in HVDC Converter Station. IEEE Lett. Electromagn. Compat. Pract. Appl. 2023, 5, 57–61. [CrossRef]
22. Kong, X.; Xing, W.Q.; Xie, Y.Z.; Zheng, Z.; Sun, Z.D. A broadband optical fiber transmission-based time domain measurement
system for nanosecond-level transient electric field. Rev. Sci. Instrum. 2022, 93, 011501. [CrossRef] [PubMed]
23. Mohamed, F.P.; Siew, W.H.; Soraghan, J.J.; Strachan, S.M.; McWilliam, J. Partial discharge location in power cables using a double
ended method based on time triggering with GPS. IEEE Trans. Dielectr. Electr. Insul. 2013, 20, 2212–2221. [CrossRef]
24. Russer, P. EMC measurements in the time-domain. In Proceedings of the 2011 URSI General Assembly and Scientific Symposium,
Istanbul, Turkey, 13–20 August 2011; pp. 1–35.
25. Braun, S.; Donauer, T.; Russer, P. A real-time time-domain EMI measurement system for full-compliance measurements according
to CISPR 16-1-1. IEEE Trans. Electromagn. Compat. 2008, 50, 259–267. [CrossRef]
26. Cheng, X.; Tian, X.; Zeng, M.; Han, X. The design of 1 GSPS real-time sampling system for transient pulsed signal. IEEE Trans.
Nucl. Sci. 2010, 57, 539–542. [CrossRef]
27. IEC 61000-4-33:2005; Electromagnetic Compatibility. Part 4-33: Testing and Measurement Techniques-Measurement Methods for
High-Power Transient Parameters. IEC Standard: Geneva, Switzerland, 2005.
28. Harrington, Z.C.T. IEEE Standard Calibration of Electromagnetic Field Sensors and Probes (Excluding Antennas) from 9 kHz to
40 GHz. In IEEE Electromagnetic Compatibility Society; IEEE: Piscataway, NJ, USA, 2013.
29. Zhang, H.; Lin, H.; He, Z.; Lu, R.; Xing, H.; Wang, B.; Wang, Z. Experimental Design for Verifying the Time-Frequency Domain
Performance of Specular Single Cone Electric Field Standard Devices. Metrol. Sci. Technol. 2023, 67, 43–47.
30. Johnk, R.T.; Ondrejka, A. Time-Domain Calibrations of D-Dot Sensors; NIST: Gaithersburg, MD, USA, 1998.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.