E-Band VNA For Automotive Market: Bumper and Emblem Measurements
E-Band VNA For Automotive Market: Bumper and Emblem Measurements
E-Band VNA For Automotive Market: Bumper and Emblem Measurements
Purpose
This application note, developed in conjunction with vendors in the automotive market, will present a
configuration used to test and measure the transmission and reflection characteristics of materials (emblems
and bumpers) placed in front of automotive radars. The configuration explained below will make it possible to
provide clear and accurate measurements for companies working in the automobile market. These will be
precise and repeatable measurements given in dBs for transmission and reflection characteristics.
Introduction
A radar placed behind an emblem or bumper of a car must transmit its signal with minimal interference.
Specifications of the materials are required by the car manufacturers to ensure the proper operation of the
radars. To test the emblem or the bumper, a ShockLine MS46522B vector network analyzer (VNA) with option 83
(5 meter E-band) is used to provide the necessary frequency range coverage, measure all S-parameters (S11, S12,
S21, and S22), and characterize the behavior of the material. In this case, instead of using a conducted setup, where
the signals are transmitted through a transmission line like a coaxial cable, two antennas are attached to both
ports of the ShockLine MS46522B-083 VNA to transmit and receive the signals. The model of antenna used must
be defined according to the test configuration, always considering a plane wave must be present in the surface
to be measured (working in the E-band [60 to 90 GHz], this should not be an issue). Specific antennas with lens
or dielectric will be chosen to form the wave.
For this test, the shape of the material under test (MUT) is something that must also be considered as it can
affect the results depending on the width, paint, manufacturing process, etc. Another question that arises is
whether it is better to use a single, large beam on the whole surface to be tested or a smaller beam and to do it
in two or three tests.
It is important to note that during these tests the radiation environment was not ideal to characterize a MUT
and that some care had to be taken. The most important issue is reflections. In the open environment used,
reflections could happen and ruin the measurements. To avoid reflections there are two options that could be
used that are not exclusive of each other. The first requires placing an absorbent material all around the room,
however, this can complicate the setup even more. The second is the use of the gating process in the ShockLine
VNA. In this case, a gate is defined around the test distance range to exclude any external interference.
The test results were quite precise and consistent, getting measurement accuracies in the order of a few
hundredths of a dB. When conducting these measurements, the ShockLine VNA did not seem to be the most
critical element affecting the accuracy of the measurement, rather it was the actual position of the tested sample for
the reflection measurement. This measurement was influenced by the angle between the direction of the
transmission and the plane of the sample. A small longitudinal displacement of the sample has little affect on
the transmission measurement and for the reflection it is on the order of ± 0.2 dB for ± 1 mm.
The measurements on the bumpers gave values around 0.7 dB for the transmission and –9 dB for the
reflection. These values, of course, being affected by the thickness and paint used.
Although traditional E-band measurement systems, which are much larger and more expensive, can yield
sufficiently accurate results, we must also consider how those systems would fair in a production environment.
The lower cost and smaller “plug-and-play” form factor of the ShockLine MS46522B model is perfectly suited
for this type of production use.
Today, radars used by automotive manufacturers are critical elements in their safety systems. In the future,
these systems will be considered even more critical for the autonomous car. At the output of the production
line, all bumpers and emblems must be measured to ensure that they comply with safety standards and the
results will be saved to maintain traceability. One must be able to leverage a quick and easy-to-use VNA
system that can be placed in a production plant environment.
Basic Setup
The following basic configuration was created to conduct the material measurement tests. The space between
the antennas had to be: 1) wide enough to allow a bumper to be placed in between them, and 2) able to reach
the far field conditions for the wave to be a plane at the sample. Precision mechanics must also be used to
maintain the MUT alignment between the two antennas. In this configuration, two dielectric antennas are
used but this could be changed depending on the requirements. To avoid multiple reflections, absorbent
material was used to improve the configuration. Some samples and bumpers were measured and the results
correlated with previous measurements.
VNA Setup
Complex configurations are not necessary for radar characterization measurements. The information needed by
bumper manufacturers is usually transmission and reflection parameters in the form of S-parameters in
logarithmic scale (plus phase). Thus, 4 traces are configured showing S11, S22, S12, and S21 in log scale.
• F
requency range: the configured frequency range will be decided by the customer, depending on their
particular requirements (typically between 75-80 GHz)
• Number of points: 201
• Power: 0 dBm
• IFBW: 10 kHz
Calibration Setup
As with most uses of a VNA, this application requires the instrument to be configured and calibrated to get
accurate results. The calibration method used for these radiated measurements is the Line, Reflect, and Match
(LRM) algorithm, in which a reflection plane is used for reflective calibration, a line for transmission calibration,
and a perfect match.
The calibration kit consists of a simple metallic plate with a known thickness so it can be considered in the
calibration process. The plate should be as conductive as possible, and with a flat response in the entire
frequency range of interest (a reflection response has been seen in which some anomalies appeared in the
higher frequency band).
2
Figure 1. LRM setup used with metal plate and antennas
For the LRM algorithm, the Reflect for both ports will be done with a metal plate placed perpendicular to both
antennas, while the Line will be done without the plate. Finally, for the Match, the metal plate placed at a 45°
incline will give no reflection back to any of the ports and achieve the same behavior of a perfect match. In this
case, the use of an absorbent material placed to both sides of the metal plate can improve calibration (Figure
1).
In our case, the distance between both antennas is 310 mm, a distance large enough to have some room to
place the samples or bumper parts and to be in the far field conditions of the antennas. The metal plate, with a
width of 2 mm, is placed in the middle.
Also, in the LRM setup, the reflection is configured to be short-like (B), and the reference plane to be situated in
the middle of line 1. Also, a Match must be configured as Cal Device Y (B), which could also be a second line if
we use LRL calibration algorithm.
3
Figure 2. LRM calibration setup
Once the system is calibrated, the measurement of the “Line” (no metal plate or plastic sample in the middle
of the setup) would give the flat trace of both magnitude and phase. Figure 3 shows an example the results
achieved after a good calibration procedure.
4
It is expected to have a magnitude ripple of less than ±0.25 dB and a phase ripple of less than ±0.2°, as
shown in Figure 4. The calibration process, however, could be affected by different parameters, like position
of the metal plate, external reflections, etc. Special care should be taken when performing the calibration
process to obtain best results.
Test Configuration
The bumper (or material sample) will be set at the reference plane, which is the middle point between the
antennas (in the case of measuring bumper parts, mechanical positioners play an important role to place the
bumper in the precise position to be measured).
The distance between the antennas is dependent on the test requirements. Far field conditions need to be
met, which is easy in this frequency range (supposing far field is ten times the wavelength, these conditions
are met when the antennas are set apart by 5 cm or more). It is also important that the radiation spot on the
sample not exceed or even be similar to the sample size to avoid dispersion. Figure 5 illustrates the situation
if a dielectric antenna and horn antenna are used (radiation pattern for both can be found at the end of this
document):
5
Using dielectric antennas with a distance of 31 cm between them, the radiation spot (diameter) is less
than 3 cm for a 3 dB attenuation (less than 5 cm for a 10 dB att). Similar calculations with the horn
antennas shows a spot of 57 mm in diameter. This means that having a sample of 15x15 cm provides
enough margin to be sure the spot will fall entirely in the sample.
In the case in which absorbent material is not available, a gate can be configured to improve the
measurement and avoid external reflections (these external reflections can appear in the trace as
additional ripple). Thus far the material position has been established as the reference plane, which is the
zero position. The gating should now be configured around the zero position. This can be seen in Figure 6
that shows the time domain response (in bandpass mode) of the S21.
Then, what needs to be discovered is the width of the window to use in the setup to try to improve the
results. This will depend on the frequency span being used in the measurement (the higher the frequency
span, the shorter the window size needed). Table 1 shows the relationship between the configured
frequency span and the window size (these values state the window size for reflection measurements; for
transmission measurements the window size must be doubled):
Nominal 2 1/BW
2 2/BW
⎛ 0.5 ⎞ π + β 2 2
⎜ ⎟ ⎛ ⎛ 10.5⎞ ⎞ ππ 2++ββ 2
Low side lobe ⎝ BW ⎠ 1.5/BW
π ⎜ ⎜ ⎟ ⎟3/BW
⎝ ⎝BW
BW⎠ ⎠ ππ
Minimum side lobe 2/BW 4/BW
2 2
⎛ 1 ⎞ π +β 2
⎜ ⎛ 0⎟.5 ⎞ π 2 + β 2 ⎛ 0.5 ⎛ ⎞1 ⎛ π ⎞π 2 + β 2 −1 α / 20
Kaiser (parameter β) ⎝ BW
⎜ ⎠ ⎟ π ⎜ ⎜ (
⎟ ⎜ ⎞⎟ ⎟ + cosh 10 ( )) 2
⎝ BW ⎠ π ⎝ π ⋅ BW⎝ BW
⎠ ⎝ ⎠2 ⎠ π
2 2
⎛ 0.5 ⎞ ⎛ π ⎞ ⎛ 1 ⎞ ⎛π ⎞ 2
Chebyshev (parameter α dB) ⎜ (
⎟⎛ ⎜1 ⎟⎞ +π 2cosh
−1
+ β 2 10 (
α / 20
)) 2
−1
⎜ ⎛ 0.5⎟ ⎞⎜ ⎛ ⎟π ⎞+ cosh 10( (
α / 20 2
))
⎝ π ⋅ BW ⎠⎜ ⎝ 2 ⎠⎟
⎝ BW ⎠ π ⎝ π⎜⋅ BW ⎠ ⎟⎝ 2⎜ ⎠ ⎟ + cosh 10
−1
( (
α / 20
)) 2
⎝ π ⋅ BW ⎠ ⎝ 2 ⎠
Table 1. Gate size as a function
2 of the frequency span
⎛ 1 ⎞ ⎛π ⎞
⎜ (
⎟ ⎜ ⎟ + cosh 2
−1
(
10α / 20
2
)) ⎛ 1 ⎞ ⎛π ⎞
2
⎝ π ⋅ ⎛BW0⎠.5 ⎝ 2⎞ ⎠⎛ π ⎞ + cosh −1 10α / 20
⎜ ⎟ ⎜ ⎟ ( ( ))
2
⎜
−1
⎟ ⎜ ⎟ + cosh 10 (
α / 20
( )) 2
6
⎝ π ⋅ BW ⎠ ⎝ 2 ⎠ ⎝ π ⋅ BW ⎠ ⎝ 2 ⎠
As an example, using a nominal window with a frequency span of 3 GHz we should use a 2/3 ns gate for
transmission, what is equivalent to, if we consider a propagation velocity of Vp= 3.108 m/s, a 20 cm gate for
reflection measurements, and, consequently, to a 40 cm gate for transmission measurements. In the trace
shown in Figure 7 (using a nominal window with a frequency span of 3 GHz, a 20 cm gate for reflection
measurements, and a 40 cm gate for transmission measurements), we can see the gate around the time
domain response in transmission.
We can also see how the gate affects the time domain (TD) response. Of course, window type can be changed
and the size modified, which will slightly vary the response. Then we can move to displaying the “Frequency
Response with Time Gate” to see the result after applying the gating to our measurements. In the following
trace we can see the transmission and reflection response of the thru with and without gate (Figure 8). We
can clearly see the fading of the ripple in both traces, but also some effect on the trace itself. Not a big
difference in transmission trace, which can be degraded by 0.02 dB, but a significant variation in reflection.
Still, reflection values are still very good numbers (almost less than –40 dB), and it could be worth trying with
different window sizes and types to enhance the results.
7
Figure 8. Transmission and reflection response with and without gate
Another point to be considered with these results is the fact that the setup of the system was not fully
optimized (precision mechanical positioners for the MUT or extender heads and antennas were not used). It
is important to reiterate that in this setup, the precise placement of the MUT, extender heads, and antennas
will play a big role in the results to be obtained. Later in this document, some pictures of the setup in some
customers will be shown.
Test Results
For traceability, storage of the results is important, and, for this purpose, the data and images can be
saved. Care must be taken when saving the data if the format is in .s2p because, by default, the gating will
not be taken into account. The results can be saved in a .csv or .txt format, or the ShockLine MS46522B
system can be setup to save the .s2p file with the applied gating in the system menu as shown in Figure 9
(System → Setup → Misc. Setup → SnP Files Setup). In the SnP Setup popup window, select Save Gated
Data to keep the gating in the .s2p file.
8
Some Examples
Recently, great interest has developed within the European automotive industry around characterizing
materials used in bumper manufacturing that could affect radar transmissions. The big automobile
manufacturers are pushing their contractors to measure the material characteristics in the E-band
frequency range on all related components used in radars. Different materials, paint coatings, plastic
injection methods, and material widths must be tested many times in a production environment to
accurately characterize their behavior. To accomplish this, simplicity and ease-of-use of the VNA system is
usually what customers require. The ShockLine MS46522B E-band system and this particular setup for
material characterization appears to be the best solution to fit those needs. A VNA that is almost plug-and-
play, a calibration method that barely takes a couple of minutes to be done, and a measurement method
that is repeatable and easy to perform constitutes the best setup for making these measurements.
Of course, and in order to obtain consistent results, precise positioning of the setup is required. This is
usually not a problem at all for these type of customers, but it is something to be considered when
approaching this setup.
One customer was a contractor working on bumpers for a major European automobile manufacturer. In
this case, they were able to manufacture a structure to hold both the VNA extender heads with the
antennas and the sample to be measured. They also built the metal plate for calibration (Figure 10).
In the picture we can see the structure to hold the parts of the system and all of them are movable so we can
adjust the length between the antennas, the position of the sample, etc. One important point to mention is
the holder of the extender head. They were 3D printed in plastic, which turned out to be a good solution as
in a second demo for a different customer this holder was manufactured in metal and was the cause of
several reflections.
9
Figure 11. Transmission and reflection behavior of a plastic sample
In Figure 11 we can see the trace with the reflection and transmission characteristic and behavioir of one of
their plastic samples.
The possibility of adjusting the vertical position of the sample gives us the flexibility to play with different
sample sizes.
A second customer visited was a contractor working in this case for PSA group. Here the mechanical setup
was a bit more complex, as they wanted to characterize not just samples but entire bumper parts (however
once again, they took care of manufacturing everything to setup the system). The results were very similar to
what we measured with the other customer, with the exception of some ripple obtained in transmission due
to the extender head holders manufactured in metal in this case. One lesson learned is that it is better to
have the parts in plastic if possible, to avoid possible reflections affecting our measurements.
Summary
The ShockLine MS46522B E-band system enables a very simple and intuitive setup to conduct material
characterization. The most appreciated feature is the simplicity of the setup and configuration, as well as the
repeatability of the measurements compared to other systems.
10
The antennas to be used can vary depending on the requirements of the customer, as we have previously
shown. One should always consider how big the spot is that we want to illuminate in the sample to select
the antenna we want to use. Following we can see different antennas that can be used in this setup:
• A
nritsu dielectric antenna: this high-gain dielectric antenna offers very high performance and
narrow beam width. It comes in WR-12 interface to be attached to the WG interface in the extender
head.
• A
nritsu horn antenna: this antenna has a lower gain and, thus, wider beam width, but still performs
properly for these setups.
The interface is WR-12 and the size of the antenna is reduced, doing this appropriate for this kind of
setups. For more information please refer to the Antenna and Antenna Kits Technical Data Sheet
located on anritsu.com.
• A
nteral: another lens antenna with typically more than 30 dB gain and a narrow beam (typically less
than 5°) width for small radiation spots. To learn more visit Anteral's website.
11
Figure 14. Anteral lens antenna performance data
CAR
Bumper
50 mm
Car Radar
7°
Note
This document was initially created by J.P. Guillemet in co-working with J. Martens, JB Soeiro, and G. Denis, as
a result coming from some tests with cars and bumpers manufacturers. This is an update coming from the
last tests done, mainly with the bumper manufacturers with additional tests performed and feedback
documented and received.
12
Note
13
Specifications are subject to change without notice.
®Anritsu All trademarks are registered trademarks of 11410-01181, Rev. B Printed in United States 2020-03
their respective companies. Data subject to change ©2020 Anritsu Company. All Rights Reserved.
without notice. For the most recent specifications
visit: www.anritsu.com