ELEC 453/6391 Microwave Engineering

Experiment #4
Network Analyzer Measurements
This experiment consists of four parts. All parts will be related to impedance
measurements. The student should be familiar with the Anritsu MS46122 Vector Network
Analyzer (VNA), which measures s-parameters based on incident and reflection. In the first part,
the student will be utilizing the VNA to measure the load impedance of the same loads measured
used in Exp. 1 and 2. In the second part, the student will measure the impedance of the four
waveguide irises measured in Exp. 3. For the third part, the student will use a slide screw tuner to
obtain a perfect match with iris D. In the final part, the student will measure an unknown load on
a microstrip line and download the data to design matching networks, which will be connected
and measured during the next experiment.

Objectives:
The main objectives of this lab are:
- Use the VNA (Vector Network Analyzer) to characterize unknown loads,
- Demonstrate the ability to de-embed the measurement reference plane arbitrarily,
- Measure impedances of mismatches using a VNA,
- Use the sliding screw tuner to match a waveguide load, and
- Design matching networks for unknown loads.

Equipment
- Anritsu MS46122 VNA,
- HP X870A slide screw tuner,
- X-band waveguide matched load,
- Waveguide to SMA transition,
- Shorting plate,
- Irises A, B, C, D used in Experiment 3,
- Microstrip line with Z 0 = 50 Ω on GML 1000 board terminated with a short, and
- Microstrip line with Z 0 = 50 Ω on GML 1000 board terminated with an unknown load.

Note: It is important to watch the videos available on Moodle about this experiment before
coming to the Lab.
Introduction
The following is an overview of a vector network analyzer.
The Vector Network Analyzer
A vector network analyzer, such as the Anritsu MS46122, has two input channels for each port.
This gives the same function of the vector voltmeter of Experiment 2 for each port in conjunction
with a frequency sweeper, one input is the “reference” channel and the other is the “test”
jφ
channel. If the voltage at the reference connector or “port” is Vref = Vref e ref , and the voltage at
the test port is Vtest = Vtest e jφtest , then the network analyzer measures the ratio of the voltage at the
Vtest Vtest j (φtest −φref )
“test” port to the voltage at the “reference” port, = e . So the network analyzer
Vref Vref
Vtest
measures the ratio of the test-port voltage to the reference-port voltage, A = , and is
Vref
sometimes called a “ratiometer”. The NA measures the difference in phase between the test-port
voltage and the measurement-port voltage, φ = φtest − φ ref . Much less expensive “scalar” network
analyzers measure the amplitude ratio only.

The RF unit supplies the signal to the “source” or “RF in” port of the
reflection/transmission test set. The test set, Figure 1, uses a directional coupler to provide a
sample of the RF signal from the generator to the “reference” port of the network analyzer. The
RF voltage is routed to the “unknown” port of the test set, and the load reflects some voltage
back. A directional coupler samples the reflected voltage and then routes the sample to the “test”
port of the harmonic frequency converter. The harmonic frequency converter down-converts to
an intermediate frequency (IF) in the kHz range, preserving the relative amplitude and phase of
the test signal relative to the reference signal. The IF frequency is more easily measured and
processed than the RF signal itself. Narrow-band detection at IF brings high sensitivity, wide
dynamic range, and immunity from harmonics and noise. The VNA extracts the amplitude and
phase and displays them as required by the user.

Fig. 1 A diagram for the directional coupler inside the VNA, the “reflection/transmission test
set”, or “S-parameter test set”.

A key element in any network analyzer is the directional coupler1, which separates the
incident signal from the reflected signal. The “directivity” of the directional coupler is a figure-
of-merit which measures how well the two signals are separated. Pozar discusses the

1
Pozar, Sec. 7.9, p.414 describes the use of directional couplers for network analyzers.
measurement of the directivity2 and the effect of coupler directivity on the accuracy of reflection
coefficient measurements3.

VNA Calibration

Fig. 2 An illustration for the measurement unknowns that requires calibration

As described previously, the process of measurement requires several steps that may introduce
magnitude and phase errors in the measurements. In addition, the use of external cables,
connectors and adapters introduce other types or measurement errors. All these errors can be
combined in the signal flow graph diagram shown in Fig. 2. Consequently, a calibration process
has to be performed ahead from the Device Under Test (DUT) measurement to identify the
unknown sources of errors, or in other words the unknown S-parameters in Fig. 2. The main
objective of this calibration step is to improve the measurement accuracy. In this experiment, we
should focus on the SOLT calibration procedure (Short – Open – Load -Thru), where we use
these different measurements with the VNA to automatically calibrate the equipment before
performing the actual measurements.

Part 1:
The goal of this part is to measure the N-Type loads that are previously characterized in
experiments 1 and 2. A comparison is required between the different measurements conducted
throughout this Lab.

Procedure:
- Set the measurement frequency range between 600MHz to 900MHz.
- Perform the VNA calibration for a one port measurement.
- Measure the short circuit load.
- Measure the different N-Type loads.
- Perform the de-embedding manually as illustrated in the related video.
- Report the load impedances with and without de-embedding.

2
Pozar page 357.
3
Pozar page 414.
Part 2:
This part is concerned about measuring the waveguide irises characterized in the previous Exp 3.
A comparison between the measured impedances is required for this part.

Procedure:
- Set the measurement frequency range between 10.3GHz to 10.8GHz.
- Perform the a 1 port VNA calibration.
- Enabling the measurement port de-embedding using the available VNA feature as
presented in the related video.
- Measure the reflection coefficient of the available waveguide irises backed by the
matched load.
- Calculate the impedance for each iris at the chosen frequency from the previous
experiment .

Hint: Compute the waveguide wave impedance to calculate the impedance from the reflection
coefficient.

Part 3:
Introduction
The sliding screw tuner of Fig. 3 is a length of waveguide with a slot cut in the wall, similar to
the slotted line. The tuner is 14 cm long. Thin metal post sticks into the slot, parallel to the
electric field of the dominant mode and can move along the slot. The depth of the penetrating
probe is adjustable with a micrometer. When the micrometer is set to zero, the probe is all the
way in (maximum length). When it is set to 400, it is fully retracted, by a 0.400 inch. A short
probe acts like a shunt capacitance and a very long post becomes a shunt inductance. The sliding
screw tuner behaves the same way as a single-stub tuner. Let Y1 be the input admittance of the
waveguide from the post to the load termination. Choose the distance of the post from the load
to make the real part of Y1 equal to the wave admittance. The post itself behaves as a
susceptance connected across the waveguide, in parallel with Y1 . By adjusting the depth of the
post, we can obtain a range of susceptance values and so “tune out” the imaginary part of Y1 .
Then, the input admittance of the post in parallel with Y1 is a “match” to the wave admittance of
the waveguide.
 Fig. 3 The sliding screw tuner.

In industry, tuning screws are often used to match a load to a waveguide. A tuning screw
behaves like the post in the slide screw tuner. To “design” the tuning screw, we must determine
the position of the tuning screw from the load and the depth of the screw. The position of the
screw is sometimes found with a small cylindrical magnet on the inside of the waveguide, held in
place by a larger magnet from the outside. Slide the magnet along the outside of the waveguide,
which drags the screw inside the waveguide with it. By observing the return loss on the VNA
screen, we can slide the magnet along until the RL shows the best match. Then we drill a hole
for a tuning screw, tap the hole, insert a tuning screw and adjust its depth to get the best possible
match.

The Experiment
The main goal of this part is to use the sliding screw turner to match iris D at the frequency
10.5GHz. You are required to report matching position and probe length.

Procedure:
- Use the previous setup with the same calibration and de-embedding.
- Connect iris D backed by the matched load and store the data.
- Remove the matched load and connect the sliding screw tuner between the iris and the
matched load.
- Allocate a marker on screen at the desired frequency.
- Search for the matching point by changing the position and the probe length as presented
in the related video.
- Save your best matching results to compare to the original case.
Part 4:
This part of the experiment is about designing a matching network for an unknown load mounted
over a PCB.

Procedure:
- Set the frequency range between 2 to 4 GHz.
- Calibrate the VNA for a one port measurement.
- Use the available short circuited microstrip line board to de-embed the measurement
plane to the tip of the board.
- Measure the reflection coefficient of the load at 3 GHz.
- Export the S1P file representing this measurement to use it in ADS.
- Design single and double stub matching networks (to be implemented in the next lab).

Note: Students will not be allowed in the lab for the next experiment without having their
matching network design and prepared.

Preliminary Exercise
Answer these questions before you come to the lab. There will be no quizzes in this lab.
Instead the lab demonstrator will evaluate your pre-lab answers. The lab demonstrator will
check that you have answered these questions before he permits you to do the experiment.

1. Calculate the effective relative permittivity and line impedance of the GML 1000
substrate material for a microstrip line 73 mils in width. You should use the information
available in the provided data sheet at the end of this manual.
2. In the framework of calibrating the network analyzer for a two port measurement, the
VNA calibration should include: (Choose the best answer)
a. Short, Open and Load at both ports
b. Short, Open and Load port 1
c. A thru
d. (a) and (c)
e. (b) and (c)
3. For a measurement case in the scenario of part 1 in this experiment where the de-
embedding is not enabled for the measurement, the measured input reflection coefficient
of an unknown load reads 0.7 with phase 180°. Calculate the actual load impedance given
that the input reflection coefficient for a short circuit reads 0.95 with phase 39°.
Hint: The N-type adapter has a characteristic impedance of 50 Ω.
4. Calculate the characteristic impedance of the WR90 waveguide at frequencies 10.5GHz,
10.7GHz and 11GHz. Comment on its variation with frequency.
Hint: You can use all the available resources to identify the dimensions of WR90
waveguides.
Tables of Data
Student name:
Student I.D.:
Lab Section:

Lab Instructor’s Signature:

Table 1
N-Type Load Measurements at the frequency ___________.
Case File Name Reflection Coefficient
Short Circuit
Matched Load
2:1 Load
1.2:1 Load
“tee 50” load
“tee 75” load
Large resistor
Small resistor

Table 2
File Names for the Iris Measurements
Case File Name
Waveguide Matched Load
Iris A
Iris B
Iris C
Iris D

Table 3
File Name for the Slide Screw Tuner Experiment with Iris D
Distance from the post to the iris=___________cm
Depth of the post=_____________cm
Case File Name
Slide Screw Tuner

Table 4
Microstrip Board Label: ________
File Names for the Microstrip Line Measurements
Case File Name
Calibration with a short circuit
Reflection Coefficient at the Load
Questions to Answer in your Lab Report
Your lab report must include a signed “Expectations of Originality” form.

Your lab report must include the tables from the previous section, filled in with your data, and
“signed off” by your lab demonstrator at the end of the lab session.

Your lab report will include answers to the following questions:

1) Use a rectangular plot or a Smith chart to show the reflection coefficient of the N-type
loads in part 1 (separate plots). Compare with your measured values from Experiments 1
and 2. Do the three methods agree?

2) Plot the admittance of each of the four irises, backed by a matched load, as a function of
frequency on a Smith Chart (4 separate plots). In Lab #3, you measured the admittance
of each iris-plus-matched-load at 10.4, 10.5 and 10.6 GHz. Include these three points on
your Smith Chart plot for each iris. Does the reflectometer measurement from Lab #3
agree with your network analyzer measurement?

3) Plot the reflection coefficient for iris-D-plus-matched-load and for iris D “matched” with
the slide screw tuner on a Smith Chart. What is the “return loss” at 10.5 GHz? What is
the bandwidth over which the return loss is better than 20 dB?

4) Design a single stub matching network for the load measured in part 4. Use ADS to
verify your design considering the discontinuities (you should use the S1P file obtained
from your measurement). You may need to tune the design parameters in ADS to
improve the design.
Plot the measured load impedance in comparison to your ADS results

5) Design a double stub matching network for the load measured in part 4. Use ADS to
verify your design considering the discontinuities. You may need to tune the design
parameters in ADS to improve the design.
Plot the measured load impedance in comparison to your ADS results