ANRITSU 37300 Vector Network Analyzer

ANRITSU 37300 Vector Network Analyzer
37000 Series
Vector Network Analyzer
Application Note
Measuring Frequency Conversion Devices
Introduction
Controllable Sources:
This application note discusses frequency conversion device
measurements with the Anritsu 37100A Vector Network Analyzer. The 37200B and 37300A VNAs will also make these
measurements, in addition to full S-parameter measurements.
Appendix F shows how to apply the material presented on the
37100A to the 372/300 series. The mixer, a typical frequency
conversion device, will be used as the DUT. However, everything covered is also valid for other frequency conversion
devices such as up or down converters, multipliers, and
dividers.
The measurements that will be discussed are the following:
• Amplitude and Phase Tracking
• Group Delay
• Port Match
• Port to Port Isolation
• Conversion Gain/Loss
For absolute conversion loss, magnitude and phase, Appendix
E discusses the NxN mixer measurement method. This powerful technique that yields real time tuning capability, takes
advantage of a special VNA calibrating software.
MULTIPLE SOURCE CONTROL
Vector Network Analyzers are instruments used typically to
make ratio measurements at the same input and output frequency. The Anritsu 37100A Direct-Access Receiver configuration is ideal for measuring frequency conversion devices.
The 37100A uses its Multiple Source Control Mode to stimulate a frequency conversion device at one frequency, and measure its response at another.
Transfer Test
Source
Switch Attenuator Attenuator
Drive
Drive
Drive
Source
Samplers
6 dB
a1
3 dB
Source
Lock
a2
Out
b2
RF
Out
RF
In
Internal
Source
Out
Figure 1. 37100A Direct-Access Receiver Configuration
Multiple Source Control allows the user to separately control
up to two sources and a synchronously tuned receiver. The
sources’ frequency range and power along with the receiver’s
frequency range and reference channel may be specified. With
the Multiple Source Control software, a sweep may consist of
up to five consecutive bands, each with independent source
and receiver settings for convenient testing of frequency conversion devices.
2
Control Formula:
Multiple Source Control is specified as a frequency range
partitioned into 1 to 5 consecutive bands. For each band, the
sources’ and receiver’s frequency sweeps may be independently specified per the following formula. A CW frequency
could also be specified.
Frequency Sweep = X/Y * (F + N (GHz)), where
N is the offset frequency in GHz
F is the displayed frequency
X, Y are integer constants
X, Y, and N may be independently specified for each source
and receiver. They may be positive or negative numbers
depending on the type of device. F is global and is the same
value in all formulas.
Example: Band 1 = 500 MHz to 900 MHz
Source 1 = 2/3 * (CW 1 GHz)
666.7 MHz CW
Source 2 = 2/3 * (F − 0 GHz)
Will sweep 333.3 MHz to
600 MHz
Receiver = −2/3 * (F − 1 GHz)
Will sweep 333.3 MHz
to 66.7 MHz
The x-axis will display the band of
500 to 900 GHz.
(Notice how the receiver is set to decrease in
frequency with respect to the increasing band
1 frequency.) Check menus in Figure 2.
Bands:
Splitter
b1
Source #1: Internal source or external Anritsu 68B, 69A,
or 6700B series synthesizer.
Source #2: Any Anritsu 68B, 69A, or 6700B series
synthesizer.
The specified frequency range may be divided into one to five
bands. Band 1 must start at the beginning of the desired frequency range and end at a user-specified stop frequency or
the end of the desired frequency range. Band 2 must begin at
the next point after band 1 and continue until a user-specified
stop frequency or the end of the desired frequency range, etc.
Band 5 must end at the end of the desired frequency range.
The unique ability of the 37100A’s Multiple Source Control
to divide the frequency range into bands makes it ideal for
testing harmonic conversion devices.
Multiple Source Control comes with easy to follow menus
containing helpful instructions. (Figure 2)
Figure 2. Multiple Source Control Menus
AMPLITUDE AND PHASE TRACKING
Amplitude and phase tracking can be accomplished in one of
two ways:
• Using a separate device as the phase lock reference.
• Using one channel of the DUT as the phase
lock reference.
For mixers or other single channel devices, a separate device
can be dedicated as the phase lock reference. Figure 3 shows
a typical mixer amplitude/phase match configuration capable
of operating from 10 MHz to 40 GHz. With the addition of a
multiplier, millimeter-wave harmonic mixers can also be tested. The measurement system consists of an Anritsu 37169A
VNA with Multiple Source Control capability, the internal
source to provide the RF input and a synthesizer to provide
the LO drive. The LO and RF are each split and applied to the
DUT and reference mixers. 3 dB pads are used at every mixer
port to improve the mixer match. Without these pads the
standing waves on the interconnect cables can significantly
affect the measurement. The IF outputs of the DUT and reference mixer are applied to a test and reference input on the
37100A. Filters at the receiver inputs will eliminate the possibility of errors due to spurs (see Appendix C on spurious
response). The reference mixer is used to provide a signal for
phase-locking the receiver and to provide a stable amplitude
and phase reference.
68/9 SERIES SYNTHESIZER
TO DEDICATED
GPIB PORT
37100A VNA
RF
b2 OUT
a1
IF
FILTERS
GOLDEN
STD
IF
RF
LO
SPLITTER
SPLITTER
LO
IF
RF
DUT
PADS
REF MIXER
Figure 3. Mixer Amplitude and Phase Tracking Configuration
3
To amplitude and phase track the DUT mixers, the first one
is connected. Multiple Source Control is used for setting the
RF, LO, and Receiver frequencies. The first mixer’s response
is stored using the 37100A’s trace memory feature. Successive mixers are then connected and compared to the first one.
Up to three DUT’s can be tested on one setup using four
way splitters and the four input channels of the 37100A,
one channel as a reference and the other three channels as
test channels.
In many cases it is impractical to dedicate a separate device
as a reference. This is true when testing multi-channel
converters. Consider a two channel converter as shown in
Figure 4. In an ideal, perfectly matched measurement system,
the only contributor to gain/phase mismatch is the DUT. In a
real system, however, mismatch can also be caused by the
power splitter and its associated cabling, and by the receiver
and its associated cabling. The source and its associated
cabling do not contribute to the mismatch because they are
common to both channels. In this system it is necessary to
remove the effects of amplitude and phase mismatch in the
test system channels. This can be accomplished in two ways,
described in Appendix D.
68/9 SERIES SYNTHESIZER
TO DEDICATED
GPIB PORT
GROUP DELAY
Group delay is the rate of change of phase through a device
with respect to frequency. Measuring phase requires a reference plane, which cannot be directly established during a frequency translated measurement. Linear applications would
use a through connection as a reference standard. The
37100A uses the approximation of a broadband mixer for the
reference standard. This technique introduces a small measurement error, proportional to the broadband mixer’s very
small delay.
To measure group delay with a 37100A, follow the previous
procedure for measuring amplitude and phase tracking. Use a
broadband mixer as the first mixer, and store its group delay
response in trace memory. Replace this reference setting
broadband mixer with the device under test, and measure its
group delay by looking at Data/Memory. Notice that all errors
due to the phase locking mixer are in common mode, thus
subtracted. The only remaining error is the small group delay
of the broadband mixer used.
PORT REFLECTION COEFFICIENT
Mixer port reflection coefficient can be measured by inserting
couplers between the port to be tested and the source, as
shown in Figure 5. With this configuration, the 37100A acts
as a reflectometer, ratioing b1/a1. The receiver is set to the
same frequency as the RF input. Port reflection coefficient
can then be measured on the 37100A VNA as a normal
reflection measurement.
68/9 SERIES SYNTHESIZER
37100A VNA
RF
b2 OUT
a1
TO DEDICATED
GPIB PORT
IF
FILTERS
37100A VNA
RF1
b1 a1
RF
b2 OUT
SPLITTER
RF2
DUT
IF
RF
CONVERTER
LO
RECEIVER AND CABLES
DUT
COUPLERS
SPLITTER
AND
CABLE
Figure 5. Setup for Measuring Port Match
Figure 4. Measuring a Two Channel Converter
Since the couplers are connected to b1 and a1, the 37100A
can be calibrated using a standard Reflection Only cal, or a
1 path 2 port cal. Notice that the receiver frequency was set
to the same frequency as the RF input, simulating a nonfrequency translated measurement. The calibration should be
performed with the attenuators installed, to compensate for
their effect.
4
PORT TO PORT ISOLATION
RF to IF isolation, or LO to IF isolation can be measured
using the same configuration shown in Figure 5. The receiver
is again set to the same frequency as the RF input. RF to IF
isolation is measured by ratioing b2/a1. If the couplers are
placed in the LO port and the receiver is set to the LO
frequency, then LO to IF isolation can be measured.
The 37100A can be calibrated for RF to IF isolation using
either a frequency response or 1 path, 2 port calibration. Note
that if a 1 path, 2 port calibration is used, the 37100A can
measure port match and port to port isolation simultaneously.
CONVERSION LOSS/GAIN
Measuring Conversion Loss is distinctly different from
the previous measurements in that the VNA is not being
used in its basic mode of making ratioed measurements.
To measure Conversion Loss, the VNA needs to make
absolute power measurements. This is accomplished by
performing a flat test port power VNA calibration, using a
power meter. (Appendix B)
The setup for measuring conversion loss is shown in
Figure 6. First the receiver of the VNA needs to be calibrated
to measure power at the IF frequency of the mixer. This is
accomplished by connecting all the cables and components to
be used in the final measurement, except for the mixer itself.
The resultant measurement (a1/1) is stored in trace memory.
Next, the internal source is calibrated for flat test port output
power at the same power level as the IF, but at the RF frequency. Using Multiple Source Control and inserting the
DUT in the chain, absolute Conversion loss can be measured
by looking at data/memory.
A reference mixer is not necessary in measuring conversion
loss or gain, eliminating the need for two power splitters. The
a1 or a2 channels should be used to make the measurements.
Either channel could be simultaneously used for locking and
measuring purposes.
The 37100A has a phase lock RF output as shown in
Figure 1. It is important to be consistent with the usage of this
output throughout the measurement. If the source lock output
is terminated during the receiver cal step and left open during
the measurement step, differences in reflection will alter the
results, making it meaningless. All aspects other than the
mixer itself have to be in common between the trace in
memory and data measured.
68/9 SERIES SYNTHESIZER
TO DEDICATED
GPIB PORT
POWER METER
TO DEDICATED
GPIB PORT
37100A VNA
SRC
LOCK
a1 OUT
RF
OUT
TERMINATION
IF FILTER
DUT
IF
RF
B
A
LO
SENSOR
DURING THE RECEIVER CAL: CONNECT POINTS A AND B.
DURING THE MEASUREMENT: INSERT THE DUT BETWEEN POINTS A AND B.
DURING POWER FLATNESS CALIBRATION: CONNECT THE POWER SENSOR
TO POINT A, FOR OPTIMUM RESULTS.
Figure 6. Conversion Gain/Loss Setup
Conversion Loss Measurement Steps:
[Channel Menu]
Ë Single Channel
[Enter]
[Graph Type]
Ë Log Magnitude
[Enter]
[S Params]
Ë S21, Fwd Trans
[1]
to redefine parameter,
a parameter other than S21 could be chosen
[Enter]
Ë S21, User 1
toggles to User 1
[Enter]
Ë Change Ratio
[Enter]
Ë numerator, a1
a2 channel could be chosen
[Enter]
Ë denominator, 1
notice display changes to a1/1
[Enter]
Ë Previous Menu
phase lock is already on a1
[Enter]
Ë Previous Menu
5
Calibrate the Receiver:
Measure Conversion Loss:
This is a direct function of the IF frequency; fixed or swept. A
key point to remember is to maintain the number of data
points consistent throughout the measurement. The VNA
makes measurements at distinct frequencies. If the number of
points varies between the stored and current traces, the VNA
will not be able to make comparison calculations.
The start and stop frequencies should be adjusted for the RF
band. The power level should be exactly the same as during
the IF Calibration.
[Data Points]
Ë 401 max Points
[Enter]
for example
• Calibrate for flatness per Appendix B.
The start and stop frequencies should be a narrow band
around the CW IF frequency, or the true band for a Swept IF
frequency mixer.
The power levels should be adjusted for the mixer and setup.
For best results, the test port where the flat output power is
calibrated should be chosen to be at the output of the RF side
pad (check Figure 6). For example, if the mixer conversion
loss is chosen to be measured at 2 dBm, and a 3 dB pad is
used at the RF input, the VNA Power Control and Port 1 Attn
should be set for 5 dBm source output power. But in the Flatness Calibration menu, the Power Target should be adjusted
to 2 dBm, compensating for the pad.
• Connect the through path between the RF and a1 channel
including all cables, pads and filters, but without the mixer.
− In case of a Swept IF mixer you are set for storing the
response; proceed to [Trace Memory].
− For a Fixed IF mixer, change the frequency mode to
CW Mode per the instructions below, then proceed.
[Setup Menu]
[Enter]
Ë CW Mode
[IF Freq]
[MHz]
[Data Points]
[401]
[x1]
[Trace Memory]
[Enter]
Ë Store Data to Memory
[Enter]
Ë View Data / Memory
6
• Re-connect the power sensor to exactly the same test port as
during the receiver calibration.
• Calibrate for Flatness per Appendix B.
You are now ready to insert the mixer and measure
conversion loss.
• Put the VNA in Multiple Source Control Mode per Appendix A,
properly setting the LO source power.
• Connect the Mixer following the Figure 7 setup and measure Conversion Loss directly.
Summary
The Anritsu 37100A VNA is a powerful network measurement tool capable of measuring amplitude and phase characteristics on a wide variety of networks. This capability is not
limited to devices with the same input and output frequencies.
With the 37100A’s direct-access receiver configuration and
built-in Multiple Source Control, this capability is extended to
frequency conversion devices as well. Devices such as mixers,
up/down converters, frequency multipliers and frequency
dividers can be measured with the 37100A VNA.
Gain and phase matching between devices, as well as port
match and port to port isolation are readily measured by the
Anritsu 37100A VNA. With the addition of a power meter for
calibration, conversion loss can also be measured with the
Anritsu 37100A VNA.
Included in the appendix is a description of absolute mixer
measurements without reference to a “golden” standard.
Appendix E describes the NxN technique. Notice that a full
capability VNA is used instead of the Direct Access Receiver
configuration. The external test set required is fully customized to each application. Please forward further interests
on the NxN technique or any other VNA measurements
requirement to you local Anritsu Wiltron representative.
APPENDIX A: Multiple Source Control’s Detailed Instructions
Notice the helpful instructions available in the various Multiple Source Control Menus. (Check Figure 2)
• Connect the external source to the dedicated GPIB connector of the VNA, making sure that the source GPIB address matches the VNA
setting. Check by pressing [Option Menu] and selecting 4Source Config. The Source Config menu can also be accessed through the Multiple Source Control and the Receiver Mode menus found by pressing the [Option Menu] button.
[Setup Menu]
Ë Test Signals
Ë Source 2 Pwr
[Pwr Lvl]
[Enter]
[x1]
[Option Menu]
Ë Multiple Source Control
Ë Define Bands
Ë Band
Ë Band Start Freq
Ë Band Stop Freq
Ë Edit System Equations
Ë Source 1
Ë Multiplier
Ë Divisor
Ë Offset Frequency
[Enter]
[1]
[Freq]
[Freq]
[GHz]
[GHz]
[Enter]
[Enter]
[#]
[#]
[x1]
[x1]
[Freq]
[GHz]
• Repeat for Source 2
Ë Receiver
Ë CW
[Enter]
toggles ON
Ë Offset Frequency
Ë Store Band 1
Ë Band
Ë Previous Menu
The default multiplier, divider, and
offset frequency are 1, 1, and 0
respectively. Modify them as required
making the source or receiver
frequency a linear function of the
band frequency. Notice that all inputs
can be negative numbers depending
on the application.
[Freq]
[GHz]
[Enter]
[2 thru 5]
[Enter]
Up to five consecutive bands could
be added, within which the three
equations could be different. Notice
that once a new band is entered,
the Start Frequency is automatically
set to the Stop Frequency of the
previous band.
when done
Ë Store Band
store each band independently
Ë Set Multiple Source Mode
Ë ON
[Enter]
[Enter]
Before setting Multiple Source Mode
ON, you may cycle through all stored
bands, and easily verify the band
frequencies and their equations on
your screen.
Notice that the x-axis of the display during Multiple Source Control reads the raw band frequency, even though the receiver
could be a CW frequency or a function of the band frequency. This is necessary to meet all possible applications of this
extremely flexible tool.
7
APPENDIX B: Power Flatness Calibration
Flat Power Calibration adjusts the source output power at each measurement point across a frequency span to provide a constant
power level at the test port. For easy referral, these instructions along with additional recommendations are displayed in the Calibrate for Flatness menu. (Check Figure 7)
Power Flatness Calibration Steps: (Check Figure 6)
• Preset, zero, and calibrate the power meter.
• Set power meter offset if required.
• Connect the power meter to the dedicated GPIB interface, and the power sensor to the test port.
[Setup Menu]
Ë Start
Ë Stop
Ë Test Signal
Ë Power Control
Ë Port 1 Attn
Ë Calibrate for Flatness
Ë 401 Points
Ë Power Target
Ë Start Flat Power Cal
[Freq]
[Freq]
[GHz]
[GHz]
[Enter]
[Atten./Gain]
[x1]
[#] of 10 dB steps
[Enter]
[1]
[Pwr Lvl]
[Enter]
[x1]
[x1]
The menu selection 401 Points will change depending
on the data points per sweep selected. For a 401
frequency point sweep, at the default of a power point
cal for every 1 frequency point, the outcome is a 401
points power calibration. This would result in the best
performance, at the expense of the longest calibration
time. If interpolation between frequency points is
acceptable, change the # of frequency points interval
for every power point calibration.
Example: 2001 frequency points per sweep, # of
frequency points interval for a power
point cal selected 20, results in 100
(2001/20) power points calibration.
Notice that the designation under “Calibrate for Flatness” changed from (No Cal Exists) to (Cal Exists). Also the “Flatness Correction” flag was set to ON. If desired this could be toggled to OFF. The calibration will remain stored for later use, until power
is cycled. The calibration can be stored more permanently via the [Save/Recall Menu].
Figure 7. Calibrate for Flatness Menu.
8
APPENDIX C: Errors Due to Spurious Signals
The 37100A’s test set is a narrow band tuned receiver. As such, it is resistant to effects of spurious signals. However, there are
times when the measurement can be affected or even destroyed by spurious signals. If a spurious signal is generated which is
the same frequency as the IF frequency, it will affect the measurement.
Spurious signals in a frequency conversion device measurement are primarily caused by three sources:
• Mixer multiple IF’s
• Mixer/37100A sampler mixed products
• Source frequency errors
Mixer Multiple IF’s:
With specific LO and RF combinations, a mixer can actually
generate two IF’s at the same frequency. For example, if a
mixer receives an LO at 15 GHz and an RF at 10 GHz, it will
produce an IF at 5 GHz. However, the second harmonic of
the RF at 20 GHz will also mix with the LO and produce an
IF at 5 GHz. Hence the 37100A would receive two IF’s at the
same frequency.
This is strictly a DUT phenomenon, independent of the
receiver. The only way to avoid this problem is to select
different frequencies. On the 37100A this can readily be
accomplished using the Multiple Source Control’s bands.
The LO and IF can be normally defined until the problem frequency is approached. Then, using the next band definition,
the LO and IF frequencies can be shifted slightly. (Enough to
avoid the problem, but not enough to significantly affect the
desired data.) After the problem frequency is passed, the LO
and IF frequencies can, using a third band definition, be
changed back to their original numbers.
DUT/37100A Sampler Mixed Product:
The 37100A Test Set, above 270 MHz, has a sampler based
front-end with a first LO of approximately 500 MHz. In
normal operation, the sampler produces harmonics of the
500 MHz LO, one of which mixes with the incoming RF
signal to produce the proper IF frequency. In a non-frequency
translation measurement, there is only one RF frequency to
mix with the signal. (The 37100A uses a unique “spur avoidance” algorithm to prevent harmonics of the input from
mixing to the fundamental IF frequency.) In a frequency
conversion measurement, there are spurious signals, such as
the RF and LO leakage, which can mix with harmonics of the
sampler to produce a spurious IF at the fundamental IF
frequency. If this occurs, the measurement accuracy will be
compromised. A measurement point affected by a spurious
IF should be easily recognized, because that point will be
significantly different than surrounding points.
To avoid the potential for sampler mixed products, the spurious signals from the DUT must be removed. This can be
accomplished with filters at the receiver input. If, for some
reason, filters cannot be used, the ability of the 37100A’s
Multiple Source Control to select bands can be used to avoid
the spur. As previously described, when the bad point is
reached, the LO can be shifted slightly (enough to shift the
spur out of the IF range, but not enough to significantly affect
the measurement). After the bad point, the LO can be shifted
back to its proper frequency. The 37100A test set has a
unique architecture which directly measures signals below
270 MHz, bypassing the sampler.
For many measurements it is feasible to keep the IF below
270 MHz. If the IF is kept below 270 MHz the sampler is not
in the measurement path and potential problems from sampler
spurs will not occur.
Source Frequency Errors:
The 37100A’s internal source is phase locked to provide
synthesizer accuracy and stability during measurements. The
source is phase-locked through the IF. If conditions are such
that the source is not at the correct frequency, but a spurious
IF is within the IF bandwidth, the 37100A may phase lock on
the spurious IF. This results in the source being at the wrong
frequency. For example, the desired configuration is: LO at
9.99 GHz, RF at 10 GHz, and IF at 10 MHz. If the source is
initially 5 MHz off, such that the IF is at 5 MHz, the second
harmonic of the IF will be at 10 MHz, and the 37100A may
lock to this second harmonic. This would result in the source
being 5 MHz off in frequency. This problem only exists at
low IF frequencies, so the solution is to keep the IF frequency
above about 50 MHz. With two external synthesizers the
37100A does not phase-lock the source, so this problem will
not occur.
9
APPENDIX D: Removing the Effects of Gain/Phase Mismatch in the
Test System Channels
There are two ways this can be accomplished:
• Characterize the amplitude and phase offset of each test system component and combine them using an
external controller.
• Measure the DUT with each channel and using an external controller, determine the amplitude and phase offset.
Characterizing the Individual Components:
Measuring the DUT with Each Channel:
Characterizing the individual components and combining
them to determine the total system effect is the traditional
method of determining system error. The mismatch of the test
system components are characterized at every frequency point
and combined to produce the total system error.
First the power splitter and its cabling (the cables between
the splitter and the DUT, including any attenuators) are
characterized for each point over the RF range. This is easily
accomplished with an Anritsu 372/300 VNA. Measure the a1
side with the b2 side terminated, then reverse the procedure.
The ratio of b2/a1 is the amplitude and phase mismatch of the
splitter and its cables.
Next the receiver and its associated cables (the cables
between the DUT and the receiver, including any attenuators)
can be characterized over the IF range. This is a two step
process:
• Characterize a power splitter over the IF range.
• Use the power splitter to characterize the receiver.
A power splitter (which is used only for the system characterization) is characterized in the same manner just described.
This splitter is then connected to the 37100A with the RF Out
of the VNA driving its input, and its outputs connected to the
a1 and b2 ports. With the receiver set to measure b2/a1 the
measurement will show the total amplitude and phase
mismatch, which is in fact the receiver mismatch times the
splitter mismatch. If this number (receiver * splitter) is divided by the splitter error just characterized the result is only the
error of the receiver and its cables.
Now that all system components are characterized, the total
system error can be calculated as:
A more straightforward and simpler way of determining
system error uses multiple DUT measurements to find the
error. An amplitude and phase match measurement can be
thought of as having two unknowns; the error of the DUT
and the error of the system. If two sets of measurements are
made, forming two sets of equations, then these two
unknowns can be determined.
First measure the DUT with channel 1 connected to a1 and
channel 2 connected to b2 and record the result. Then measure
the DUT with channel 2 connected to a1 and channel 1
connected to b2. The system error is then determined, as
detailed below, from the formula:
RF splitter error x Receiver error = Total system error
With the total system error known, it can be ratioed out of
any subsequent DUT measurements.
(b2/a1)1 * (b2/a1)2 = System amplitude/phase error
where: (b2/a1)1 is the result of the first measurement
(b2/a1)2 is the result of the second measurement
Once the system error is known it can be ratioed out of
subsequent measurements.
If desired, the DUT difference is determined from the formula.
(b2/a1)2 / (b2/a1)1 = DUT amplitude/phase difference
( ab ) = SS
2
1
1
b2
a1
Sa1
DUT1
=
Sb2
DUT2
*
DUT1
DUT2
*
( ab )
1
( ba ) * SS
a1
1
Sb2
DUT1
=
Sa1
DUT2
*
1
1
1
* DUT1
b2
2
2
2
( ab ) = SS
=
b2
Sb2
Sa1
*
2
a1
DUT2
( ba )
( ab )
1
2
2
2
( ) * ( ) ( SS )
b2
a1
b2
a1
1
=
2
a1
( ba ) * ( ba ) =
2
1
2
1
1
b2
2
Sb2
Sa1
Sb2
b2
Where: S is the system error in the
a1 the measurement.
a1
DUT is the response of a channel in the converter.
10
1
2
2
Appendix E: Measuring Absolute Mixer Parameters, without Referencing to a Golden
Standard, the NxN Method
This application note describes two mixer measurement procedures; the Golden Piece method and the direct conversion
loss measurement technique. The Golden Piece method measures amplitude and phase match between the Device Under
Test (DUT), and a Golden Standard mixer. The accuracy of
this method is wholly dependent on the accuracy of the Golden Piece data. The direct method of measuring conversion
loss yields absolute magnitude but no phase information.
If absolute and accurate measurements including phase information are necessary, for applications such as phase distortion
measurements, the NxN technique is the recommended
method.
The NxN technique uses a standard VNA to make 12-term
error corrected S-parameter measurements on a Frequency
Translated Device. This method yields absolute and accurate
magnitude and phase information, with measurement speeds
suitable for real time tuning. Figure 8 shows the standard
setup for the NxN technique. Notice that with mixer 1 downconverting the signal, and mixer 2 up-converting it back, the
coherent frequency detection criterion of the VNA is
preserved.
68/9 SERIES SYNTHESIZER
TO DEDICATED
GPIB PORT
37300A VNA
PORT 1
PORT 2
MIXER 2
MIXER 1
IF
RF
LO
IF
BPF
RF
LO
P.D.
Figure 8. NxN Technique Setup
Fixed attenuators are recommended on the RF and IF side of
both mixers to improve matching and reduce SWR interaction
between the mixers. If enough power level is available, 6 to
10 dB attenuators should be used. In cases where excessive
loss cannot be tolerated, broadband isolators may be used.
The amplifiers in the LO path provide LO drive amplification,
and improved LO isolation. The band pass filter in the IF path
provides rejection to spurious signals from the mixers.
The Mixer 1 location in the setup is where the DUT will
eventually be inserted and measured. Note that a down-converter assembly could be inserted and measured also, provided that mixer 2 properly up-converts the signal back, meeting
the coherent signal detection criterion of the VNA.
Once a VNA is calibrated with standards at the end of its test
port cables, the measurement reference planes are established
at those two points. In order to measure the DUT at location
1, the calibration reference plane must be moved to before
and after mixer 1. A software package is available that can be
used for this specific calibration process. Once the calibration
reference planes have been effectively moved to the desired
new locations, the software package is disengaged, and the
DUT is measured in location 1.
Mixer Calibration Assistant software is available to guide the
user through the NxN calibration technique. The software
characterizes the IF path, comprised of the two IF side attenuators and the band pass filter. The two RF side attenuators are
calibrated out as part of the VNA measurement system, leaving the task of characterizing mixer 2. Using the above IF
path, RF Cal, and Mixer 2 data, the software creates a new
calibration file and loads it into the VNA for measurements at
location 1. Specifically, the new calibration file is equal to the
RF calibration including the RF attenuators, compensated for
the IF path and mixer 2.
The NxN technique is used for characterizing a mixer for
location 2. The technique uses a VNA to provide an inferred
response from several measurements, involving the swapping
of three test mixers. In particular, the technique is based on
the concept of performing three sets of measurements using
pair-wise combinations of three devices. The result is a set of
three simultaneous equations from which the individual
device responses can be determined. Once the three test mixers are characterized, either one can be chosen as the up-converting mixer in location 2.
The three mixers should have similar operating parameters,
but they need not be identical from a performance standpoint.
An important requirement of at least one of the three mixers
is reciprocity. Since one mixer is used in both forward and
reverse mode during the pair-wise combination measurements, it needs to display the same response in both configurations. Commonly used double and triple-balanced mixers
exhibit this property when operated linearly with their ports
terminated properly. A simple way to verify reciprocity is to
check the equality of s21 and s12 of a combination of two
mixers in back to back configuration.
11
Calibration Process using the Mixer Calibration
Assistant Software (p/n 2300-232):
Install the software on a PC running the Windows operating
system (16 or 32 bit), and containing a National Instruments
GPIB board. Configure the system as shown in Figure 8 and
attach the PC to the VNA GPIB port. Start the Mixer Calibration Assistant software program by double-clicking on the
icon. Figure 9 shows the first screen. Use the Sys Config icon
to change the GPIB address of the VNA.
Measure Filter and Interconnect Step:
This step characterizes the IF chain comprised of the IF attenuators, band pass filter, and any other components between
the two mixers. This characterization occurs at the IF frequencies, the operating range of these components.
If we assume that the final measurement will be made at 401
data points, all calibrations and measurements must also be
made at 401 data points. If the IF is a CW frequency, set the
VNA to 401 data points CW measurement.
Figure 10. Measure Filter and Interconnect
Figure 9. Mixer Measurement Calibration Assistant Main Screen
The following steps are used to calibrate the VNA for a mixer
measurement: Measure Filter and Interconnect, Measure Reference Mixers, and Measure DUT Mixers. The applicationspecific step Measure Attenuator and Amplifier must be
specifically selected. Full 12-term Calibration is the only calibration that can be used with this software. Other calibrations, such as Transmission Only, are not allowed and will
result in a GPIB error.
The software will prompt you to calibrate the VNA appropriately, and select Continue when ready, as shown in Figure 10.
It is important to wait until the VNA has completed a good
measurement before pressing Continue. If inaccurate data is
downloaded due to a rushed selection, the overall measurement accuracy will degrade significantly. Wait for two complete sweeps before pressing Continue, giving the VNA a
chance to update all forward and reverse terms' coefficients.
This warning is valid throughout this calibration process.
General Warning
Measure Attenuator or Amplifier:
In order to get accurate results, the response of the calibrated
components must remain constant before and after calibration. This consistency requires the utmost care during the
entire process. All connections should be torqued to the manufacturers' specifications. When adapters are required, precision components are recommended. These adapters might
sometimes need to be swapped due to different requirements
between the calibration and DUT measurement steps. In order
for the calibration to remain valid, the two different precision
adapters used must be electrically interchangeable. Anritsu
Calibration Kits contain different style Phase Equal Insertables (PEI) that have identical electrical length; therefore, calibrating with one and measuring with another will result in
accurate measurements.
This is an optional step as mentioned before. Select it and
press Start if needed.
12
This step is required when a component in the RF chain
needs to be characterized separately and compensated for in
the final calibration. In the normal setup where two RF attenuators are used, this step is not required. The power loss of
the RF attenuators does not cause an insufficient level at port
2 of the VNA, which would result in a low measurement Signal to Noise Ratio, (SNR).
Consider a case where the final DUT is a down-converter
assembly with high gain. In order to test this device, a mixer
for up-converting the signal back to the original frequency is
required at location 2. With a total of three of this type of
mixer, the system can be calibrated and the down-converter
can be measured. Due to the high gain of the DUT, significant
attenuation is needed at its input. During calibration that
attenuator is present in the system, but the high gain DUT is
replaced with a standard lossy mixer. Without this special
step, the SNR during calibration will cause degradation in
measurement accuracy.
Follow the instructions in the software, as shown in figure 12.
The calibration in this step is performed at the RF frequencies. During the RF calibration, the RF attenuators should be
installed on the VNA test port cables, and calibrated out as
part of the VNA system.
Figure 11 shows the step to characterize up to two RF chain
components separately.
Figure 12. Measure Reference Mixers
Figure 11. Measure Attenuator or Amplifier
Measure Reference Mixers:
The software will now guide you through three measurements, in order to characterize all three mixers. Check the following table:
Measurement #1
Measurement #2
Measurement #3
Location #1
Mixer #1
Mixer #2
Mixer #3
Location #2
Mixer #2
Mixer #3
Mixer #2
Labeling the three Mixers is recommended. At the completion
of the characterization steps, you will be asked to select one
of the three mixers as the up-converting mixer in the DUT
measurements. If the selected reference mixer does not match
the actual mixer inserted, the final measurement accuracy will
be inaccurate.
Notice that mixer #2 is the only mixer that is used both in the
forward and reverse direction. It is the only mixer that is
required to exhibit the reciprocity property discussed earlier
in this appendix.
Since the mixers are in the measurement path, the LO source
now comes into play for the first time. If a fixed-LO measurement is being made, the VNA does not have to control the
source. Any synthesized source may be used and controlled
manually. If the VNA is asked to control the source, as is necessary in a swept-LO measurement, the source has to be a
68/9XXX series El-Toro synthesizer. Refer to Appendix A for
instructions on how to use the VNA's Multiple Source Control
Mode to control an external source, and set the proper RF and
LO frequencies. In this case the receiver frequency equation
is set equal to the RF frequency, as the VNA is not aware that
a frequency translated device is being measured.
Remember to always use the same number of data points as
in the previous stages, and to always wait for two full sweeps
of the VNA before pressing Continue.
Measure DUT Mixer:
This is the final step before measuring the DUT. The screen
shown in Figure 13 allows the selection of one of three characterized mixers as the reference mixer to be used in
location 2.
Figure 13. Choosing the Reference Mixer
13
At this point the software is ready to download a calibration
from the VNA, compensate for the IF chain, reference mixer,
and optional attenuator, and then upload a new calibration
into the VNA. It will give you an opportunity to recalibrate
the VNA, if the current calibration is not suitable for the software to modify. This is rarely the case, unless another power
level calibration is required. Once Continue is pressed, the
software displays a message that the new calibration has been
successfully uploaded into the VNA, as shown in Figure 14.
You may remove the PC at this point, and measure DUTs at
location 1.
a1 LOOP
b1 LOOP
b1
37300A TEST SET
BLOCK DIAGRAM
AMP
LOOP
b2
SAMPLERS
a1
a2
INTERNAL SOURCE
PORT 1
PORT 2
b1 LOOP
SWITCH
a1 LOOP
SWITCH
MIXER 1
MIXER 2
BPF
P.D.
EXTERNAL SOURCE
Figure 15. Adding Isolation and Match Measuring Capabilities to the
NxN Measuring Technique
Figure 14. Perform final cal screen with cal uploaded message.
Figure 16 shows mixer conversion (magnitude and phase),
group delay, phase linearity, and input match measurements.
Custom systems that can be used for port to port isolation,
and all three ports match measurements are also available
from our Custom System Solution Group. A typical test set
block diagram is shown in Figure 15, where back to back
couplers are added in the IF and LO paths of mixer 1. For
more information about custom solutions for your measurement needs please call your Anritsu representative.
Figure 16.
Typical mixer
measurements
after and NxN
calibration.
14
Appendix F: How to Apply the 37100A Material to a 37200B or 37300A Anritsu VNA
The 37100A and 37300A block diagrams are shown in Figures 1 and 9 respectively. The 37300A’s reference loops in the back
of the instrument allow direct access to the a1 and b1 samplers. The 37200B series does not offer the step attenuators. Only a
reference loop to the a1 sampler is available as an option. However, additional reference loops on either series can be quoted as
a special. The advantage of these two configurations is the ability to make S-parameter measurements without the need of an
external reflectometer test set.
Amplitude and Phase Tracking, Group Delay:
With the reference loops removed, Port 1 becomes equivalent
to the 37100A RF Out port. Ports a1 and b1 will be available
at the back. The equivalent setup for Amplitude and Phase
Tracking, and Group Delay measurements is shown in
Figure 10. Notice that the b1 port is used instead of b2.
68/9 SERIES SYNTHESIZER
to a1 in
to b1 in
TO DEDICATED
GPIB PORT
PORT 1
PORT 2
IF
FILTERS
37300A VNA
Conversion Loss:
For Conversion Loss measurements use port 1 as RF out,
and feed the IF output directly into the a1 sampler at the
back.When the sampler loop is removed, the second port
should be properly terminated into 50 ohms.
Reflection Coefficient, Port to Port Isolation:
In the setup shown in Figure 11, the Source 1 and Receiver
frequencies are set equal. When set to the RF frequencies, the
RF Port Reflection and RF to IF Isolation can be measured
directly. When set to the IF frequencies the IF Port Reflection
and IF to RF Isolation can be measured.
Rotating the mixer such that the LO port is connected to the
VNA’s port 1, will result in the LO Port Reflection, and IF to
LO and LO to IF Isolations.
The 37300A has a front panel amplifier loop where an
external amplifier can be inserted to properly drive the
LO port of a mixer.
GOLDEN
STD
RF
IF
68/9 SERIES SYNTHESIZER
TO DEDICATED
GPIB PORT
LO
SPLITTER
SPLITTER
LO
RF
IF
37300A VNA
DUT
PADS
PORT 1
PORT 2
REF MIXER
Figure 10. Mixer Amplitude and Phase Tracking, and Group Delay
Setup using the 37300A.
DUT
RF
IF
LO
Figure 11. Mixer Reflection Coefficient, and Port to Port Isolation
Setup using the 37300A.
15
All trademarks are registered trademarks of their respective companies.
Sales Centers:
United States
(800) ANRITSU
Canada
(800) ANRITSU
South America 55 (21) 286-9141
April 1998; Rev: B
Data subject to change without notice
Microwave Measurements Division • 490 Jarvis Drive • Morgan Hill, CA 95037-2809
http://www.global.anritsu.com • FAX (408) 778-0239
Sales Centers:
Europe
44 (01582) 433200
Japan
81 (03) 3446-1111
Asia-Pacific
65-2822400
11410-00197
AN360B/37XXXA/B-1 /GIP-E
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising