Amplifier Testing
with the 37300C VNA
APPLICA TION NOTE
Lightning™ VNA
Linear T ests
Linear tests are standard S-parameter measurements. They
provide information on the following amplifier parameters:
S21 Gain- magnitude as a function of frequency as well as drift
performance. Phase and Group Delay.
S11, S22 Input and output impedance and match.
S12 Isolation.
The important factor is that the above S-parameters are made
with the amplifier operating in its linear range. This range is
usually known from amplifier specifications that typically
specify operating gain, frequency range, and power supply
requirements as shown in Table 1.
Linear operation can be confirmed by increasing
the input power by 3 dB and observing that S21
remains the same. The requirement to operate in
the linear range often dictates and input level far
below the default power of a VNA, particularly for
high gain amplifiers. The Anritsu 37300C VNAs
include a test port power specified in dBm as well
as a built-in step attenuator that simplifies setting
the appropriate input signal level. On the output
side it is important that the power level input to
the VNA does not exceed its capability either for
power handling or measurement linearity. The
standard 37300C systems can handle a maximum
input of 1 watt (+30 dBm).
Typical Amplifier Specifications
Frequency 8 to 12 GHz
Minimum Gain 22 dB
Gain Flatness ±1 dB
Maximum SWR
Input::
Output:
2:1
2:1
Output Power @ 1 dB Gain compression +8 dBm
Table 1.
Introduction
Vector Network Analyzers (VNA) are designed to measure S-parameters of RF and Microwave devices and circuits.
S-parameters are by definition ratios and as such the actual test power utilized is not critical. When amplifier S-parameters
are measured it becomes important to know approximate input and output power levels so that the user can insure that
the amplifier is operating in its proper range and that the instrument's power handling capability is not exceeded. If
the user wants to measure non-linear effects such as the 1 dB compression point or AM to PM conversion, it is necessar y to
know power levels more accurately and it becomes desirable to sweep the power over a desired range. The Anritsu 37300C
has a number of features: nominal port power specified in dBm, power sweep, power meter calibration and a user
interface that simplifies and increases productivity for testing amplifiers.
There is a built-in step attenuator associated with the S21
measurement receiver that can reduce the receiver input
power to less than –10 dBm which insures linear instrument performance (see figure 1).
Figure 2 shows a four S-parameter display of amplifier
parameters.
Group Delay
Group delay is an important parameter for amplifiers used
in communications systems. It is a measure of the time it
takes for energy at a given frequency to travel through a
device. Unless all important frequency components have
essentially the same delay, distortion occurs in the output
signal. Group delay is derived from the transmission phase
response by the expression:
The actual appearance of the group delay display is very
dependent upon the aperture (∆ϖ) selected. Aperture
and smoothing are terms that are used in describing delay
measurements and they are essentially the same from a
functional point of view. The 37300C gives the user control
of this parameter. It's very important that apertures be
close to identical if group delay measurements are to be
compared. Figure 3 shows the group delay of an amplifier.
The aperture for this display was set to 3% of the sweep
width: 420 MHz.
2
Figure 2. Four S-parameter display of amplifier parameters
Figure 3. Group delay of an amplifier
Figure 1
Front Panel
Amplifier Loop
a
1
a
2
S
=
21
Rear Panel
Loops
Port 1
b
1
Receivers
b
2
Port 2
DUT
NOTE:Port 2 Coupler Loss is 13 dB
0.000 dB
.5
.2
0
-.5
OUTPUT IMPEDANCE
ISOLATION
LOG MAG.
1
GHz
1
1
-1
20.000 dB/DIV
18.000000000
2
-2
5
-5
MARKER 1
ALL DISPLAYED
CHANNELS
CH 1 - S11
10.000000000 GHz
101.544
5.373 j
CH 2 - S12
10.000000000 GHz
-64.674 dB
CH 3 - S21
10.000000000 GHz
18.977 dB
CH 4 - S22
10.000000000 GHz
30.649
-10.238 j
MARKER TO MAX
MARKER TO MIN
MARKER READOUT
FUNCTIONS
PRESS <ENTER>
TO SELECT
INPUT IMPEDANCE
.2
0
-.2
-.5
S21
0.000 dB
4.000000000
.5
1
-1
LOG MAG.
1
GHz
GAIN
1
10.000 dB/DIV
18.000000000
S12
2
5
-5
-2
4.000000000
S22
Z
-.2
Delay=
Ư
∆ϖ
b
2
a
1
Ω
Ω
Ω
Ω
Nonlinear T esting
Active devices are usually designed to provide linear,
distortion free operation for a specific level or power range
of input signals. If the input level is increased there will a
transition from linear to non-linear operation - the output
will start to compress when the device can no longer
provide its nominal linear gain. A common specification is
the 1 dB compression point, which indicates the level at
which the gain has decreased by 1 dB. As the input power
is further increased the gain will continue to drop and
depending upon the device and application, the user may
want to observe performance at these higher compression
levels. As more and more power is driven into the input
the gain will continue to drop until the point of saturated
output power has been reached (Psat). At Psat no further
increase is output power can be obtained.
Nonlinear testing can be done with a number of
instruments including Vector Network Analyzers.
Parameters that must be considered for this type of test
include: Test Port Power, DUT input power, power sweep
range, and the output power of the amplifier. The
block diagram below is representative of this type of
test environment.
In most applications the user is interested in measuring
both S-parameters and nonlinear performance. As
described above, S-parameters are typically measured in
the amplifiers linear region. Input power is then increased
so that the amplifier's nonlinear performance can be
measured. It is recommended that the frequency and
linear operation power parameters be established before
performing the S-parameter calibration. This will minimize
the possibility of inadvertently overpowering either the
device or the receiver during testing.
Power and VNA’s
It is necessary to measure absolute power to determine
Gain Compression. VNA receiver channels are typically
downconverters and do not measure power directly. They
are; however, linear so that an accurate power calibration
at one level will result in a receiver channel that will
accurately indicate power in dBm. This is accomplished by
calibrating the VNA using an external power meter that in
effect transfers the power meter/sensor accuracy to the
VNA. The 373XXC firmware will support calibration with
the Anritsu ML2430A.
These meters differ in the way they handle sensor
efficiency (consult the power meter manual). The
37300C does expect to receive corrected data from
the power meter.
Errors can result if the proper correction factor isn't
applied by the power meter.
The vector error correction available in VNA's is
dependent upon ratioed S Parameter measurements.
Power is measured using a single unratioed channel;
therefore, when power is being measured error correction
must be turned off.
3
Figure 4. Block diagram of test environment
Correction Factor Error (dB)
1% 0.043
3% 0.128
5% 0.212
10% 0.414
Table A.
4
Gain Compression
As mentioned above there are a number of ways to
measure Gain Compression. Power meters can be used,
although this is often a tedious procedure, scalar analyzers
can also be used and VNA's can do the job as well. Each
method has some limitations in either convenience or
accuracy. The VNA does have the advantage that it can
also provide S-parameter data. The phase information
associated with S21 can provide group delay and AM/PM
data, which is not available using the other instruments. It
should also be noted that VNA's are tuned receivers and
do not respond to harmonics that may be present in the
device output. When using a VNA two approaches can be
used: Swept Frequency Gain Compression (SFGC) and
Swept Power Gain Compression (SPGC). The 373XXC
offers a very straightforward approach to each of these
measurements.
It is normally desirable to make S Parameter measurements in the linear operating region of an amplifier and
then observe Compression or AM/PM characteristics by
increasing the input power sufficient to drive the amplifier
into its nonlinear region. The characteristics of the
amplifier to be tested (AUT) dictate the operating power
levels required for the tests. Prior to making measurements
on a specific amplifier the user must determine the
desired operating levels. A recommended level for linear
region operation is:
P = PGC - Gain - 15dB (PGC = Nominal 1 db compression of the AUT)
The actual level is constrained by the power available
from the VNA in conjunction with the built in 10 dB step
attenuator. In the case of the 373XXC this is easily supplemented by the addition of an external amplifier/attenuator combination. The power input to Port 2 must also be
considered, as the test should not drive the VNA into
nonlinear operation. Typical specifications show 0.1 dB
compression at a VNA receiver input level of –10 dBm.
The receiver signal (this is the signal at b2 in Figure 2) is
derived through a 13 dB coupler from the Port 2 signal.
The 373XXC includes a 10 dB step attenuator in this path
that enables linear operation with input signals as high as
30 dBm (1 watt), the maximum signal level that should be
input to Port 2. Higher power levels can be measured by
attenuating the signal prior to Port 2.
The following sections will show results obtained from tests
on a specific amplifier performed on an Anritsu 37347C.
Appendix A includes a complete description of the set up
used for the tests.
Swept Frequency Gain Compression
This is a manual procedure in which the user obtains a
normalized amplifier response as a function of frequency
at Pstart and manually increases the input power while
observing the decrease in gain as the amplifier goes into
compression. This permits the user to easily observe the
most critical compression frequency of a broadband
amplifier. The SFGC process is implemented in the
37300C by following the procedure outlined after
accessing the Applications (APPL) menu and selecting
Swept Frequency Gain Compression. This brings up the
menu shown in Figure 5a and the user follows the menu
instructions that would normally include a power meter
calibration detailed in Figure 5b at the power level defined
by Power Target (Power Target is usually set to Pstart).
The menu guides the user through the steps required to
obtain a result such as that shown in Figure 6. This clearly
indicates that for this specific amplifier, the 1 dB compression point is reached first at the high end where the gain
is maximum.
Figure 5a.
Figure 5b.
- SWEPT FREQUENCY GAIN COMPRESSION -
PORT 1
AMP ATTN
NOMINAL OFFSET
1. TEST PORT 1 POWER SHOULD BE APPROXIMATELY =
AUT (x dB compression spec) - AUT (gain) - 10 dB
2. PORT 2 INPUT POWER SHOULD BE LESSTHAN 0 dBm
(UNLESS OPTION 6 IS INSTALLED). FORWARD DIRECTION
1-PATH 2-PORT ORTRANSMISSION FREQUENCY RESPONSE
CALIBRATIONS ARE PERMITTED, BUT NOT NECESSARY.
3. NOMINAL OFFSET = APPROXIMATE GAIN (OR LOSS)
OF EXTERNAL DEVICES PRECEDING THE AUT.
4. DEFAULT DISPLAY IS DUAL CHANNEL 1-3 IN WHICH
CHANNEL 1 = b2/1 [dBm] AND CHANNEL 3 = S21.
- MEASUREMENT INSTRUCTIONS -
1. AFTER THE AUT IS CONNECTED, NORMALIZE S21.
2. INDICATETHE GAIN COMPRESSION POINT VALUE (x dB)
AND SELECT <TEST AUT>.
3. INCREASE TEST PORT 1 POWER UNTIL A 1 dB (or x dB)
DECREASE IN S21 IS OBSERVED.
TEST PORTS
12
- CALIBRATION INSTRUCTIONS -
ATTN
PORT 2
SWEPT FREQUENCY
GAIN COMPRESSION
NOMINAL OFFSET
0.00 dB
CALIBRATE
FOR FLATNESS
(NO CAL EXISTS)
FLATNESS OFF
CORRECTION
CALIBRATE
RECEIVER
(NO CAL EXISTS)
NORMALIZE S21
(NOT STORED)
GAIN COMPRESSION
POINT (0 dB REF)
1.00 dB
TEST AUT
EXIT APPLICATION
Loading...