Gain Compression for Amplifiers GCA (Opt S9x086A 086)

• Features, Requirements, and Limitations

• Gain Compression Concepts

• Understanding the GCA Displayed Traces

• Gain Compression Parameters

• Compression Methods

• Acquisition Modes

• Using Gain Compression App

• • Frequency tab

  • Power tab

  • Compression tab

  • Safe Sweep Mode dialog

• Compression Analysis

• Saving GCA Data

• GCA Measurement Tips

• Macros

See Also

• Gain Compression for Converters

• GCA Calibration

• Programming commands

• App Note Amplifier Linear and Gain Compression Measurements

Other VNA Applications

Features, Requirements, and Limitations

Features

• Fast, easy, and complete Gain Compression measurements for amplifiers.

• Many compression parameters to choose from, including gain, input power at compression, output power at compression, input match, and compression level.

• Several compression methods to choose from, including deviation from linear gain, deviation from max gain, back-off, and X/Y, and compression from saturation.

• Three acquisition methods to choose from: Power per Freq, Freq per Power, and SMART Sweep

• SMARTCal Calibration Wizard to guide you through Full 2-Port or Enhanced Response calibration, plus Source Power calibration.

• Compression Analysis allows traditional power sweep view at a selected frequency.

• Receiver Leveling provides continuous source power accuracy.

• Supports Wideband (NOT Narrowband) Pulse measurements using the new integrated Pulse setup dialogs.

Requirements

• Option S9x086A (software option only) must be enabled.

• When performing an optional calibration:

• • ECal module or Calibration Kit

  • Power meter/sensor

Limitations with GCA

• Number of points limited to 100,001 for two-dimensional acquisitions, 50,000 points for SMART Sweep.

• Standard CW power sweep is NOT supported in a Gain Compression channel.

• Independent IFBW, Power Levels, Shift LO, or Sweep Time in a segment table is NOT supported.

• Stepped sweep mode only.

• Linear, Log, and Segment frequency sweep modes only.

The following VNA Features are NOT Available in a Gain Compression channel:

• Unratioed receiver measurements (A, B, R)

• ECal User Characterization

• Some Fixturing Features

• FOM or FCA

• External Test Set Control (Option S93551A/B)

• CW Time sweep

• Balanced measurements

• Save Auto Formatted Citifile data.

• Narrowband Pulse measurements using the Integrated Pulse App

Gain Compression Application Concepts

What is Gain Compression

An amplifier has a region of linear gain, where the gain is independent of the input power level. This gain is commonly referred to as small signal gain. As the input power is increased to a level that causes the amplifier to approach saturation, the gain will decrease. The 1 dB gain compression is defined as the input power level that causes amplifier gain to drop 1 dB relative to the linear gain.

You can quickly measure the gain compression using a compression marker on a power sweep trace.

Terms used in GCA

Linear Power Level  The specified input power that yields linear gain (also known as 'small-signal gain') in the amplifier.

Reference gain  The measured gain that is used as a reference for determining compression level.  The Compression Method that is used could cause this value to be different.

Compression level  The specified amount of gain reduction from the reference gain.

Target gain  The gain at the specified compression level.  Although this term does not appear in GCA, it is important to understand when discussing the various compression parameters.

For example, when using Compression from Linear Gain method with the following settings:

• Linear gain (measured at Linear Input power) = 10.2 dB

• Compression level (specified) = 1 dB

• Target gain = 9.2 dB

This is called 'Target' gain because GCA will search for the closest measured gain to 9.2000 dB. It may not measure this gain exactly.

Compression point  The operating point at which the measured gain is closest to the Target Gain. All compression parameters report data for this operating point.

Understanding the GCA Displayed Traces

One of the most important concepts to remember with GCA is that, each frequency data point represents many measurements using different input power levels.

Some things to notice about how GCA displays compression data:

1. The X-axis values are ALWAYS frequency. Imagine behind each frequency data point, a traditional power sweep curve with corresponding measurements and calculations to find the specified compression point.

2. The Y-axis values are always reported at the compression point. The value that is displayed depends on the compression parameter that you choose.  The S-parameters that are displayed in a GCA channel are always measured at the linear and reverse power level.

Example: Five of the six GCA compression parameters are displayed in the above image. The missing trace, DeltaGain21 is discussed below.  

• Markers are placed at 4.549 GHz for all of the parameters.

•  Tr 3 CompIn21  (Input power at the compression point) shows the marker value to be -5.4117 dBm. This is the power into the DUT that was required to achieve the compression point   Notice that this is about the same input power required to achieve the specified compression at ALL frequencies.

•  Tr 5 CompGain21  (Gain at the compression point) shows the marker value  9..6443 dB . This is the measured gain at the compression point.

• To see the gain at a different input power at this frequency, use the Compression Analysis feature.

Create a GCA Measurement

1. Press Meas > S-Param > Meas Class....

2. Select Gain Compression, then either:

   • OK delete the existing measurement, or

   • New Channel to create the measurement in a new channel.

Gain Compression Parameters

There are several Gain Compression parameters, as well as standard S-parameters and ADC parameters, that can be measured in a GCA channel.

Linear S-Parameters

For convenience, the standard S-parameters are offered in a GCA channel. S11 and S21 are measured at the specified Linear Input level. S22 and S12 are measured at the specified Reverse power level.

Note: When a DC meter is added, it will be displayed in the New Trace dialog and SMART Sweep Safe Mode dialog (in DC Parameters pull down menu).

ADC Parameters

Four ADC analog-to-digital converter measurements are offered in a GCA channel:

• AI1 and AI2 are measured at the specified Linear Input level.

• CompAI1 and CompAI2 are measured at the specified compression point.

These DC measurements, along with Equation Editor, allow you to make PAE measurements at the Linear Input level and compression point.

These measurements are made at pins 7 and 8 of the VNA Power IO connector.

Compression Parameters

The raw data for these parameters are always measured.

Compression Methods

GCA offers the following methods to find the compression point of an amplifier using GCA:

Compression from Linear Gain

The Reference Gain is measured using the specified Linear (Input) Power Level. The Target Gain is calculated as the Linear Gain minus the specified Compression Level. For example 8.3 dB - 1 dB = 7.3 dB.

Compression from Max Gain

The linear region of an amplifier gain may not be perfectly linear.  The highest gain value that is found at each frequency is used as the Reference (S21) Gain.  The Target Gain is found in the same way as Compression from Linear Gain.

Compression from Saturation

This method is used to better find the compression point when measuring amplifiers with non-linear gain as shown in the following image:

The Max power out value * is found at each frequency. Then input power is lowered until the output power decreases by the specified 'From Max Pout' value. This is the compression point.^

Backoff and X/Y method

These two compression methods are very similar.

• Both methods specify a difference in input power (X axis) between the linear region and compression point.

• For the Y-axis difference:

• • Backoff method specifies Compression Level which is a difference in Gain.

  • X/Y method specifies Delta Y which is a difference in Output Power.

GCA searches for these points differently for 2D sweeps and SMART sweep.

The following images show how Backoff and X/Y method is calculated at ONE frequency.

Acquisition Modes

The GCA offers three modes for data acquisition: Two 2D sweep modes, and SMART sweep.

To see a traditional power sweep at a single frequency, use the Compression Analysis feature. Learn more.

2D (two-dimensional) Sweeps

This is the easiest method to understand, and the least efficient for finding the compression point. Both 2D sweep modes work as follows:

1. All GCA measurements begin by measuring S-parameters at the specified Linear Power level. Reverse parameters are measured ONLY if Full 2-port calibration is applied or if a reverse parameter is displayed. Learn more about Cal choices.

2. Gain measurements are then made at ALL of the specified frequency and power values. Although these are conceptually 2-Dimensional sweeps, a single sweep is constructed in firmware.  See Data Points Limit.

3. After data has been measured, a search is performed to find the compression point. You can choose to interpolate between the two measured points closest to the target gain. Learn more.

As each sweep is performed, dots are plotted next to the Ch indicator in the lower left corner of the display to indicate progress for the current sweep.

Note: For Backoff and X/Y compression method, GCA does not verify that the specified Start - Stop power range is at least the size of the specified Backoff or X value. The closest compression point is always reported.

Note: SMU Hardware List trigger mode is NOT supported in GCA 2D sweeps.

2D Sweep Modes

• 2D Sweep Power per Frequency -  Input power is stepped from Start to Stop at each specified frequency.  From the following example you can see that the device is exposed to the highest power level (p3) at the first frequency (f1). This could heat the device early in the measurement and affect compression results.

The following examples show (frequency, power) values for three frequency points and three power points, resulting in a total of 9 measurements:

• 2D Sweep Frequency per Power -  Frequency is swept from start to stop at each specified power level as follows:

Viewing and Saving 2D Data

It is NOT possible to plot ALL of the 2D measurement data on the VNA display. However, it can be saved to a *.csv file and then read into an Excel spreadsheet.  The initial S-parameter measurement data is not saved to this file. Learn more.

You can also view on the VNA all power sweep information at a selected frequency using the Compression Analysis feature.

SMART Sweep

SMART Sweep is usually the fastest and most accurate method to measure Gain Compression. Unlike the 2D acquisition modes which measure all of the specified frequency / power points, SMART Sweep performs a series of power search iterations. At each frequency, an 'intelligent guess' of input power is made to find the compression level that is within tolerance. This guess is further refined with each successive power search iteration sweep.  

SMART Sweep continues to iterate until one of the following conditions occur:

1. ALL data points are within tolerance. When the compression level for a data point achieves the specified tolerance, it continues to be measured and input power changed to improve the measurement within tolerance.

2. The specified compression level can NOT be achieved for the remaining frequencies that are not in tolerance. Either the Start power is too high or the Stop power is too low.

3. Maximum iterations have been achieved.  If a measured gain is not within the specified tolerance before the specified Max number of Iterations has been reached, then the last power reading is used as the compression point.

The Iteration Counter, Dots, and Bangs(!) Next to the Ch indicator, in the lower left corner of a GCA window, the following annotation appears: An iteration counter is incremented each time input power is adjusted. A dot appears when another 10% of the frequency points are within tolerance. ! (bangs) are displayed after the last iteration. Each bang represents 10% of the data points that are NOT within tolerance.

SMART Sweep and Compression Method

The intelligent guess process works differently depending on the compression method.  This is important because Backoff and X/Y compression methods subject the DUT to significant changes in input power during an iteration sweep.  This can affect the DUT and the measurement results.

Learn all about Backoff and X/Y compression methods.

ALL GCA measurements begin by measuring S-parameters at the specified Linear Power level. Reverse parameters are measured ONLY if Full 2-port calibration is applied or if a reverse parameter is displayed. Learn more about Cal choices.

• Backoff and XY   Because both compression methods specify the separation between the 'linear" region and the "compressed" region, each iteration requires a single sweep at two dramatically different power levels over the same frequency range. The first half of the sweep measures the DUT at the Backoff or X power level. The second half of the sweep measures the DUT at the compressed power level, specified by the Start and Stop power range.  At the beginning of the second half, the power level rises by the Backoff or X value. The specified Settling Time is applied at this point to allow the DUT time to react to this significant change in power level.  Safe Sweep does NOT minimize this change in input power. However, Safe Sweep with Backoff and XY methods DOES prevent the DUT from being exposed to too much input power.

• Compression From Linear Gain  After the reference gain is measured at the linear input power, the next iteration measures the DUT at a higher power level which attempts to push the DUT well into compression. Subsequent sweeps, depending upon the compression level of the DUT, either increases or decreases the power in order to reach the desired compression level. Usually, by the third iteration sweep, a curve-fit algorithm is utilized to precisely find the compression point.

Note: The DUT can be subject to significant changes in power from one iteration sweep to the next. This can be minimized by the use of SAFE Sweep and careful selection of the corresponding settings.

• Compression from Max Gain  The maximum gain that is found at each frequency is stored and used to calculate the compression point. SMART Sweep does NOT perform extra iterations to search for the maximum possible gain of the amplifier at each frequency.

• Compression from Saturation  The maximum power out that is found at each frequency is stored and used to calculate the compression point. SMART Sweep does NOT perform extra iterations to search for the maximum possible power out of the amplifier at each frequency.

Using the Gain Compression Application

The following is a general procedure for performing a GCA measurement. The challenge with GCA is configuring a measurement that yields the true compression performance of YOUR DUT. This requires knowledge of the Gain Compression settings and knowledge of the DUT.

See specific dialog boxes below.

1. Disconnect the DUT if preset or default power levels may damage the VNA or DUT.

2. Preset the VNA, or configure a suitable User Preset that will be safe in case the DUT is connected.

3. Create a GCA channel. Learn how. The default trace is S21.

4. Start GCA Setup dialog and configure the measurement settings based on the DUT, adapters, attenuators, booster amplifiers, and fixtures to be used in the measurement.

5. Save the instrument state (optional).

6. Connect DUT and apply bias and RF power as appropriate. The default measurement for a GCA channel is S21 (amplifier gain). Inspect the gain measurement to ensure the DUT is operating as expected.

7. Add GCA compression parameter traces. Learn how.

8. Adjust the measurement settings to yield satisfactory compression parameters. See GCA Measurement Tips.

9. Start and complete the GCA Calibration wizard.

Compression Analysis

Compression Analysis changes the current trace into a power sweep trace at a specified CW frequency . The current parameter and acquisition method is unchanged. For example, with a CompGain21 trace displayed and SMART Sweep selected, enable Compression Analysis. The trace becomes a power sweep trace at the specified CW frequency. The Y-axis displays S21 Gain at each X-axis power point.

When Smart sweep is used, a complete power sweep is not performed, but only the data points that are required to find the compression point. To see a traditional power IN vs power OUT compression sweep, use one of the 2-D acquisition methods.

You can create PNOP or PSAT markers on a CompOut trace with Compression Analysis mode ON. Learn more.

Saving GCA Data

Beginning with VNA release A.08.20, GCA data can be saved to a *.csv file in both 2D and SMART Sweep modes (previously only 2D modes). Also, a Delta Gain, AI1, and AI2 columns have been added to the data. Learn about ADC parameters.

Notes

• This data type can be read by spreadsheet programs, such as Microsoft Excel.

• Data from the last complete sweep is saved to the specified *.csv file.

• If calibration is turned ON when the file is saved, then all data is calibrated. Otherwise, raw data is saved.

• All *.csv data saves include a reference power level sweep at the beginning of each frequency data.

• GCA Sweep Data (*.csv) includes sweep data.

SMART Sweep data with 5 iterations and 3 frequency points. The yellow highlight is added here for readability.

When saving or recalling 2D data:

• When Linear Input Power EQUALS Start Power, then the number of data points (rows)/ freq = num power points.

• When Linear Input Power does NOT EQUAL Start Power, the number of data points (rows)/ freq = num power points + 1.

• Make these selections on the GCA/GCX Power tab dialog.

GCA Measurement Tips

There are many settings in the Gain Compression Application. Here are a few tips when using GCA to learn as much as possible about the compression characteristics of your DUT in the most efficient manner.

DUT Compression Characteristics and GCA  

Although GCA provides excellent results with a wide variety of amplifiers, it works best with amplifiers which have a monotonic compression curve.  In some cases where the compression curve is not monotonic, for example if the amplifier gain expands before it compresses, the correct compression level may not be found.

To help a SMART sweep find the correct compression point, limit the Start and Stop power levels around the anticipated compression point.  Learn more.

The following two power-sweep traces are examples of non-monotonic gain:

DeltaGain

A DeltaGain trace is the best way to see how closely GCA is actually measuring to the desired compression level.  In addition, you can view the phase of DeltaGain to see the phase deviation between the compressed gain and the reference gain. DeltaGain is calculated as:

• DeltaGain = Measured Gain (watts) / Ref Gain (watts)

• In LogMag format: DeltaGain = (Measured Gain) - (Ref Gain)

With SMART Sweep, DeltaGain (in LogMag format) shows how soon certain frequencies achieve the specified tolerance.  Learn more.

Some other settings which may be helpful:

• Trigger source: Manual allows you to analyze data and make adjustments while allowing the device to cool.  

• Construct Limit Lines around the compression point at the tolerance level.

The following image shows a DeltaGain21 trace using SMART Sweep. The Limit Lines were added manually.

In the above image:

SMART Sweep Tips

• Compression from Linear Gain is the easiest compression method to understand and control in SMART Sweep. Learn more.

• If SMART Sweep requires more than twenty iterations, this is an indication that something is wrong. Try changing the Tolerance setting, Frequency Range, Start / Stop power range, IF bandwidth, or Dwell Time.

• If the number of iterations required to achieve the desired compression level changes significantly from one set of measurements to the next, this could be due to other effects, such as heating. Try increasing the dwell time or using a wideband pulse measurement configuration.

• If the DUT should not be significantly overdriven into compression, or the changes in the input power should be limited, use Safe Sweep mode with Deviation from Linear Gain compression method.

Single Frequency Macros

Note: Beginning with VNA rev. A.09.00, the Compression Analysis feature provides an easier method of viewing a traditional power sweep at a single frequency than the GCA macros. However, the Macros are still maintained on the VNA hard drive.

The macros perform a single power sweep on the DUT using a standard channel with corresponding stimulus settings. The macro can show measurement differences from the compression analysis traces due to bias/thermal/settling effects of the DUT. So, the macro can help confirm a DUT is exhibiting some type of settling behavior which will need to be handled in some way.

Also, the macro is a great GCA programming example.

With a 2D sweep (NOT SMART Sweep) a script that is stored on the VNA hard drive automatically creates a traditional power sweep measurement in a standard channel using the same stimulus setting as the GCA channel. Use a marker in the GCA channel to specify the frequency for the measurement.

The script has two modes of operation:

1. View Mode displays all of the previous 2D sweep data at that frequency.

2. Measure Mode performs a new measurement at that frequency.

Both modes create a new S-Parameter channel using the same stimulus settings as the GCA channel, including port power, attenuator, IF Bandwidth, and dwell settings. The new channel does not support calibration or pulse characteristics.

To see noise on a measurement, use the Measure macro in continuous sweep. Adjust the IFBW and averaging until the noise versus sweep speed meets your needs.

To see other effects of your DUT at a specific frequency, use the View macro and the Measure macro with 2D sweep mode. Both macros present data using a standard channel. The View macro shows 2D data at a specific frequency, while the Measure macro shows freshly-measured data at the same frequency. Ideally, the data from these two would be identical. However, changes in your DUT behavior due to heating or other effects can cause these to be different. If significant differences exist, try:

• Using the 2D Frequency per Power setting rather than Power per Frequency

• Adjusting the dwell time

• Adjusting IFBW

• Use a wideband pulse configuration

How to setup the Macros

Each macro must be setup separately.

1. Press Macro > Key Setup > Macro Setup....

2. Select a blank line, then click Edit.

3. In Macro Title, type a short description such as Meas GCA or View GCA.

4. Click Browse, then navigate to C:\Program Files(x86)\Keysight\Network Analyzer\Applications\GCA\GCA.vbs

5. In Macro run string parameters:

   1. Type M for the Measure macro or V for View macro.

   2. Optional: Supply the following additional parameters in any order:

1. • To run the program from a remote computer, specify the full computer name of the VNA .

   • Channel in which to create the measurement. If not specified, Measure is created in Ch30 and View is created in Ch31.

   • Example: Run string parameters for the Measure macro run from a remote computer in Channel 5.---- M MyVNA 5.

6. Click OK.

How to run the Macros

On a GCA channel:

1. Create a 2D sweep. Either Power per Freq or Freq per Power. Both macros always create a power sweep at the frequency of interest.

2. Create a CompIn trace.

3. On the CompIn trace, right-click and select Add Marker. Drag the marker to the frequency of interest.

4. Press Macro, then select either by the short description your provided in Step 3.