Complex Stimulus-Response Measurements

The 89600 VSA software provides the ability to measure the input/output relationship of a power amplifier, stimulated with complex signals. The VSA software characterizes this distortion or compression with AM Amplitude Modulation - CW modulation using amplitude variation in proportion to the amplitude of the modulating signal. Usually taken as DSB-LC for commercial broadcast transmissions and DSB-SC for multiplexed systems./AM, AM/PM, and Gain Compression traces, and with the differential Error Vector Magnitude (EVM Error vector magnitude (EVM): A quality metric in digital communication systems. See the EVM metric in the Error Summary Table topic in each demodulator for more information on how EVM is calculated for that modulation format.) metric. These are summarized in Measurement Results below and covered in detail in the Graph Traces and Graph Results topics. Precisely aligned stimulus and response waveforms are also provided for use in external algorithms to study and correct amplifier distortion.

See the Measurement Overview topic for procedures on making complex stimulus-response measurements.

Complex vs. CW Signal (Sine wave) Effects on Distortion

CW signals are insufficient for testing amplifier distortion for several reasons.

The response of a nonlinear device depends on the signal, which excites it in a generally nonlinear way and also a non-local way. Non-local means that the response depends not just on the instantaneous values of the signal at each time, but also on the shape of the signal over time, e.g. the waveform. Specifically, values of the excitation at previous times, that is, the history of the excitation, can affect the nonlinear response. Complex signals look very different in time from CW signals, and hence the resulting distortion will be different.

That the ACPR Adjacent Channel Power Ratio - A measurement of the amount of interference, or power, in the adjacent frequency channel. ACPR is usually defined as the ratio of the average power in the adjacent frequency channel (or offset) to the average power in the transmitted frequency channel. ACPR is a critical measurement for CDMA transmitters and their components. It describes the amount of distortion generated due to nonlinearities in RF components. The ACPR measurement is not part of the cdmaOne standard. of the same amplifier is different when stimulated by different modulation formats is a consequence of this fact. There are even significant differences between the response of a nonlinear device to one-tone and two-tone excitations. The 1dB compression point can be worked out and it is different in each case.

The nonlinear response of the device also depends on the termination impedances at all harmonic and intermodulation frequencies. A complex excitation will generate a much broader spectrum of response, and therefore the effects of these impedance terminations can be very different from a simple CW tone. This is true even independent of true "memory" effects at the device level.

Memory can make the results even more different. The low-frequency beat of a two-tone excitation at sufficiently small tone spacing can effectively modulate the bias of the device, which causes excess distortion. This effect is also present in complex signals (there is modulation near DC which modulates bias). Modulating the bias in response to a complex signal also modulates the temperature of the device, which makes its own contribution to the overall distortion. A CW tone, by contrast, excites the device faster than the local temperature can be modulated, so there is no information in this response to predict the temperature contribution, which is present in a more complex (realistic) signal. (There is still some temperature effect on the device at CW, since the DC power is being converted to RF Radio Frequency: A generic term for radio-based technologies, operating between the Low Frequency range (30k Hz) and the Extra High Frequency range (300 GHz). and dissipated outside the device. That is, a class A amplifier cools when hit by more RF power).

The statistics of the complex signal determine the probability and size of the peaks in amplitude over time. They may be large but may occur with small probability. The peaks can create much distortion, raise the temperature of the device, which then relaxes slowly but influences the waveform for a period related to the thermal relaxation times. These times involve the "time constants" related to the device, the package, and the thermal coupling to other heat sources and sinks. These time scales can exist over a very broad range, overlapping much of the spectrum of the complex signal. However, the CW high-frequency signal will be generally above all thermal time constants, hence will not excite these degrees of freedom that will be excited by the actual signal.

Complex waveforms can also affect the way in which a device "breaks down," because breakdown is history and thermally dependent. Breakdown (even soft breakdown) can be an additional source of distortion.

Finally, charge-trapping phenomena can significantly affect devices (notably GaAs FETs). A complex signal will have slow components that modulate the trap states, causing excess distortion and long history dependent currents (such as "gate-lag") from the device. That is, complex excitations couple the slow trap dynamics to the fast intrinsic device dynamics. Under CW conditions, however, the traps are essentially de-coupled (in time) from the large RF CW signal.

Measurement Overview

A measurement is taken first of the Stimulus (input) signal and then the Response (output) signal. Although these two measurements can be synchronized using an external trigger signal between the signal generator and the measurement hardware, precise frequency and time alignment is needed. This alignment is accomplished in the DSP code of the 89600 VSA software and is configured with the Compensation parameters.

If you have a multi-channel instrument or two instruments, you can measure the stimulus and response signals simultaneously.

The steps in the alignment include:

Amplitude normalization so that the absolute gain can be measured
Alignment of the time samples of the input/output time samples so a mag/phase comparison at every instant of time can be accomplished
Precise frequency lock in case there is a slight error in the generation of the stimulus frequency

Once the Stimulus and Response waveforms are precisely aligned, they can be saved and used in Multicarrier Power Amplifier Predistortion R&D and manufacturing processes.

For a step-by-step procedure for setting up a Complex Stimulus-Response Measurement, see the Making Measurements section of this documentation.

Measurement Results

Measurements on the aligned waveforms provide the following results:

Graphs of AM/AM, AM/PM, and Gain Compression data
Polynomial curve fit to the AM/AM, AM/PM, and Gain Compression data
Coefficients of the polynomial curve (order 3 to 12)
Composite signal EVM, a measure of the average magnitude and phase error of the output with the stimulus signal as a reference
Gain and Delay