R&S^®ZVA Vector Network Analyzers

Pulsed measurements on amplifiers and mixers

The R&S^®ZVA offers various capabilities for testing amplifiers and mixers under pulsed conditions.

Average pulse measurements

Without any options installed, the R&S^®ZVA offers power and S-parameter as well as conversion loss measurements on amplifiers and mixers versus frequency and power using the average pulse measurement method. This method requires a periodic pulsed signal and operates on the assumption that the measured parameters do not change during the pulse.

The average pulse method has no limitation concerning the minimum width of the pulse and requires only the R&S^®ZVAB16 option in order to generate a pulsed RF signal by means of a pulsed generator or an external pulse modulator.

VNAs using the average pulse method measure only the main carrier of the signal spectrum. Therefore, spectral lines of the pulsed signal close to the main carrier that are not completely suppressed due to the side lobes of the digital IF filter will cause interference in the wanted measurement signal.

The R&S^®ZVA overcomes this problem by offering high-selectivity IF filters that have so far only been used by spectrum analyzers. Complicated and time-consuming techniques such as “spectral nulling” are a thing of the past.

Normal-selectivity and high-selectivity filters of the R&S^®ZVA

Point-in-pulse measurements

The average pulse measurement has some disadvantages, e.g. low S/N for low duty cycles or duty-cycle-dependent results for absolute power measurements.

The average pulse measurement also delivers unsatisfactory results if the behavior of the DUT after settling, overshoot, or ringing of the pulse has to be analyzed.

The point-in-pulse measurement method overcomes these problems because the measurement value is monitored only during the on-phase of the RF burst. It allows accurate power measurements independent of the duty cycle, as well as shifting of the measurement to the desired part of the pulse by selecting a suitable trigger delay.

The point-in-pulse method requires that the data acquisition time of the receivers (i.e. the sampling time) be shorter than the pulse width t_on.

Sampling time and trigger delay for point-in-pulse measurements

With the R&S^®ZVA-K17 option installed, the R&S^®ZVA performs point-in-pulse measurements on pulses with pulse widths down to 430 ns (down to 230 ns with some degradation of performance) using measurement bandwidths of 5 MHz (10 MHz for 230 ns).

Pulse profile measurements

To analyze the time-dependent S-parameters and wave quantities of a DUT during a pulse, the R&S^®ZVA performs pulse profile measurements using the R&S^®ZVA-K7 or R&S^®ZVA-B7 option. Data is recorded versus time at a sampling rate of 80 MHz. The recording time per sweep may vary from 25 ns to 3 ms (R&S^®ZVA-K7 option) or 25 ms (R&S^®ZVA-B7 option). The time resolution is always 12.5 ns, corresponding to the A/D conversion rate. Quasi-continuous display update in the ms range makes it easy to adjust the DUT. IF bandwidths in the range from 100 kHz to 10 MHz (30 MHz with restrictions) are available. For a minimum pulse width consisting of the pulse rise time plus a number of data points sufficient to represent the shape of the flat pulse top, pulses down to a length of 50 ns to 100 ns can be measured.

Unlike existing techniques, this new technique is very easy to configure (only a few parameters have to be set in the dialog mode), requires minimum instrumentation effort, has no loss of dynamic range caused by desensitization, and also features very high measurement speed, since wide IF bandwidths can be used and no averaging is required.
Furthermore, measurements of double pulses and arbitrary pulse trains are possible.

Below is a test setup for pulsed measurements that uses a modulator (switch). Alternatively, a pulse generator can be used.

Test setup for pulsed measurements

Pulse profile measurement on an amplifier

For DUTs with a large group delay, i.e. a group delay longer than the pulse width, it is difficult to measure S21 because the reference stimulus signal a1 may already be off by the time the transmitted pulse b2 reaches the receiver.

Wave quantities a1, b2, and S21 of a DUT with 100 ns group delay

The R&S^®ZVA solves this problem by mathematically shifting the waves quantities by the DUT’s group delay before calculating the S-parameters. Thus, DUTs with a large group delay can be analyzed using even very short pulses.

S21 after shifting b2 by 100 ns