You can use the vector averaging functionality in the OFDMModAcc measurement when testing devices like power amplifiers and front-end modules in a test setup that uses an NI signal generator for generating an RF stimulus signal for the device-under-test (DUT) and an NI signal analyzer to perform measurements on the RF output signal of the DUT. You can use vector averaging to reduce the impact of uncorrelated noise on the measurement, originating from anywhere in the entire signal path.

Vector averaging averages the acquired I/Q data across averaging counts after aligning the data in time and phase. I/Q data for each averaging count is obtained by processing I/Q data in a record in a multi-record acquisition. Correction of relative time offsets between records may cause the number of samples available for the OFDMModAcc measurement to be lesser than the number of samples in a record, depending on the absolute value of the maximum time offset between records. Vector averaging assumes that all records are aligned in center frequency and sampling frequency.

Therefore, center frequency alignment and sampling frequency alignment are not performed on acquired records. You must make sure that the frequency reference is locked between the signal generator and the signal analyzer. Assume that si[l] denotes acquired I/Q samples of record i for 0 ≤ l < L, where L is the number of samples in a record, with 0 ≤ i < N, where N is the averaging count. Let s'i[m] denote I/Q samples of averaging count i after correcting relative time and phase offsets of si[l], with 0 ≤ m < M ≤ L, where M is the number of samples available for averaging across all averaging counts after alignment. The output samples after vector averaging are given by

The OFDMModAcc measurement is performed on the averaged I/Q data s[m].

Vector averaging reduces uncorrelated noise in the signal. In the vector averaged signal s[m], the power ratio of the signal of interest and the uncorrelated noise is increased by 10log(N) dB, as compared to the same power ratio in each of the acquired I/Q data si[l]. However, after vector averaging, the improvement in the overall signal-to-noise ratio and thus the EVM may be limited by correlated noise components such as non-linearities. To determine the averaging count, you must check the amount of EVM improvement for your device-under-test and test system for different averaging counts.

To perform vector averaging, you must set the OFDMModAcc Averaging Type property to Vector. You can control time and phase alignment operations through the OFDMModAcc Vector Averaging Time Alignment Enabled and OFDMModAcc Vector Averaging Phase Alignment Enabled properties, respectively. You can disable time alignment if your test setup uses a precise and repetitive triggering mechanism where all acquired records are triggered at the same relative position within the signal. You can disable phase alignment if you are sharing the LO between the NI signal generator and the NI signal analyzer. Disabling time and/or phase alignment may improve the measurement time.

In contrast to vector averaging, RMS averaging performs OFDMModAcc measurement on each averaging count and then averages the measurement results. RMS averaging can improve the standard deviation of the measurement results such as EVM, whereas vector averaging can improve the mean value of the measurement results.