Using the EVM to Predict BER Performance of QPSK Modulators with Phase Noise and Timing Jitter Impairments

Abstract

The use of vector signal analyzers for measuring digital signal impairments is standard in the wireless business and is being adopted for various communication links in the satellite business. Boeing Space & Intelligence Systems introduced a comprehensive test and measurement process for NASA Geostationary Operational Environmental satellites using this instrument with support and co-development with the Government program office. It is fast and accurate and has many diagnostic advantages. But using the instrument’s primary measure, the error vector magnitude to correctly predict the bit error rate performance is difficult because many different effects are all included in the one measure. Modifications of a digital modulator probability of error equation are used which capture the effects of phase noise and jitter. Parameter values of the new equation are determined for a specific modulator. The equation for the QPSK modulator is in this paper. The actual probability of error performance results from two modulators are measured in the laboratory using an error counting method. Then the modulators are measured using the vector signal analyzer. Also the phase noise, phase imbalance, and amplitude imbalance are measured using a vector network analyzer. A Monte Carlo simulation of the modulators and demodulators (end-to-end) with measured phase noise and jitter impairments was developed. This allows comparisons of actual, simulated, and predicted performance. The results in this paper show that the accuracy of predicting BER from the EVM measurements is within 0.1 dB of the actual un-encoded measurements and the simulations, at a Pe on the order of 10 -8