Improvements in estimating the spectra of random PWM waveforms

Accurate estimates of the power spectral density of random-pulse-width-modulated waveforms (RPWM) are essential in reducing the perceived acoustic emissions from modern motor control drives. It is shown that by refining the estimate of the autocorrelation function of the phase of the RPWM waveform about the origin, the resulting power spectral density is greatly improved, giving excellent agreement with the observed spectrum.