Reply to thread

I don't understand what you mean with "too slow". Loop gain measurement with injected signal should be done in steady state. One problem is selection of injected signal level, it must not exceed linear range of the modulator. Other problem is suitable filtering of response signal out of pwm noise, your results suggest that you are failing at this point. There are commercial instruments like Ridley's AP300 performing similar analysis, in so far there's no doubt that it can work.