With only an AP to measure, and without watching for both common mode and differential mode signals (or inverting and noninverting channels separated), especially using it with many ms scale, you are practically blindfolded. Oscilloscope is required for ClassD development.
Why don't you use post-filter feedback ? It does of course have the disadvantage that you will introduce more instability if not properly done. OTOH you have some control on output filter distortion (linear and non-linear) and the output filter might also function as anti-aliasing filter for the ADC , at least to some degree.
You said that you simulated the DAC on "transistor level". Did that also include the decimation filter ? The latency of such a filter might be detrimental to the stability of a feedback loop.
And yes like Pafi mentioned one has to take care that the on-time of each output state of the switching stage does not fall below a certain value. Or latch up and going into protection might happen.
Regards
Charles
You said that you simulated the DAC on "transistor level". Did that also include the decimation filter ? The latency of such a filter might be detrimental to the stability of a feedback loop.
And yes like Pafi mentioned one has to take care that the on-time of each output state of the switching stage does not fall below a certain value. Or latch up and going into protection might happen.
Regards
Charles
- Status
- Not open for further replies.