I have a question... Has anybody measured the current of CLASS-D prior to LC filter directly at the PWM output? Has anybody seen any real advantage of measuring the speaker current in CLASS-D before the LC filter vs after the LC filter.
After the LC filter the common mode is not as bad as after the LC filter.
After the LC filter the common mode is not as bad as after the LC filter.
It depends on what is going to be done with the current monitor signal.
For example, for cycle-by-cycle peak limiting purposes, sensing before the capacitor includes capacitor current to ensure the limit is not exceeded at high frequencies. But the cost effective evolution of this is sensing before the LC filter, across FET Rds-on, because "floating" sensing uses more parts (or requires "floating" power supply). An exampe of this implementation can be seen in IRS20957 FET driver IC datasheet (and there is an app-note too), although datasheets from this manufacturer are not particularly easy to understand.
For slower gain-reduction based current limiting, or speaker monitoring purposes, sensing after output capacitor tends to be easier. Sensing at speaker ground lead even easier, for single ended applications. Bridging requires some form of floating sense circuit and analog level shifting.
Another option that I have seen is an auxiliary winding in output inductor feeding an integrator, to reconstruct inductor current waveform, but to make this circuit accurate the integrator needs variable compensation depending on output inductor winding resistance, and copper resistance almost doubles from ambient temperature to approx. 100C, so implementation becomes complex and not very accurate.
For example, for cycle-by-cycle peak limiting purposes, sensing before the capacitor includes capacitor current to ensure the limit is not exceeded at high frequencies. But the cost effective evolution of this is sensing before the LC filter, across FET Rds-on, because "floating" sensing uses more parts (or requires "floating" power supply). An exampe of this implementation can be seen in IRS20957 FET driver IC datasheet (and there is an app-note too), although datasheets from this manufacturer are not particularly easy to understand.
For slower gain-reduction based current limiting, or speaker monitoring purposes, sensing after output capacitor tends to be easier. Sensing at speaker ground lead even easier, for single ended applications. Bridging requires some form of floating sense circuit and analog level shifting.
Another option that I have seen is an auxiliary winding in output inductor feeding an integrator, to reconstruct inductor current waveform, but to make this circuit accurate the integrator needs variable compensation depending on output inductor winding resistance, and copper resistance almost doubles from ambient temperature to approx. 100C, so implementation becomes complex and not very accurate.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.