Crossover makes amp go into protection mode

[EarlK]

Unfair.
While you had one approach helping, I had another.
1. I reverse engineered the filter based on the PCB layout and created a scematic. How is that not helpful?
2. I presented the hypothese that the 500kHz switching residue from the class D amp might be the problem and suggested a fix for this with the 2uH inductor + 2 ohm resistor. (It even worked). How is that not helpful?

You have indeed been extremely helpful Armand.

Now the task at hand ( IMHO ) could be to determine if the inductor ( any inductor ) is absolutely mandatory for stability ( or is it simply a design preference derived from "Best-Practice" techniques espoused by EE text-books ? ) .

 

[Defo]

You have indeed been extremely helpful Armand.

Now the task at hand ( IMHO ) could be to determine if the inductor ( any inductor ) is absolutely mandatory for stability ( or is it simply a design preference derived from "Best-Practice" techniques espoused by EE text-books ? ) .

I'm currently running with a 1 ohm resistor in series with the cap.
The amp seems to be stable.

 

[EarlK]

I'm currently running with a 1 ohm resistor in series with the cap.
The amp seems to be stable.

Thanks for that update!

 

[Armand]

Thanks.
The inductor solution is probably not neccesary for stabilty. I would not be surprised if 0.5 ohm in series cures the problem. Such a small resistor will not alter the frequency response either.
The inductor solution will give a much higher impedance over 100kHz and can potentially help amplifiers with poor output filters. The series resistor is however still needed to avoid a resonance peak at 112kHz so this is not a solution without its drawbacks either.

 

[EarlK]

Thanks.
The inductor solution is probably not neccesary for stabilty. I would not be surprised if 0.5 ohm in series cures the problem. Such a small resistor will not alter the frequency response either.
The inductor solution will give a much higher impedance over 100kHz and can potentially help amplifiers with poor output filters. The series resistor is however still needed to avoid a resonance peak at 112kHz so this is not a solution without its drawbacks either.

Thank-you for all your thoughts Armand.

 

[Snickers-is]

The reason why the inductor was considered in the first place is because when you put a series resistor together with a capacitor, it changes the Q dramatically. An inductor could, in theory, have extremely low impedance at the XO-point, but relatively high impedance outside the audible range, where it is needed in order to break down the resonance point of the two capacitors.

The difficulty of this is, off course, that while the two capacitors have a resonance point where the impedance is the lowest, and phase angles near this point tend to be quite aggressive (making amplifier selection a rather random exercise), it is also very low in a very broad frequency range from below 100kHz and upwards.

This is a bit like leaving the EMI-components out from a high BW signal circuit. It might be handling a lot of unwanted signals, and the behavior is not always that predictable either. It could sound just fine, but when you remove the unwanted contribution that adds to the work load of the circuit, you might also hear that this was actually rather important.