Very simple bridged amplifier 170W/4ohm

[indra1]

Some measurements in time domain...

Very nice, enjoy.

 

[PMA]

Thank you . It is not bad.

 

[mjona]

Attached is the simulated result with CV list 5nF 10nF 100nF 220nF. Real measurements are in a good conformance, but sorry I am not going to repeat and save every measurement to bring an evidence.

Test with big capacitances >= 1uF is impractical, such situation will never happen with a standard electro-dynamic speaker. Such test also stresses output devices, even the rugged ones like those 21194/3 used. And it does not speak about amp stability, rather about L/C/R ratios.

In my experience, it is best to make conclusions on real circuits, real amplifiers and real measurements. Qualified predictions and guesses may be helpful, but also misleading, as something may be easily not taken into account. So I am very careful when making assumptions about circuits I have not built and tested. And then, every realization is unique.

The clearance between the peak output voltage and the supply rails voltages is sufficient to allow a decent sine wave however the peak current demand for this is momentary in comparison for the that of a square wave over a period of 50uS.

If the required output (over 20V peak) is too close to the supply rails it should come as no surprise there is some ripple in the waveform on square wave although the absence of any overshoot looks to be a good sign.

With parallel capacitors in a load the charge/discharge properties/rates of these in response to a step function, and an amplifier on the cusp of saturation, one could perceive this resulting in a negative feedback signal that is intermittent in form.

If you reduce the input signal the ripple should disappear and you could try increasing the parallel capacitor keeping the values below 1uF.