I have tried with source resistors of 0.22 Ohms in my setup today and the thermal runaway with the verticals was gone.
The BJT for bias compensation was still in place, no overcompensation at all, @AKSA. In this setup, the bias stayed at 130mA as I had set it, without signal and after running 500mV input for a while and then letting it rest again.
My board was not planned with room for source resistors, so I cut the source pin and soldered the resistor in for the test.
The BJT for bias compensation was still in place, no overcompensation at all, @AKSA. In this setup, the bias stayed at 130mA as I had set it, without signal and after running 500mV input for a while and then letting it rest again.
My board was not planned with room for source resistors, so I cut the source pin and soldered the resistor in for the test.
Of course not 😀. The above pic is just to show how I added the resistors. Afterwards I mounted the board on the heatsink again.
I never got any clear answers to why Pass F4 can use TL431 for bias control, while this amp supposedly can't. Is it because the F4 runs on a highe quiescent current? If thats the case, I guess I can do the same with my amp?
Pass F4:
Ultra HEXFET:
Time will show, because I am (very) slowly progressing with my project:
🙂 morten
Pass F4:
Ultra HEXFET:
Time will show, because I am (very) slowly progressing with my project:
🙂 morten
TL431 could be noisy if used in the front stage.
The Vbe multiplier position is at the output stage. It seems to be fine for me. As it is not possible to thermal couple with TL431, you need a big heat sink.
You are limited by the voltage rating of the JFET, paralleling 2 mosfets is overkill.