[hsnam]

hi all,
I have recently built a +/-16V power supply to use with Kevin Gilmore's discrete dynamic headphone amp. It utilizes LM317 and 337 to regulate down from a 24VDC, ~0.8 A supply. When the volume is low, the amp sounds fine, but if i play it too loudly, then the power supply seems to shut off. i can only get the amp to work if i power cycle the supply.
the amp should draw no more than 200-300 mA, if I remember KG's comments correctly.
so what's the problem? LM317/337 are heatsinked, but perhaps i need a larger heat sink? how do i test?

thanks!

once i have time, i shall design a PCB for KG's ultrareg PS and build one...

 

[aos]

My KG amp doesn't draw more than that either and while I have two regulator stages, the first one is using LM regulators with fairly small heatsinks. They get very hot (too hot to touch) but still work.

Are your regulators too hot to touch? I have another version of PS that has HUGE heatsinks and is driving a class A mosfet stage equivalent of a headphone amp and heatsinks and chips get only lukewarm. I think thermal shutdown kicks in only close to 100C (as the maximum junction temperature in silicon is usually 125C). If that's not the case, and your amp doesn't draw much current, then your 24/0.8V supply cannot deliver the juice.

 

[kevin gilmore]

Its possible you have a bad transistor somewhere in
one of the amplifiers. Put a 1 ohm resistor in series
with each power supply voltage. The next time it locks
up, measure the voltage across these resistors to determine
if its a bad amp, or a bad power supply. As the thing runs
class A 99% of the time unless its clipping very very hard, the
power supply current should not fluctuate.