PS Diagnosis?

[sam9]

Description of PS:
A standard unregulated power supply ~42-0-42 VDC under load. Uses 25A 600V bridge (with subbers), four Nichicon 10,000uF filter caps, followed by a pair of .1uF film caps, and a pair of 4.7kR bleed resistors. Transformer is A-L 35-0-35 500VA.

DC voltage measurements show everything to be just as desired and expected, no off set everthing steady whether under constant load or no load. BUT ( AC while voltage measured at the posative DC rail show nothing except perhaps a mV of ripple, at the negative rail I read a pulsing AC reading - the DMM cycles between 1-2Vac to 23Vac constantly. I suspect a bad filter cap that won't hold a charge, hence the rapid pulsing. But before I tear everything apart and start replacing caps can anyone confirm or refute my diagnosis.

Oh yes I have disconected the PS from the amp boards to isolate the problem. I did not check the AC at the rails before installing and connecting the amp boards. I actually discovered this by tracing back why the MOSFETs in the amp were getting too hot. The thermal switch on the back of one opened which was the first clue something was amiss.

[Enzo]

Your output devices won't overheat because the power rails are hummy. You have disconnected the amp circuits, but is there anything else riding on those rails, no matter how seemingly innocent? Disconnect EVERYTHING from the power supply, even the LED power indicator.

SCope it to see what the ripple looks like. Beyond that, either the cap is leaky on that rail, or the rail is being loaded down. Isolate the problem. Swap the filter caps and see if the problem moves with them. If so, you have a bad one. If the caps are OK, then go back to basics. Remove the bleeder and film cap from the bad side. Then pull the snubbers off the bridge. If the bleeder is mismarked or if you inadvertently put in a wrong value or if the film cap is bad, there could be a load on the rail not present on the good side. Likewise a leaky snubber could be passing some AC on through past the bridge.

And check the bridge. If one leg is open, then you essentially have half wave and it will have more ripple.

I doubt it is the AC mains, that would affect both rails. Well, unless the power mains waveform was grossly assymetrical.

[sam9]

Thanks, you are confirming my approach but have a couple of additional suggestions.

One help I just remembered is that I built two identical power supplies, the second one for a follow on project. I've swaped them and No. 2 shows no problems so far. Once I've completly checked out No.2, I'll reconnect the amp PCBs. Then if all else is OK, I can come back to PS No.1 and figure out the problem at my leisure.

[AndrewT]

Hi,
your DC voltage is very low for a 35vac 500va transformer.
I would expect +-48 to 50.5v on the rails when loaded. How big is your load? When you have completed the caps checks, then check you have the same plus & minus Vrail voltages.
regards Andrew T.