ATX PSU and motherboard

AT one time many mother boards were failing because the capacitors on them were defective. Look at some of the big ones tops, especially near th eprocessor and those coils of wire. If they are sort of dome shaped instead of flat, that would be the place to start.

Sometimes you could heat the board and it would start and then run for a long time. Cut it off and you had to heat it again to get it to start.

I sent a couple of boards back for replacement due to this problem years ago.

There's a lot that can be done as far as diagnosis, but learning to diagnose computer problems is ... better left to folk who do it fulltime.

Power supply problems are testable without having a new power supply. Motherboard problems are testable and detectable, but rarely fixable (the bulging-capacitors issue of 2000-2010 vintage being a notable exception).

Your best plan, if you suspect a hardware problem, is to FIRST rule out software (and firmware) issues. Then, visually inspect (for loose or dirty connectors, burned-up components, stalled fans, etc.). Then run test software for common issues (memory, hard disk). At this point, though, you're left with having to repace major parts (like a motherboard or CPU), which you already know.

Inspecting the electrolytic capacitors on both items is the place to start.

An ESR meter would take the guesswork out of it, but the good ones aren't cheap. The most usual sign is bulged tops on the capacitor cans, but they can also appear slightly sunken - if they're running hot enough to build up a head of steam, they can lose electrolyte through the end seal - when they cool, a vacuum forms inside. Since dodgy electrolytics tend to get hot, the plastic covering sleeve can deteriorate giving the capacitor a tired shabby appearance.

If an ATX PSU is reluctant to start; keep trying is a bad idea - they can suddenly burst into life without any regulation!

can

Thanks - wow, concentrating on the PSU stand-alone and not plugged into the m/b seems like a Good Idea!

Mike.

You need a dummy load for that.

Only the 5V & 3.3V rails are regulated, once they come up to spec, the PWM throttles back to idle - the other rails will collapse if you try to draw current.

My dummy load uses both filaments of a H4 headlamp bulb for the 5 & 3.3 rails, an indicator bulb for the +12V and dash bulbs for the rest.

That loads it well enough to scope all the rails for ripple.

There's usually a very small electrolytic close to the SMPSU chip that samples the error voltage - if that's one of the expired, the PSU can go bang.