PC Hardware → Re: SDS (Sudden Death Syndrome)

Do you have a power supply tester? I have one that displays the various voltages so if you powered it up for a while you'd be able to see the voltages going wonky UNLESS for some strange reason it only showed up under load.

I would suspect one of the rails is slowly losing voltage until the PC goes 'oopsie' and reboots.

I only found my sleep problem by backprobing the connector on the mobo and seeing the 5v rail slowly lose voltage after I put the PC to sleep.

Have a Fluke and at least half a dozen HF multimeters

Anyhow, on a Molex I measured 5.04V and 11.98V, both steady within +/- 0.01V.

The PS is a bit dusty, but based on a cursory look, I can't see anything obviously wrong with it.

I think I might want to probe the power good signal. And I wonder if there is an ATX interposer that would give easy access to all power lines.

I'll upload the pictures later.

I have a Fluke too, but it's a clamp-type I inherited that's only so useful.

You'll have to check more than a molex; there's many rails of power to check. You might've only checked 10% of the rails on the power supply you have.

Presumably when my PSU was bad, there was a leaking capacitor... I inspected the boards carefully but noticed no visual leaks. It must've been leaking at the bottom where I couldn't see it.

Unlike the one computer I had. That one started acting strange, then I heard what sounded like popcorn popping. Shut it off, opened the cover and it was very obvious capacitors were missing, popped right off the board...

you want 10mV ripple or less on the lines, using an Oscilloscope.. that's the only instrument to use.. a voltmeter can be fooled.. and the frequency(60KHZ) is way to high for a voltmeter anyway..

Since the PC usually boots when it's initially turned, high frequency noise on the power lines is not likely to be the root cause of what I've been seeing. The measurements were done to ensure that there are no gross defects with either +5V or +12V rails.

I need to take it apart for closer visual inspection and testing.I have a capacitor tester that to some degree can test caps in-circuit. I don't like to blindly replace parts unless they are shown to be bad, or all other avenues are exhausted.

I just need some spare time to do it.