ATX PSU and motherboard

[MJC]

I don't remember seeing a question here about ATX PSUs and motherboards.
So if there is a better forum to ask my question no doubt someone will
tell me where...

My previous computer became progressively more reluctant to boot up and
eventually I gave up and bought a new ATX PSU in case that was the
problem. That was not a cure so I removed the motherboard as well and
bought a new one. That combination has been working ever since, with
enough problems to be interesting, of course...

But I would like to know if anything can be done with the old PSU and
motherboard combination. If I power up the PSU alone by shorting the
green PS_ON connection to ground the fan turns and voltages appear
briefly, but then it goes dead again.
If I connect the PSU to motherboard and short the front panel "on"
contact the same thing happens. The CPU fan turns as well as the PSU
one, but only momentarily.

I took the cover off the PSU but the inside shows nothing untoward.

Is there anything that I can do for some more diagnosis? The
motherboards is a GA-8ISXT-FS and the PSU FSP300-60ATV with a 20-pin
main power connector.

[Ralph Mowery]

In article <MPG.33194c...@news.plus.net>,
gra...@mjcoon.plus.com says...

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.

[whit3rd]

On Thursday, February 23, 2017 at 11:45:21 AM UTC-8, MJC wrote:

> My previous computer became progressively more reluctant to boot up and
> eventually I gave up and bought a new ATX PSU in case that was the

> problem. That was not a cure so I removed the motherboard as well and ...

> Is there anything that I can do for some more diagnosis?

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.

[Benderthe.evilrobot]

"MJC" <gra...@mjcoon.plus.com> wrote in message
news:MPG.33194c...@news.plus.net...

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!

[MJC]

In article <Jt1sA.573666$he2....@fx22.am4>,
Benderthe...@virginmedia.com says...
> ...

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

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

Mike.

[Benderthe.evilrobot]

"MJC" <gra...@mjcoon.plus.com> wrote in message

news:MPG.331ac1...@news.plus.net...

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.