System abruptly powers down -- an update

Here's an update for those interested. It looks like I have a PSU
problem.

I opened up the box and turned it on and just left it sitting in the
BIOS. It was on for maybe 8-10 min, which is considerably longer than
it took to halt when it tried to run Windows. The CPU temp was around
33-35 and the mobo around 28-30, according to the BIOS temp monitor,
and the temps climbed slowly but not dramatically over the time it was
on. All the fans were spinning -- case, CPU and PSU -- and it was
somewhat dusty inside but not terribly.

But I kept thinking I smelled something burning. And the smell kept
getting stronger, but I couldn't figure out where it was coming from.
Then I sniffed around the PSU exhaust and that was it. A definite
burning electrical smell.

It still hadn't shut down, but at that point I turned it off. I
hadn't tried to boot so I don't know what would have happened if I
did.

(My guess about why it initially failed is that maybe the voltage to
the HDD wasn't right and the drive wasn't reading properly and that's
why Windows never came up fully, though I still don't know why it
would have shut down.)

But in any event, there's certainly a problem in the PSU, so I'll
replace it. I just hope the failure didn't send the voltages so out
of spec that something else was damaged.

Thanks to all who replied.

John wrote:
Here's an update for those interested. It looks like I have a PSU
problem.

I opened up the box and turned it on and just left it sitting in the
BIOS. It was on for maybe 8-10 min, which is considerably longer than
it took to halt when it tried to run Windows. The CPU temp was around
33-35 and the mobo around 28-30, according to the BIOS temp monitor,
and the temps climbed slowly but not dramatically over the time it was
on. All the fans were spinning -- case, CPU and PSU -- and it was
somewhat dusty inside but not terribly.

But I kept thinking I smelled something burning. And the smell kept
getting stronger, but I couldn't figure out where it was coming from.
Then I sniffed around the PSU exhaust and that was it. A definite
burning electrical smell.

It still hadn't shut down, but at that point I turned it off. I
hadn't tried to boot so I don't know what would have happened if I
did.

(My guess about why it initially failed is that maybe the voltage to
the HDD wasn't right and the drive wasn't reading properly and that's
why Windows never came up fully, though I still don't know why it
would have shut down.)

But in any event, there's certainly a problem in the PSU, so I'll
replace it. I just hope the failure didn't send the voltages so out
of spec that something else was damaged.

Thanks to all who replied.


If the PSU is out of warranty, you can remove the screws on the
cover, and have a look. Don't touch anything in there. It could be
that one or more capacitors is leaking. I actually got a little
puff of smoke one day from mine, but it continued to run. And
the root cause there was bad caps.

(The leaking is the orange-brown stuff, shown on four caps here.)
http://en.wikipedia.org/wiki/File:PSU_Caps.jpg

I replaced mine, before it failed completely. It was an Antec True480.
It did not have a lot of service hours on it, so it wasn't the
stress of operation that killed it. The capacitors failed due to
their internal chemistry.

Paul

On Fri, 22 May 2009 13:17:50 -0400, Paul wrote:

John wrote:
Here's an update for those interested. It looks like I have a PSU
problem.

I opened up the box and turned it on and just left it sitting in the
BIOS. It was on for maybe 8-10 min, which is considerably longer than
it took to halt when it tried to run Windows. The CPU temp was around
33-35 and the mobo around 28-30, according to the BIOS temp monitor,
and the temps climbed slowly but not dramatically over the time it was
on. All the fans were spinning -- case, CPU and PSU -- and it was
somewhat dusty inside but not terribly.

But I kept thinking I smelled something burning. And the smell kept
getting stronger, but I couldn't figure out where it was coming from.
Then I sniffed around the PSU exhaust and that was it. A definite
burning electrical smell.

It still hadn't shut down, but at that point I turned it off. I
hadn't tried to boot so I don't know what would have happened if I
did.

(My guess about why it initially failed is that maybe the voltage to
the HDD wasn't right and the drive wasn't reading properly and that's
why Windows never came up fully, though I still don't know why it
would have shut down.)

But in any event, there's certainly a problem in the PSU, so I'll
replace it. I just hope the failure didn't send the voltages so out
of spec that something else was damaged.

Thanks to all who replied.


If the PSU is out of warranty, you can remove the screws on the
cover, and have a look. Don't touch anything in there. It could be
that one or more capacitors is leaking. I actually got a little
puff of smoke one day from mine, but it continued to run. And
the root cause there was bad caps.

(The leaking is the orange-brown stuff, shown on four caps here.)
http://en.wikipedia.org/wiki/File:PSU_Caps.jpg

I replaced mine, before it failed completely. It was an Antec True480.
It did not have a lot of service hours on it, so it wasn't the
stress of operation that killed it. The capacitors failed due to
their internal chemistry.

Paul

The PSU was still under warranty, so I didn't dismantle it. Got an
RMA from Thermaltake and sent it back.

But I decided I didn't want to wait for however long it might take for
them to turn it around and in the meanwhile ordered a replacement. I
installed it yesterday and it looks like all is well. (The repaired
one, when I get it back, will be a spare like the dozens of other odds
and ends of computer stuff I've got lying around :-))

To my relief, it looks like whatever badness happened to the supply
didn't damage anything else. The systems's been running fine (knock
wood) for about 28 hours now. I'm thinking that it wasn't a thermal
shutdown of the processor after all but the PSU's own overvoltage
protection or something of the like kicking in and shutting the supply
itself off. (Maybe one voltage, or one rail, at a time, so that the
video card was the first to go? I don't really know how a supply
reacts when it senses a failure -- does it shut everything down, or
only the voltage that it senses is bad, or what?)

In any event, "all's well that ends well." Thanks (to you and others)
for the replies and suggestions.

Reply-to address is real
John