OT: system shuts off sporadically

[Craig White]

On Mon, 2006-01-16 at 12:35 -0700, Vaughn Treude wrote:
> Hello everyone:
> I've been experiencing a problem with one of my PC's.  Eight months ago
> I upgraded it to an Abit VIA KM400A motherboard with an AMD processor
> and new RAM.  In the last couple of weeks it's started shutting itself
> off sporadically, for no apparent or consistent reason.  I know it's not
> a software problem (though if anyone wonders, I've got Ubuntu on it, and
> have been quite happy with that flavor of Linux) because it has at times
> shut itself off while I'm in the BIOS setup screen.  I have replaced the
> power supply and that didn't help.  I suppose it's probably a
> motherboard problem, and I believe it's still under warranty, BUT
> sending it back will be a hassle and I'll lose the use of that machine
> for a few weeks at least. Most odiously, I'll have to reinstall the
> processor and mess with that heat sink junk once again. So I'm wondering
> if anybody knows of anything else that could cause this problem.
> 
> Some more info:
> - The problem appears to be totally random, sometimes happening within a
> minute of startup, sometimes hours later.
> - Sometimes the system will not restart for a few minutes after this
> happens.  Or sometimes it will start but not boot.
> - Often when this happens the BIOS checksum gets corrupted and I have to
> re-enter my settings.
> - In BIOS, I've taken care to turn off features I'm not using, such as
> APIC and SATA support. I've turned off power management to the extent
> the BIOS allowed me to do that. I've also set the power-off switch mode
> to a 4 second delay, in case there could be noise spikes on the power
> switch line.  I've turned off the "halt on errors" feature.
> - In the BIOS PC health screen the CPU temperature appears to be well
> within the good range.  However, I've only checked this shortly after
> power-on.
> - The motherboard is rather compact, and is mounted in a maxi-tower case
> with extra fans.  I don't know for a fact if the airflow is optimal.
> - I've checked the system logs and there is absolutely no record of any
> kind of error at the time it powers itself off.
> 
> Any ideas of what I might try, or any resources I could use?  I googled
> for it under "PC sporadic shut off" but didn't get any useful hits. 
> Half of the problem with googling for a PC problem is knowing what to
> call it!
> 
> Thanks in advance for any suggestions you might have,
----
I have had that happen from 3 different causes that I can recall.

The first was a failing power supply which apparently you have already
replaced.

The second was with a bad cpu fan where the temperature of the cpu would
cause it to shut down. You should be able to monitor the temperatures
but acpi would have to be turned on and with Linux, you would probably
have to install and configure lmsensors and probably something like
gkrellm.

The third was most recently when a Dell system had capacitors that were
leaking - you could see the gook on the top of the capacitors when you
looked at the motherboard. This is most likely the issue and I would
recommend that you look carefully at the capacitors on the motherboard
and if there is anything other than dust on the top of some of them,
replace the motherboard as there is no other option.

Also - as a method of control, if the computer is plugged in to a wall
outlet, I would plug a plug strip into the wall outlet and plug the
computer into the strip along with something else that would turn off
(but not back on) if for some reason the power dropped momentarily.

Craig