Computer Shutting off for no reason

[scratched]

About a week ago my PC began shutting off with no warning ahead of time. It would stay on a couple of hours, and then it would just croak and turn off. As the week went on, my computer could stay on for progressively shorter amounts of time. I thought that the power supply was the problem, so I bought a Thermaltake 430W power supply off of newegg to replace my old Powmax 450W PS since it had been making noise for a few weeks now and I suspected that was the problem. Unfortunately, the power supply didn't help at all. It still shuts off.

I looked inside of the computer and noticed that the CPU overheat LED turns on whenever the computer shuts off. I know it isn't a heat problem because I went into my BIOS and looked at the temperature and the CPU temperature wasn't above 51 degrees celcius when it shut itself off on me. I also had the overheat shutoff feature disabled so I don't think it is the CPU overheating.

I'm not sure what is the problem. Anyone have any ideas.

If a part went bad I'd like to replace it instead of buying/building a whole new computer.

Specs:
Shuttle AN35-Ultra Mobo
AMD AthlonXP 2500+ "Barton" processor
512MB Kingston Memory
ATI Radeon X700 256MB GPU
Windows XP Home

[peterb]

Well, it isn't shutting off for no reason.. either it is a a real heat problem and the CPU or mobo is overheating and shutting itself down, (and the temperature reading from the bios is inaccurate because of a faulty heat sensor - if it is on on the mobo), or the CPU isn't overheating but the bios thinks it is and shutting the computer down (again an inaccurate heat sensor) Difficult to tell which without putting a separate temp probe on the CPU. Does the mobo have a separate heat sensor? - the CPU usually has one built onto the chip die.

Is it a controlled shutdown or is it just switching off? How old is the mobo? Have you altered anything on it just before the problem started? (knocked the heat sink or something) Are any dust filters clean - is the CPU cooling fan OK? Does it ake any difference if you leave the sides off the case?

[scratched]

The mobo & processor are about 2 years old. It could be possible that the heat sensors are faulty I guess.

I as far as I can tell the heatsink/fan on the processor are attached correctly and the fan is definitely spinning. I also have a fan on the back of the computer sucking hot air away from the processor. I've left one side of the case open and it doesn't seem to help any. When it does shut down it just switches off immediately, it doesn't give me a blue screen or a shut down prompt.

It could be a heat problem, but would the computer overheat within 5 minutes of turning it on if it has been left off overnight?


When I went to replace my power supply the computer had been turned off for the night. When I went to turn it back on after it had been replaced it turned off within just 5 minutes. I just came back from work and now the computer has been running close to 15 minutes without problems.

EDIT: The computer stayed on for about 15 minutes, then I started playing Warcraft and it shut off about 3 minutes into playing. Afterwards I turned the computer back on and went into the BIOS to check the temp. The "System Temperature" was 38 (celcius), and the "Processor Temperature" was 54 and the "PWM Temperature" was 42(This may be a dumb question, but what does PWM mean?) Are those temperatures normal? I've never paid much attention to PC temperature until now...

[Mimp]

This happened to me and it turned out that my motherboard was shorting against my case and causing the instability. I'd missed a screw when i put it in like a fool..
Anyway, try running it outside the case on something non conductive (ie not an antistatic bag) and see if your problems disappear. Actually id wait a little before you try that as it's pretty time consuming and there are lots of other things taht could cause it but worth a try.
Um if your on xp theres a setting somewhere in system properties that you have to uncheck or it will reset every time theres a blue screen, so tinker with that to see if it changes things any..

[herulach]

Its possible its a gpu overheating problem, get some canned air and put that through all the heatsinks.

Also, have you specifically told windows to give you bluescreens at some point in the past? If not you need to go to control panel -> system ->advance-> startup and recovery, then uncheck the button marked automatically restart.

[scratched]

I know it isn't a blue screen of death because a lot of the time it shuts down in the BIOS when I'm looking at the temperature, not while in windows.

[ultimate]

Have you taken out the cpu cleaned the old paste off the heat sink and cpu and renew it with a decent paste like arctic silver etc.?

I renew mine once a year and by doing this I have constant low temps.

[peterb]

My (AMD) CPU runs at anything between 55 and 65 C without any bother (and has done for about 3years). I did have a problem with an Asus terminator system running Linux, which shut down when I was running MP3 ripping software - that was a PSU overheat problem solved by changing the cooling flow through the PC.

In your case it appears to shut down when you are running processor intensive stuff - so some of the suggestions such as cleaning the coolers would be worthwhile. I'm not sure about changing the heat sync paste - if you aren't overclocking, and you are using mfr stock parts with the mfr fitted paste, and you haven't disturbed the heatink, then I'd leave well alone.

[scratched]

It hasn't just been shutting down doing processor intensive things. It also has shut down when all I'm doing is running windows and nothing more. I don't have a spare mobo or processor to swap out to check parts with. I might try the video card since I have a couple of those but I can't test anything else without buying new parts which I only want to do if it is the exact part I need...

Everything for the processor is stock (fan, heatsink, etc.) and it's been running fine for the last 2 years untouched. I don't think it has been disturbed, although I can't know for sure since the computer is mostly used by the rest of my family. I usually only use it when my family wants me to fix it and they won't tell me the truth if they bumped anything so I can't be sure if that is the problem. I doubt they bumped anything internal though. If anything was bumped it would have been on the outside of the computer, nothing inside...

[peterb]

Yes - realised that - just seems to happen (if I read the post correctly) quicker when you start intensive stuff, indicating that the CPU is getting hotter. I'd start by checking the inside of the case for dust and fluff blocking any air vents or cooling fins.

[Koolpc]

Sounds like a temperature problem to me or you could have a faulty PSU?