18 replies to this topic

[eniparadoxgma]

What up, y'all.

Well, one of my neighbors has enlisted me to fix their computer.

Here's the computer:

HP desktop
AMD Athlon XP 3200
512mb Ram
160gb HD
crappy video card
etc

Here's the problem:

Computer just up and turned itself off on them. When trying to boot back up and trying everything from safe mode to last known working condition it would just die on them in the middle of booting and turn itself off.

I went over and turned it on, and it booted, went into windows, and was alright for about 10 minutes. Then, it just died. Poof. Couldn't get it restarted.

My initial opinion was that the cpu was overheating, and possibly the cpu fan was dying. After bringing it home, I noticed when trying to boot it that the fan barely spun. When in Windows (when i got it back in windows for awhile) I noticed that the fan would spin real fast sometimes, then real slow, then real fast, etc. I then booted into the BIOS, checked the hardware monitor and noticed that the cpu temperature was about...oh 90C!

I tried plugging the cpu fan into a different fan connector and had the same problem, in my mind eliminating the possibility of a faulty fan connector.

SO, I then surmised that it was a faulty CPU fan and had them buy another. I took the old one off, ISO'd the crusty thermal paste off, applied some AS5, and slapped the new one on there.

No difference in temperature at all. I installed Speedfan to make sure to keep the fan spinning at its maximum RPM (4200) and it still idles at 87C. Think that's bad, when it's under load I saw it up to 117C! These are the readings in the BIOS and speedfan, so I'm assuming it's not a faulty temperature reading.

The computer will seem to be alright for awhile under no load (idling at 87C) but I tried playing some XP Pinball and it shut it down within about 10 minutes.

So, my next two guesses are either the cpu is already screwed or that the power supply is going bad.

Anyhow, sorry for the novel but I was hoping someone might have some insight into this or how I might go about finding out the problem. I don't have a mulitimeter to check the PSU voltages, but one thing the guy told me was that once a few months ago the computer had some kind of problem that was similar and that he "jiggled the power cord" and it seemed to fix it...until now.

Again, any help/advice/suggestions are appreciated.

[m.oreilly]

switch out the psu, and see what happens?
edit: doh!

another edit: did you reset the bios to default?

[eniparadoxgma]

Yeah, I don't have a spare PSU atm....I don't think. I will definitely check around.

Thanks for the heatsink tip, scara. I will give it a try.

Report back in a few, y'all. Thanks for the advice.

[m.oreilly]

i forgot about them non ihs amd jobs...years ago patty's son had one of them, and had a heck of a time getting the thing to seat right, as the rig was always running hot. i wonder also if there could be an issue with the boards voltage regulators...old stuff, indeed...

[eniparadoxgma]

Well, holding down the heatsink a bit didn't change the temp in BIOS.

I don't have a PSU anywhere that I can find, but I have a friend paying a visit tomorrow and he does have a spare to bring.

So, I guess I'll go with PSU until I can test her out. Thanks again, folks.

[eniparadoxgma]

m.oreilly, on Jul 29 2008, 09:06 PM, said:

i forgot about them non ihs amd jobs...years ago patty's son had one of them, and had a heck of a time getting the thing to seat right, as the rig was always running hot. i wonder also if there could be an issue with the boards voltage regulators...old stuff, indeed...

Yeah, I was wondering about seating the heatsink, as it was a complete PITA to accomplish. However, I just think it'd be strange if after 4-5 years all of a sudden the heatsink un-sat(?) itself, as no one has even opened the inside of this dinosaur for that many years. Then, I would have to had made the exact same mistake when I put it on...as the temps haven't changed at all.

[m.oreilly]

at first, i had an image of someone doing an unauthorized OC, then the psu going ape, then all of the above, plus the board regulators dieing off to boot

[Tweak]

Perhaps additionally a flash of the BIOS to the latest final (non-bate) wouldn't hurt either as the temps situation may have always been a partial issue that only now got noticed...the PSU is a definite item to eliminate though.

[m.oreilly]

Tweak, on Jul 29 2008, 07:46 PM, said:

Perhaps additionally a flash of the BIOS to the latest final (non-bate) wouldn't hurt either as the temps situation may have always been a partial issue that only now got noticed...the PSU is a definite item to eliminate though.

hehe, finding a "latest" bios for that rig might be a whole other issue...

[m.oreilly]

so, enip...sup? have you found the answer(s)?

[eniparadoxgma]

My friend forgot to bring the PSU up yesterday, so I haven't had a chance to do anything with it yet.

As far as flashing goes, I think it's a bad idea being as that the main symptom of the current issue is turning itself off without warning. I don't want to turn this thing into a paperweight.

Dude's supposed to bring the PSU tomorrow, so hopefully I'll be able to figure out if it's the power supply tomorrow.

[Tweak]

eniparadoxgma, on Jul 31 2008, 03:23 PM, said:

My friend forgot to bring the PSU up yesterday, so I haven't had a chance to do anything with it yet.

As far as flashing goes, I think it's a bad idea being as that the main symptom of the current issue is turning itself off without warning. I don't want to turn this thing into a paperweight.

Dude's supposed to bring the PSU tomorrow, so hopefully I'll be able to figure out if it's the power supply tomorrow.

The BIOS flash was a suggestion for after the primary problem was solved.

[eniparadoxgma]

Switching the power supply out seemed to have positive results, as the temperature at idle dropped from 87C to 60C.

However, running StressPrime caused it to freeze within about 15 seconds with a temperature of about 125C.

I'm about to tell them it's just screwed (aka motherboard or cpu is fried).

Anyone have any better ideas?

[m.oreilly]

well, trying some sort of diagnostic app to see what the voltages are now (the mobo could have power issues, causing higher temps because of unstable voltages). they should upgrade really, as stuff is so cheap right now (a "basic" system of today is much more powerful and far less likely to have issues...). i just had a similar job, and after going over the pros, they decided to go w/new.