HELP: Is this a blown motherboard?

[blackbeltninja]

Hey guys

Hoping someone can give me some advice here. One of the machines in my office froze this morning. I rebooted it and it started fine. About 3 minutes after logging in to Windows, it froze again. I rebooted and repeated the procedure. Again, less than 3 minutes after Windows login, it froze.

I booted into Safe Mode and did a system restore from Thursday's checkpoint, even though there has been no update/anything new installed in three weeks (and it has been fine since then) - same problem, Windows freezes almost immediately.

I booted into Safe Mode again and it is perfectly stable - been on and fine for almost 3 hours now.

So... any ideas what the problem could be? I know Safe Mode doesn't use all the drivers and whatever, but it doesn't sound like faulty RAM and the hard drive appears to be intact, so I'm guessing it's the board.

I have hauled all the data off in the meantime, so that's all safe and it's no huge disaster if the machine is scrapped (it's old; nobody has Socket 478 boards any more), but if it is salvageable that would be 1st Prize.

Any ideas from you tech-minded people out there?

Thanks
-d-

 

[Corny]

If it runs fine in safe mode it most likely is not the board.
It could be the graphics card, it could be something overheating.
Check the temperature sensors if you have any.
There are boot-cds to test your ram (memtest86).
if nothing works - you need to start taking the machine apart to find the culprit.

 

[omminc]

Basically what Corny. memtest86 is defiantly a good indicator if the ram is going bad.

It could be from a case of something over heating, too possibly the PSU failing (if you have a spare try, if it still fails then it's not that).

good luck.

 

[jft245]

Had the same problem with a machine at work this week. Turned out it was overheating. Took a can of compressed air and blew the dust out of the fins of the heatsink over the processor and it has worked fine. The machine was a Sony Vaio and it has a built in diagnostic program which shows fan speed and temp and the numbers for the temp were showing up in red so I assumed that meant it was running hot.

 

[blackbeltninja]

Thanks for the advice, guys - about to run memtest.

However, just looking at the hardware monitor in the BiOS has revealed something scary - the CPU is sitting at 112 degrees with its PSU at 65 degrees, just in the Bios running after the POST screen. An identical system two stations down is sitting at only 61 degrees at POST and its PSU at 61 degrees. Both cpu fans are running at about 2500 rpm.

I cracked the case open to have a look - the CPU fan is spinning fine. I then powered down and took off the fan and heatsink to make sure there wasn't too much dust on the heatsink (like I had at home which caused my CPU to overheat - no air getting to the heatsink AT ALL there was so much dust) but it is clean and certainly doesn't feel like 112 degrees.

There is some sort of mammoth fuckery afoot here. Any other ideas? CPU fucked, perhaps?

-d-

 

[Corny]

CPU fucked = machine will not even boot.

what kind of cpu is it? some do get warmer than others.
try running the machine with an open case, see if the temp goes down

 

[blackbeltninja]

^That was with an open case.

It's an old Celeron 2.6Ghz single core, socket 478, probably a Prescott but maybe a Northwood. As I said, though, the identical machine 2 stations down is running at half that temperature (65 degrees vs 112).

It didn't quite pass Memtest - whole thing froze at 59% complete, even though it can sit for hours in Safe Mode - not sure how this works. I will try hauling out the ram modules next to see what happens there and run memtest again. I'll also see if I can find a desktop fan and blow some cold air in there manually.

But first... samples to prep. The joys of assisting on a drug trial, eh?

-d-

 

[blackbeltninja]

^I didn't think of that. Closer inspection reveals no obviously popped capacitors.

However, I have now also hauled out the ram - a 256mb and a 512mb. The 512 stick seems to freeze and not even start windows on its own, the other appears to be okay, just extreeeeeeeeeeeemely slow under XP/sp3.

I tried replacing the 512 with two other brand new sticks I have in my drawer and it freezes. If I use any 512 on its own, it freezes, in both ram slots.

It only seems to work with the 256... but the 256 still freezes with memtest.

Chances that both ram slots blown, both in one day?

-d-

 

[blackbeltninja]

^I don't think it's that since the box runs fine (but veeeery slow) in normal mode with the small ram in it.

I don't particularly want to have to reinstall and go through our network service implementation queue thingy to get critical updates, the virus scan engine and the updated virus dat files, the official coporate email system and Office only to find that I'm still stuck with the bloody thing. So we'll call than plan D for now.

Plan A is you guys magically solve it Plans B and C are still being formulated.

-d-

 

[Corny]

If the box runs fine in safe mode, perhaps you just have some corrupted files in your operating system. It could be that simple.

still the most likely thing in my opinion

 

[blackbeltninja]

Fine, you win. I'm not formatting, though - just a reinstall in a new directory. It's not my machine and I want to make sure everyone has their shit off there before I blank the hdd.

I backed up the user directories (the my documents folders) but who knows what else might be stashed there?

-d-

 

[star-warrior]

Maybe it's a corrupted system registry entry.

 

[mike_n_herrin]

I often say this, but it's true, i would try running a live Linux CD/DVD and see if the computer freezes while running off a live CD with the memory in.

did you check the CPU fans to see if they are clogged with dirt/dust? if they are and it's not spinning as effectively as it should it could be over heating. perhaps in safemode cpu utilization is lower than regular but then when you run memtest it's shoots up again and over heats>