Spontaneous power-offs: uptime less than 10 minutes

unprovoked

Junior Member
May 31, 2005
10
0
0
We have a computer which spontaneously powers off within 10 minutes of booting.

If I immediately try to reboot afterwards, the system will either 1) hang just before the point when normally searches for a boot record (i.e. it doesn't display normally seen "Searching for boot record from floppy" message) or 2) it will not even POST.

I have to wait a while before I can sucessfully boot to XP again. After that, I get another few minutes before the next power off. And so on...

Do you guys think this is a cooling problem or a power supply problem or something else? I replaced the power supply and the case fan, but the problem persists. There is a dusty looking AMD heatsink sitting over the CPU -- perhaps that is the culprit? However, the temperatures seem stable (although they are not low: 57oC CPU, 45oC case).. System info follows:

Antec SmartPower 350W
Gigabyte GA-7VTXH
Athlon XP1600 (1.4GHz)
XP Home SP1

BIOS hardware monitor:
CPU Temp: 57oC
Case Temp: 45oC
Fan Speed: ~4800RPM

Vcore: 1.728 or 1.712 (switches between these two values)
Vtt: 1.232 or 1.248
+3.3V: 3.280 or 3.296
+5.0V: 4.838 or 4.811
+12.0V: 12.096 or 12.160
5VSB: 5.053 or 5.026

Thanks for your suggestions.
 

aGreenAgent

Senior member
Apr 25, 2005
274
0
0
Overheating is a possibility, as well as some sort of short or capacitor leak.

Right when it turns off, reboot and get to the BIOS temp monitor. If it's more than a degree hotter than your idle temp, I would say heat.

Otherwise, I would be looking into a new mobo.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
This time, instead of power-off, the system just locked up (after about 5 minutes up) and a funny crispy smell started to emanate from the case! Upon reboot, it would not POST.

I waited a few minutes, then tried to reboot. New error:

System beeps, and the displays the error: "CMOS/GPNV Checksum Bad"

I'm starting to think it's not the temps but the MB itself. What do you think?

 

unprovoked

Junior Member
May 31, 2005
10
0
0
Update:

After the CMOS checksum errors, I hit F2 to load the default CMOS values. XP stayed up less than 2 minutes, but I was able to get into the BIOS immediately after this power-off. Vcore had now dropped to 1.392 and within three minutes even the BIOS Setup froze. Hmmmmm....
 

birdpup

Banned
May 7, 2005
746
0
0
I believe the CMOS are the electrical circuits that hold the bios instructions in the motherboard. Try flashing the latest bios and if that does not work, you may need a new motherboard. Additionally, the "funny crispy smell" implies burnt circuits.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
Thanks for the suggestion, but I am afraid to try flashing the BIOS at this point. The system is too unstable that it is likely everything wil freeze in the middle of the BIOS flash, and then everything will be corrupted.
 

Smilin

Diamond Member
Mar 4, 2002
7,357
0
0
A system that is really cranky when trying to boot is often a bad power supply.

What are the other components in this system? If you're running a high end video card and some extra drives that 350 may be straining.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
It's a low end system -- minimal ATI graphics card (don't recall exact model, probably Radeon 7500 - it is a $40 card with DVI out). It only has a single HD and two optical drives. It's my wife's machine, and all she uses it for is www, email, and Office. It's been running great for several years, until now.

The funny thing is that I just replaced the generic 300W PSU with the Antec SmartPower 350W. I tested both units with the Antec PSU tester and used a multimeter to measure the voltages.

My first reaction to the power-offs was that it must be the PSU. But the problems persist (in fact, they are now getting worse), even after replacing the PSU. So either both PSUs are bad, or there is something wrong with my MB. Or something else!

 

KGB

Diamond Member
May 11, 2000
3,042
0
0
Try removing the HSF from the CPU and cleaning it.
While it's off, clean off all of the old thermal paste (also from the CPU) and install it with fresh paste.

Then check the stability and temps.
 

helpmeout

Senior member
Sep 24, 2001
540
0
0
Has everyone missed your second post where you said you were getting a "funny crispy smell" from inside the case? Shut it down, open it up, and look for a burned component. My guess is it's something on the MOBO. When you first get a "crispy smell" from a computer, you really should turn it off and investigate (sniff around), LOL.
 

stockjock

Diamond Member
Aug 29, 2000
4,205
2
76
There is a dusty looking AMD heatsink sitting over the CPU -- perhaps that is the culprit?
Let's hope the funny crispy smell isn't your CPU...one thing I found on my Nephew's computer last week (he was having same problems as you) was that his Heatsink was sooooooooo clogged with dust bunnies that the fan was not able to cool the heatsink and his CPU was overheating....check your heatsink to makes sure its not clogged. If it is pull it off and clean it up good and then slap some thermal paste on.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
I will try to replace the heatsink next, and if that doesn't fix things, I'm probably going to upgrade the MB and CPU.
 

2cpuminimum

Senior member
Jun 1, 2005
578
0
0
This is definitely bad karma, classic symptoms. The only way to fix it is to figure out what burnt up on your motherboard and replace it. Or if possibly what burnt up was the powersupply itself. If one of the power connecters to an IDE device was loose, that could short out and kill a powersupply. If it was the cpu, next time use arctic silver III or V and clean the heatsink every few years. Air filters on the case air intake that you wash once a month also help prevent heatsink dustbunnies.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
Just an update:

I replaced the heatsink and used arctic silver 5, but the symptoms were still occurring. At that point I decided to upgrade the motherboard and the CPU, and now all is fine.

Before all the power-off problems began, I had not opened the case on that machine for years. So I don't think I could have accidentally shorted something out. That leaves me thinking the "explanation" was either random failure of the MB or all that nasty dust.

Thanks for all the suggestions.
 

unprovoked

Junior Member
May 31, 2005
10
0
0
I got a Sempron 2800+ (Palermo) and MSI K8N Neo Platinum. Idle temps 32oC (CPU) and 36oC (North Bridge) using stock cooling and no o/c. Nothing fancy, but my wife seems happy :)