Mysterious shutdowns...

JanHammer

n00b
Joined
Jan 31, 2007
Messages
7
Lets start off with specs:

Thermaltake Armor (the black heavy model)
Athlon X2 6000+
Two gigs of OCZ Ram
Coolermaster Extreme Power Duo 600w power supply
PNY 8800GTS 640MB

Right, anyway, I keep having these shut downs, which tend to be 12-24 hours apart. I have a few theories for this, since I have ruled out it's my CPU or video card overheating (I've ran prime95 and superpi for about an hour.) Basically what happens is it just shuts down, no blue screen, no real warnings or reason for it. When it shuts down the power light blinks and I can't just turn it back on, so I have to power cycle the power supply then turn it back on.

Basically theory one is that the motherboard's shorting where it sustained an injury from a rogue screw driver. What happened basically, I was getting frustraited while prying off the old cpu cooler for my Athlon 3800+ and decided to use a screwdriver to get it off. Well, it ended up slipping and injuring about three printed wires that look like they're related to the cpu.

And theory two is the power supply is slowly overheating due to its design and the case's position for the powersupply. There's only one fan on the power supply and it's on the bottom (one big 120mm). And it's almost completely blocked by the full hard drive cage. The power supply auto controls the fan speed and it sounds like it's running max speed constantly. This seems the most possible and I'll probably end up flipping the power supply so it sucks air from the other side (there's actually more room over there :p).

EDIT:
Here's the low-down on the current situation, and it's gotten a lot weirder. If I reinstall the nvidia drivers in safe mode I can game once I reboot into a normal windows session. But, I play Simcity 4 a lot and if I press "Don't save and go to region" in the options, it'll shut down the machine and break the nvidia drivers again (that's the weirdest part) to the point where I can't do any gpu intensive things (and I have to repeat the process again). I don't know if this is the only program which can break the drivers.

It literally works every time, all I have to do is just press "Don't Save and go to region" and the whole machine just shuts down, no errors, no BSODs. This wasn't like this yesterday and I haven't done anything to this machine since then.
 
Have you got another PSU to test with?

Try to move the drive cage out of the way, even if it isn't the problem then it will improve the PSU's lifetime.
 
I don't have another PSU that can handle my current hardware :p. But what I did just now is turn the power supply to the other side where there's a lot more room for air. And so far the power supply sounds a lot less loud and I feel more air current in the back. If it shuts down again I'll flip it back then move the hard drive cage temporarily.
 
The shutdowns are getting worse now, and this time they're happening when I do anything graphics intensive. But I don't need to power cycle to turn back on the machine now. I'm still not getting any error message from anything so I honestly have no idea what this could be. I assume since I don't need to power cycle the power supply it's not causing this.
 
What about just actual heat? Have you tried running without the side on? If you hit the motherboard with a screw driver you might just have to replace the board.
 
Oh there's no heat problem on anything with a thermal censor (I've ran stress tests and watched CPU temps, ambient temp, etc). The power supply is the only thing I'm clueless on the heat, but once I flipped it it seems to be blowing a lot more air through it and the air itself doesn't seem very hot.

Plus if it is a board problem because of that injury why is this just now developing? I've been running this setup for about a week and a half now and things are suddenly starting to act odd.
 
Try running a memory tester.

Do you have any other hardware you could substitute in like graphics cards or RAM?
 
what kind of motherboard? i had this same problem witha biostar mobo i had last year.....baffled me for the longest time (about 3 months) and after replacing 90% of my computer out off a whim i decided to flash the bios with a update and it fixed it.....don't know, couldn't hurt (unless u fubar the flash..lolz)
 
Uh it's an MSI K9n Platinum (non-SLi, weird). A bios update is probably a good idea actually :p.

Weird, I just reinstalled the video drivers and it's not doing it anymore o_O.

EDIT: And it's doing it again...

Here's the low-down on the current situation, and it's gotten a lot weirder. If I reinstall the nvidia drivers in safe mode I can game once I reboot into a normal windows session. But, I play Simcity 4 a lot and if I press "Don't save and go to region" in the options, it'll shut down the machine and break the nvidia drivers again (that's the weirdest part) to the point where I can't do any gpu intensive things (and I have to repeat the process again). I don't know if this is the only program which can break the drivers.

It literally works every time, all I have to do is just press "Don't Save and go to region" and the whole machine just shuts down, no errors, no BSODs. This wasn't like this yesterday and I haven't done anything to this machine since then.
 
- Google, download, and burn Memtest+ to a CD and run it for at least three hours. If a single error shows up, then the RAM is bad and you need to RMA it.
- If you're OCing, set everything back to stock and see if it happens again.
- Running Prim95 and SuperPi for over an hour isn't a really good indication for CPU stability. Run Orthos for 4+ hours if you really want to know how stable your system is
- Try another power supply.
 
- Google, download, and burn Memtest+ to a CD and run it for at least three hours. If a single error shows up, then the RAM is bad and you need to RMA it.
- If you're OCing, set everything back to stock and see if it happens again.
- Running Prim95 and SuperPi for over an hour isn't a really good indication for CPU stability. Run Orthos for 4+ hours if you really want to know how stable your system is
- Try another power supply.

-I'll try the memtest, that could be the problem.
-Yeah I was overclocking the video card but I immediately set that to default settings once these drivers started killing themselves.
-I was actually just running it to get the temps as high as I could to see if anything would overheat
-I can't imagine that being the cause of the current issue.

EDIT: It isn't the memory I did memtest86+ all night, no errors. Then I swapped out the new stick of ram and left the proven working ram in there, still shuts down :/. I'm betting on a bios update.

EDIT: Bios update didn't help either.
 
My computer also shuts down during 30+ minutes of heavy graphics (ie; playing a game).

I also use an 8800 GTS 640mb too.

Taking the side of my case off solved it for me. I just left it off, and I've had no problems since.

The fan on your video card won't do much good if theres no air flow in your case, but hot air.

I just have to get around drilling a couple 80mm holes in the side of my case and installing fans someday.

Anyways, try that. Play with your computer for awhile with the side of your case off. I bet that solves it.
 
Do you have a spare vid card to swap in?

Yeah, a 7600GS, I'll try that in a bit.


Also, Ahzar, I have very very good airflow in the case and the temperature has never gotten that high on my 8800GTS (upper 60's to very low 70's), well, in G80 standards that is :p. Plus I have the video card set to run at 100% fan speed all the time.
 
Back
Top