Random System Shutdown

[Katsushiro]

I have a 3 year old system that has recently started shutting down randomly. It shuts off immediately with no warning signs on screen. No reboot.

Behavior:
- Full shutdown without warning.
- Frequency is currently at about once per 8 hours of use, when OC it was about 60-90 minutes per shutdown.
- No automatic reboot.
- When I manually restart the machine after a shutdown it will tend to not load the BIOS. Some combination of power supply on/off, cable reconnection, memory stick swapping, mobo CMOS clearing, & waiting till the next day usually gets the BIOS to load.
- One time after a shutdown when I got it to come back up it reported the Q6600 as having two cores. System ran well (gamed slower which confirmed only 2 cores used). Next reboot had all 4 cores reporting & working.


I know it isn't a:
- Software Issue: I dual boot an old HDD with Vista and a SSD with Win7. Happens on both drives
- Overclocking Issue: I used to OC my q6600 to ~3.1Ghz with no issues. Since the shutdown I've run everything stock clocks. I never had success OCing the memory so it is stock.
- Electric Company issue: Nothing else in the house shuts down. Monitors don't lose power.

I doubt it is:
- Temps: Aftermarket CPU cooler, new case with extra fans (Raven 3), I boltmodded my motherboard 6 months ago (more pressure on heatsinks for northbridge & southbridge). Under load my CPU temps don't go over 60C. And then there are times when the PC is idling and it will shut down.
- Power Supply: I put in a new Corsair HX750 3 months ago. Been working fine until a couple weeks ago with the shutdowns. I do have my old TX550 that I can put back in but I don't want to do unnecessary things.
- Memory: I used to run 4 sticks of 1 GB. This required a voltage bump from auto (1.8) to 1.95 or so for stability. Since then I am now running just 2 sticks in an attempt for max stability at mobo stock settings.
- Graphics Card: I put in my old 8800GT and it didn't seem to help. The 5850 is back in at stock clocks.

I suspect it is:
- Mobo: When running OC I had to bump up voltage by 16%. Perhaps that did damage?
- CPU: The one time it reported 2 cores. Could all be mobo issues though.

When I OCed:
- CPU max load temps were typically 72C, I never let it go above 74C.
- 5850. I mined Bitcoins this summer. I had the fan up pretty high with a high OC on the card. Temps never exceeded 85C. Normal load was 79C. I ran it this way for about 6 weeks straight. No artifacts in games since then.

Relevant Parts:
Q6600 G0 Stepping Processor
MSI P6N Platinum Motherboard (boltmodded)
2x1GB DDR2-800 G-Skill RAM (with the red heatspreaders)
HD 5850 Sapphire
HX750W Corsair Power supply
Hyper TX2 Cooler
Raven 3 case with extra fans

I wouldn't mind upgrading in anticipation of BF3, but I am afraid I would be a fool not to wait for Bulldozer to possibly impact prices. If I could get this fixed I really would like to wait for Ivy Bridge (3D transistors FTW).

Any recommendations?

 

[Katsushiro]

I figure I'll put the TX550 power supply back in tonight.

 

[suteck]

That's what I would try first. Especially since you removed any overclocks or extra power usages. I have seen issues from other posters about the corsair 750.

 

[tsnor]

Agree re power supply, then would run CPU tests to see if MB/CPU hurt by the overclock.

I'd pull one memory stick then the other. ("Since then I am now running just 2 sticks in an attempt for max stability at mobo stock settings. ") Did you swap which pair you used?

I would have guessed temps were the problem, but you've looked hard at that. Any chance your MB is covered with dust on parts that don't have a temp monitor? ("Some combination of power supply on/off, cable reconnection, memory stick swapping, mobo CMOS clearing, & waiting till the next day usually gets the BIOS to load." could be temps, maybe in PSU, maybe MB ??)

 

[jrwizbang]

Dead mobo. Well, not dead, but failing. The battery could be crap, or it could be a short in the power supply,or somewhere in the mobo. Try a few different power supplies. If the problem persists, then it is the mobo.

 

[Katsushiro]

Looks like its the PSU. I put the old one back in. Then, for the hell of it, put all four sticks of RAM in and OC'd to try to entice a failure more quickly.

System ran clean all night.

Thanks everyone.

 

[Katsushiro]

Best answer selected by Katsushiro.