My computer is randomly restarting cold (no BSOD or anything), about twice a day now for the last 3 days. I ran a memory check and Windows said it's fine (defective memory is the most likely culprit), so that leaves a few culprits.

-Virus: Ran Malwarebytes, nothing came up. There aren't any processes running that suggest I'm infected.
-Overheating: It's summer, and it's hot out. Speedfan tells me everything is fine. Less than 35 celcius on all components.
-Processor or Processor Fan: I ran a test on this, a program ran for six hours last night that did some basic functions (calculate pi, sorting algorithim, et cetera) - my CPU fan is running at the right RPM and the processor worked straight for six hours without a restart. I'm not inclined to believe my CPU (i5) is malfunctioning.

I'm inclined to believe something is wrong with the motherboard or the power supply, but how do I determine which? It could also be a software issue with Windows 7, but looking in the Windows Event Log doesn't shine any light - it just says there was an improper shutdown with a few nondescript details.

I guess I have to reformat and reinstall W7 to be sure it's a hardware issue, but before I do that, is there anything else I should run?

I'm not doing anything in particular when a restart randomly occurs. Browsing the internet, mostly. I've had all the parts for this computer for a little over 1.5 years with no issues.

 

When your computer restarts does it do anything particular, or just act like you hit the restart button? Does there appear to be any amount of time the computer is usually running before it resets?

 

It's seemless, it just restarts, like I forced a hardware reset. There's no delay. It boots up normally and I can go back to what I'm doing. It won't even restart again for at least a few hours or more.

The moment before it restarts, there's no indication of slowdown or jarring graphical glitch or anything, either. Everything is normal.

 

Apart from buying a replacement motherboard or PSU, I'm not aware of any way to check which one is the problem.

Edit: I had this happen to me last week, and it turned out to be my graphics card. It was only during games though, and I would get minor lag/graphics glitches right before the crash, so that's probably not the problem you're having.

 

I'd suspect the PSU, but it is only a guess. It's also hard to verify without installing a different PSU.

1. If your PC is from a major manufacturer [eg. Dell], then look up that model number for other people having system instability and random reboots.
2. If not, then look up your motherboard model number for other people having system instability and random reboots.
3. It's possible that your PSU fan has stopped working, which is causing it to overheat. Do your temperature readouts include one for the PSU? Is the fan still spinning?
4. Buy a can of air duster [it has many brand names, but always contains compressed noble gases - compressed air has water in it, so don't use that cheaper option] from a computer, electronics, or electrical trade store. Blow all of the dust and crap out of your PSU that you can.
5. You could always strip, clean, and then re-assemble your PC; make sure to re-silver your heatsinks before putting them back on the board.

All of those are good ideas, but there's no guarantees that any of it will work. If you've got a friend who's an electrician, then they might be able to load test your PSU for you and see if it provides stable power.

 

Thanks for the tips Yak. I'll definitely try them. Speedfan can't tell me the temperature of the power supply unit, I think it's because there isn't a sensor in it. It can only tell me the voltage it puts out, which is stable. I check the temps of everything after a forced reboot and the temperatures are normal, but the reboots are so random I can't stare at the program for hours to look for irregularities.

By the way, this is from the event log, just so I can clarify how unhelpful it is:

Kernel-Power Event (63)
Source: Kernel-Power
Event ID: 41
Level: Critical
User: SYSTEM
OpCode: Info
Task Category: (63)
Keywords: (2)

EventData:
BugcheckCode 0
BugcheckParameter1 0x0
BugcheckParameter2 0x0
BugcheckParameter3 0x0
BugcheckParameter4 0x0
SleepInProgress false
PowerButtonTimestamp 0

The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

 

You don't have to reformat. Create a live cd with Knoppix, Ubuntu or whatever you like and test if the PC reboots there. If it does, you have an hardware issue and if it doesn't, then it's Windows/Software.

You can also try to benchmark you GPU with 3DMark, since your CPU seems stable, the GPU is the next in the line of suspects even if you don't get glitches. Also, clean your PC. Someone I know had reboots like that recently and even though Aida64 showed good temperatures, after he cleaned his PC everything was good again.

 

Googling Kernel-Power Event 41 (63) turns up a lot of interesting discussion threads. It's a fairly generic 'unexpected shutdown' error, but many solutions were found in software, such as disabling all power saving, wake-on-lan, wake-on-device, sleep, etc. options. Another person discovered they had two audio drivers in conflict which was causing the crash.

The possible software and hardware problems are myriad. I'd suggest trawling through those conversations if you haven't already.

Definitely try running Linux for a few days first, as Sacro suggested. Determining whether this is a hardware or software fault is probably the best first step.

 

I guess I'll run a livecd linux distro then for a while, thanks for the help everyone. Hopefully it is software related, I'd rather not take apart my whole computer to replace the PSU or Motherboard.

I did clean out some dust, hopefully it was just that, otherwise.

 

Ah, I think I may have found my problem. My PSU is only putting out 11.0-11.1 volts when it should be putting out around 12.0~ volts. Granted, I've read 11.5 is tolerable, but my PSU isn't putting out a whole volt. I got a restart to occur by toggling from League of Legends to desktop (back and forth 3 times), while launching Firefox at the same time (which I guess was enough load for a reduced voltage to kill the computer?).

Does this make sense? Or should I not even be getting problems for 11~ volts?

 

I don't rightly know. You'd probably be better off taking that specific question to a dedicated hardware forum.

According to this page:

+12VDC can go from +11.400 VDC to +12.600 VDC
-12VDC can go from -10.800 VDC to -13.200 VDC

I don't know why there's a difference between +ve and -ve tolerances; nor which one you're looking at. Either way, 11VDC is either outside, or very close to, the acceptable bounds.

 

It's gotten so bad the computer won't even stay on for more than a few minutes (can't even get past log-in screen). I did a memtest86, ran the computer in linux, and updated the UEFI, to no effect. Thanks for everyone's help, but I strongly believe it's the PSU, as dusting the thing out did improve the situation for two days. My limited warranty on the thing probably doesn't cover what's wrong with it, but I'll check with the company anyway before buying a new one. Hopefully that fixes the problem.

 

Best of luck. It sounds like a PSU problem to me too.

 

Figured I should just post here just in-case anyone ever has these symptoms: It was definitely my PSU. Company gave me a new one for free and I realize I got used to a very odd sound in my old one (I had convinced myself it was the processor fan, since it was just a slight buzzing sound), as the new one is completely silent. There was definitely something defective with it. I haven't had any issues for a week since I replaced my PSU.