PC Master Race
A community for PC Master Race.
Rules:
- No bigotry: Including racism, sexism, homophobia, transphobia, or xenophobia. Code of Conduct.
- Be respectful. Everyone should feel welcome here.
- No NSFW content.
- No Ads / Spamming.
- Be thoughtful and helpful: especially when new beginners have questions.
First check. PSU. They fail a lot since they are "always on"
Ive had this multiple times.
The only way of checking the PSU I've found was by basically shorting pin 16 and 17 on the 24 pin cable.. but does that check anything other than if it turns on? Other test suggestions?
Stress test your GPU and see if it crashes.
Then use somethibg line LACT to limit GPU wattage draw and see if that prevents crashing. Prime example of bad PSU if so
Testing the external questions:
A stupid one, but check the power cord is in securely and nothing is pressing against the power switch on the power supply!
Plug a lamp into the same extension lead and see if it flickers when the reboot happens.
Plug a lamp into the same extension lead and see if it flickers when the reboot happens.
The cables are fine but I'm keeping this one in mind
Make a bootable USB or CD with memtest86 and boot from it. Then let it test your memory. Based on what you're describing, it's a good candidate for a fault.
Usually a memory fault results in random blue screens on most crashes, not in a consistent pattern of going dark and booting again. Still, it could be memory, and it doesn't hurt to try it.
That's a fair comment, but there's a secondary reason I suggested it, eliminating the OS as a potential cause. If it still crashes while booted into memtest86, it's much more likely to be a hardware fault, on the other hand, if it stays running overnight, it's more likely to be software.
Well it ran fine for 2/4 passes in under 4 hours before I got fed up with waiting and canceled the rest to do it tomorrow while I'm gone so that seems promising!
Good point.
I believe the default behavior is to crash dump and restart on BSOD. Holding on BSOD display must be configured.
Usually shows on the screen for at least a moment, but yeah, if it writes the crash dump fast enough you might not see it.
Got it, thanks! Trying that now
Does the restarts happen on idle, gaming/heavy load, or just randomly?
Idle/Random tends to point to ram sticks being faulty, run memtest for few hours or so?
Heavy load crashing could mean flaky power supply.
Really anytime; first I only noticed it had happened when I left it idling and came back to a login screen, but other times I was just watching youtube and scrolling Lemmy.
I've had it happen during gaming once I think but it didn't seem to be because it was too much; it was just a random moment without anything special happening.
I'd throw memtest on a throwaway usb stick and let it run tests overnight or so, unless errors appear immediately.
Easier than doing a reinstall. :)
Got it, will do. Thanks.
Ah there goes my idea, I thought it was gonna be down to Microsoft's inability to not fuck sleep up every couple of years
Memory is probably the next port of call if you don't see anything in event viewer to indicate a driver issue.
Another commenter suggested memtest which is a good shout, might be also worth putting a Linux distro onto a flash drive or partition and try running that for a couple of days to see if it does it under Linux, that will at least help inform you as to whether it's hardware or software
~~Double post.~~
Are you running Linux or Windows?
Windows 10
Edit: no double post, just a crosspost
Windows 10
Yea that’s your problem.
And I’m not trying to be a “hurt durr use Linux” but my wife’s windows laptop will do that from time to time and none of my Linux PCs does anything other than what they’re told.
I understand the sentiment but the weird thing is that nothing like this has ever happened before.
Got it, my first guess was Windows Update, but that definitely isn't the case. When you say restart, did you actually see the whole shutdown and startup process, or did you just come home to find it on the login screen with nothing running? Also, have you checked the Event Viewer to see if anything was recorded there?
Thought it was a double post because it showed up twice on my feed within seconds of each other, and I didn't notice different communities, sorry.
Yeah I saw it restart. Screen just goes black and PC turns off, then starts back up as if I booted normally. Event viewer didn't really seem to log anything specific around the time of crashing :(
Is this a new problem?
If so what was the last change made?
Does it happen under heavy load or does it happen when idle or both?
Does the OS restart or it it a total power loss?
Yes a new problem, started just a few days ago. I haven't changed anything as of late; last one was installing a game about a week or so earlier.
OS restarts, no errors or anything.
Maybe faulty ram? Pull out all ram sticks but one and see if it happens again if it does swap ram sticks until it doesn't.
If that doesn't solve it it may be a failing PSU or other major component like the GPU or HDD.
A good starting point is removing everything not needed to boot and slowly adding them back again until it fails.
Those were my ideas as well. Here's hoping it's the RAM because I could work with less and don't have the money for a PSU replacement.
Yeah, PSU is my top guess based on what all you've said. My next step would be to swap to a test PSU to see if it behaves, but since that's not an option for you now, testing the other things people have suggested is worth a shot and can't hurt.
If there are absolutely no errors in event log before the restart, PSU really is a top contender. The system will have had no warnings of any kind. If the PSU stops delivering adequate power, it's likely to restart, and this is at a low level (i.e. the motherboard restarts the system).
It's difficult to diagnose too. For mine, I was able to get more-or-less consistent restarts by requiring more PSU current by putting the system under heavier load. Once I saw the restarts occur as fans / drives / GPU were spooling up, I swapped my PSU. That was the issue.
The good news is that (well-made) PSUs usually fail in a way that won't damage components. And yes, even good PSUs can fail, especially if they're being used above their rating. And even the best PSUs don't last forever -- best practice to change them out every few years, in any event.
Do you live somewhere where the temperature changed recently? I had a friend who's PC would thermal limit and shut down if the room went above 80F in temperature.
Western Europe; not that exotic, really.
You could try swapping out the CMOS battery on your motherboard. A dying battery can cause unpredictable reboots.
Not something I'd considered but I'll put it on the list of potential fixes if others fail, thanks.
It's worth trying before replacing a PSU or memory.
Dirty cpu filters can do this, especially in the spring (as winter brings extra crap in the air, and suddenly it gets warmer).
That would be more obvious if temps seemed off but there's nothing that suggests that.
UPDATE: Memtest ran all day without any problems; no issues found and it hasn't rebooted during so I'm guessing it's a software issue, rather than the PSU, right?
OS Windows 10
Normal operation, unfortunately