PC randomly powering off

My PC has started randomly powering off after, at most, 10-20 minutes. Once it's "warmed up" it generally powers off after only a couple of minutes. When it powers off, it behaves as though the power was simply unplugged or there was a sudden power outage. No BSOD, etc. My only guess is that it's an issue with the power supply, because Windows sees nothing wrong with hardware (aside from reporting that it failed to boot or crashed unexpectedly).

So, a couple of questions:

1. Does this sound like a power supply issue?
2. How can I test my power supply?

As a side note, this of course happened after I started planning a PC upgrade and then decided against it. Murphy's Law, eh?

Did you check your CPU and GPU temperature before the system shut down? There is also some advice on testing a PSU here.

complexmath wrote:

Did you check your CPU and GPU temperature before the system shut down?

This 100%. Seriously sounds like an overheating problem from a poorly seated CPU/GPU heatsink or whatnot. Have a gander in the system and if there's nothing obvious reseat everything and clean (with compressed air) everything out....

HWinfo will let you do logging to a csv file, which you can load in excel and watch for problems in it's dying seconds. It covers pretty much everything I'd want to monitor.

Duoae wrote:
complexmath wrote:

Did you check your CPU and GPU temperature before the system shut down?

This 100%.

doogiemac wrote:

Once it's "warmed up" it generally powers off after only a couple of minutes.

Definitely an overheating issue, more likely CPU than GPU.

psoplayer wrote:
doogiemac wrote:

Once it's "warmed up" it generally powers off after only a couple of minutes.

Definitely an overheating issue, more likely CPU than GPU.

The problem is practically every component warms up in use.

This is actually (potentially) good news. My power supply is pretty new and of good quality. I'd not be surprised if the CPU or GPU is fairly dirty by now. I'll trying cleaning and reseating everything and report back!

doogiemac wrote:

This is actually (potentially) good news. My power supply is pretty new and of good quality. I'd not be surprised if the CPU or GPU is fairly dirty by now. I'll trying cleaning and reseating everything and report back!

Make sure the tabs that hold down the heat sinks are not broken. I've run into that a few times with friends computers.

The machine should not be powering off from a CPU overheat, period, unless you're using a very, very old AMD motherboard/chip. What will happen instead is that the CPU gets too hot, and then it self-throttles. That emergency powerkill sensor was on the AMD motherboards, and because it was under the CPU, and took a really long time before it triggered, it was barely any good at all anyway.... under most circumstances that would cause thermal damage, the CPU would be melted way before that sensor had any idea anything was even wrong.

I'm not familiar with current AMD behavior, but I don't think Intel chips even have an emergency thermal shutdown anymore; they just self-throttle, and stay right at 90C.

You should never get a hard poweroff, in other words, from CPU temps on a current system, unless you're running some kind of software utility that does this. If you are, you should be able to find it.

AFAIK, GPU temp spikes can still cause emergency shutdowns if the video driver chooses to do so. I think your most likely culprits are either GPU overheat or power supply failure, your original thought. In the case of any software doing this, whether video driver or some kind of CPU temp monitor, there should be traces in your system logs of what happened and why the machine was powered down. If you don't see anything, it's much more likely to be an actual hardware fault. I'm thinking what you're thinking, that the power supply is probably borked.

I'm not sure why everyone is jumping on 'overheat!' first thing. Do AMD boards still do that?

I think overheating gets kicked around a lot because it's an easy one to diagnose, and in the case of CPUs could be where the person building the PC messed up a bit.

Even if the heatsink is totally wrong, with a modern CPU, it should just self-throttle, it shouldn't do a hard poweroff, unless some monitoring software is getting trigger-happy. Your computer will be slow as heck until you fix the heatsink, but it shouldn't just turn off.

Modern bios you can set alarms for when the PC reaches high temps...

I've had laptops shut down if they became overheated. Either way though, it's the easiest thing to check.

I just went through this with my son's computer. had to do a rebuild with a new mobo, cpu and ram. everything else was recycled.
I'll see if i can find my thread. It may help.

http://www.gamerswithjobs.com/node/1...

Was not the PS as i had replaced it too but then chose to re-use.

Instead of starting a new thread, I figured I'd piggy-back on this one.

My laptop just went poof on me while just web-surfing. It refuses to power on at all. I have it plugged in while using it, most of the time. I have noticed that the little green light on the "pack" on the charging-cord goes out every time I plug it into the laptop (stays on when not plugged in) even whe the laptop is off.

My best guess is a borked PS: some kind of "short" causing the charger to trip off internally.

Any thoughts?

Battery-powered stuff generally runs off the battery even when plugged in, so it could be either the power adaptor or the battery.

I've tried plugging the cord in with the battery removed; same thing... the little green light on the power cord shuts off as soon as it's plugged into the laptop.

This has been happening with my new PC build as well. It restarts at random times and sometimes it just shuts down and I have to wait 20 seconds before I can boot it again. This happens when I am simply browsing the web, or playing a game. It happens once a day, multiple times per day and sometimes not at all.

I dont have any overheating issues as the temps are well below normal. I had the 3750k at a slight .4 overclock but it was still well under the acceptable load temps. I tried turning the OC off last night and it did not shut down or restart for an hour but I have to give it more time.

I have a 4 month new Seasonic PSU so I cant imagine thats the problem. I tried updating the video drivers but that didn't help either. Maybe a reformat is in order to wipe out the drivers. Any idea what it might be?

TempestBlayze wrote:

I have a 4 month new Seasonic PSU so I cant imagine thats the problem.

Can you make sure of this? Even good brands have a failure rate. If they're really a good brand then it'll extend beyond just their hardware but also to their customer service and replacements if you can point the finger at it.

Making assumptions when troubleshooting can come back to bite you.

I am going to try and increase my voltage on the RAM a little bit. My motherboard did not detect my 1600 RAM as 1600 so I had to put it that way in BIOS. Maybe by doing that, it now need more voltage.

Also, I will try a different power cable. Start small.

Any chance the house circuit is overloaded? That could give you an intermittent undervolt.

TempestBlayze wrote:

I am going to try and increase my voltage on the RAM a little bit. My motherboard did not detect my 1600 RAM as 1600 so I had to put it that way in BIOS. Maybe by doing that, it now need more voltage.

For what it's worth, neither did my two 1155 motherboards auto-detect my RAM speed, defaulting to 1333. I had to load the XMP profile to get the right speed.