[SOLVED] Friend's computer is cutting power randomly
So my friend has a computer she put together, and after replacing what feels like every single part on the rig, multiple trips to the repair shop, and calling a priest wrestling the demons out of it, it is randomly cutting power and we think the replacement power supply might be just as busted as the last one. Are there decent odds of that being an issue, or could it be something that we are overlooking?
EDIT: So, I had a car issue pop up and I won't be able to to take a look at at it tonight. Will download the tools you all mentioned to a jump drive and will keep you posted.
Was there, stayed up the whole time. Really thinking it might be a bad power switch that stuck, so it would turn off randomly. Thanks for all your help.
FINAL EDIT: So I narrowed it down to the Graphics Card and/or the Cooling System. Running the Heaven Benchmark on Extreme pretty reliably cleans it's clock, especially turning it off and then turning it on again. This happens on the latest Windows Updates, with latest Nvidia Drivers. (RTX 2080) Pulling the card and running the benchmark anyway causes the same issue, and this time it powered on and then back off in a loop. CPU-Z stress caused it to crash, so I'm assuming it's the cooler. Thanks again.
How "random" are the power cuts? For example, if left on but completely idle (not even Chrome or other large apps running in the background) for several hours, will it lose power then? Or does it only occur when being used? Is it more likely to happen under heavy usage (lots of Chrome tabs or playing a large game) vs light usage (playing Minesweeper)?
What specifically happens when it loses power? Does it shut down, or just reboot? Is there a Windows BSOD screen or anything like that?
Assuming this is Windows, are there any errors logged in Event Viewer?
Prime95 and memtest86+ are two useful tools for putting some stress on the hardware to try to flush out any thermal issues or other purely hardware issues, such as bad DIMMs.
CoreTemp will show CPU and other temperatures. It's possible that the system is running hot already, then exceeds the thermal shutoff point seemingly randomly. That would point to issues with the CPU cooler or thermal paste.
Yeah, that sounds like the most reasonable thing after testing. Hopefully that sorts it out.
When you say "cutting power", do you mean it immediately shuts off mid-use? That generally happens due to a thermal issue. Is there a proper CPU cooler on the CPU? Is there thermal paste? Are the fans running?
Checked all those things, headed over to check again in about an hour.
Could it be overloading the PSU? If the PSU can't handle to load it will "gracefully" shut off (more graceful than frying your mobo, that is). It's possible the PSU isn't bad, just that it can't handle the load it is getting.
Run HWMonitor on it while using it, and keep an eye out for any strange temperature fluctuations.
Does the PSU have enough wattage for the rest of the PC? It might be fine otherwise, but simply be overloaded. That's a common enough cause of random shutdowns.
Okay, I am over there now. Checked power supply, 12V line is pushing 12.38, idles at about 34C doing nothing in BIOS for half an hour. CPUs are gated at 40C with Prime 95, RTX is at 71 on a benchmark, it seems to be working aok, but a lot of the issues seemed to focused around the internet. Either an internet game or an internet browser, because Disconnected, it's working fine. Going to check event logs, try a wifi adapter instead of the motherboard Ethernet, see what happens.
EDIT: Hasn't died since I got here. Maybe I have a passive tech buff? Going to Restart.
Use a hard drive health meter like crystaldiskmark. I once had a similar issue and was baffled for quite some time. Turns out it was one my hard drives that was bad and causing the entire computer to power off randomly.
Also double check and replace cables. They are cheap and it's an easy thing to overlook.
I agree with @TheJorro, getting two busted PSUs in a row should be a miniscule chance (unless you're buying shit parts), and if the PC is just shutting down immediately, it might be thermal.
I think we'll need more info on this, what parts are in the PC? What have you attempted so far?
I can tell more once I'm there tonight, but what I remember is that the GPU is an RTX 2080, the motherboard is new and the initial build was in April.
If the PSU really was bad, there's a good chance it took something else in the build to the grave.
Post your PSU model. The RTX 2000 series takes huge wattage and under-powered (not defective, just too weak) power supplies are a common problem.
Nvidia recommends 650W but that's the bare minimum.
800W, EVGA 850 G3, has been clocking away at benchmarks for 4 and a half hours. So I don't think it's hardware... Directly.
If a power supply has gone bad, there's a good chance that it took something else with it. And I also notice that most things that work for a while and turn off tend to have problems with the capacitors in it. This will more than likely be a problem with the motherboard. Check to see if any of the capacitors on the board are expanded or malformed. If so, there is your problem.
I saw something recently where the switch on the case was faulty. try bypassing that.
https://www.instructables.com/id/Replacement-PC-Case-Power-Switch/
That may actually be it, since she's been complaining about needing to press and hold on the switch, it would usually set down <10 minutes, and sometimes it would turn over like a car engine.
It's happening again. Yesterday, no issue, today, issue. Doesn't seem to boot loop in BIOS, but in Windows, not a fan.
So I narrowed it down to the Graphics Card and/or the Cooling System. Running the Heaven Benchmark on Extreme pretty reliably cleans it's clock, especially turning it off and then turning it on again. This happens on the latest Windows Updates, with latest Nvidia Drivers. (RTX 2080) Pulling the card and running the benchmark anyway causes the same issue, and this time it powered on and then back off in a loop. CPU-Z stress caused it to crash, so I'm assuming it's the cooler.
If it overheats even with the graphics card pulled out, you can definitely blame the CPU cooler. I'd recommend getting this, removing the CPU cooler, cleaning the existing thermal paste off the cooler and the lid of the CPU, then re-applying brand new thermal paste.
I had an issue over the summer of getting random power cycles that slowly increased in frequency as I tried to track down the source over the course of about a month and a half of isolating different factors. Rebuilding the windows image became a daily routine.
the tldr of it all was that it ended up being a buggy-as-hell audio driver that an audio plugin installed. iirc the only sign of the driver not working correctly was a minor warning (not even an error) in event viewer. The reason it took me so long to find was because it was reported so infrequently that it got lost in the noise of all the other mundane warnings that happen as a matter of course.
So, in addition to all of the hardware related things you're trying, don't forget software.