PC powers off randomly, need to power cycle PSU to turn PC back on

T

timg1281

n00b
Joined
Dec 27, 2010
Messages
12
Here are my PC specs:
Processor: i9-10850k
Motherboard: Z590 Aorus Xtreme Waterforce
GPU: Aorus GeForce RTX 3090 Xtreme Waterforce 24G
RAM: 4x 8GB G.Skill F4-3600C16-8GTZNC
PSU: Corsair AX1600i
Hard Drives / SSDs:
Windows - T-Force TM8FP7001T
Other: T-Force TM8FPL500G, ST4000DM004

The PC was fine for a while but has started completely turning off at random, as if someone has pulled the power cord. The only thing that remains on is the RGB on my reservoir (EK-Quantum Reflection PC-O11D). I also have to power cycle the PSU using the switch on the back to turn the PC back on as the power button, etc. is not responsive. More often than not, this happens when I've walked away from the computer while it's idling although it happened for the first time yesterday while I was actively using it (just doing some web browsing and light text editor stuff, nothing taxing). Temperatures are very low across the board so it's not a temp. thing as it has done it maybe 30 minutes after powering on in the past or many, many hours into the day. Voltages from the PSU look ok too.

I've gone up to 2 weeks without this happening and the PC was on for 8-12 hours pretty much each day, with a mix light duty stuff and taxing games on all those days. I've also had it happen a couple of days in a row so seemingly no pattern to it other than it's usually when I walk away and the PC is just idling. Nothing in Event Viewer when going back to try to diagnose. The only non-stock thing I've done is enable an XMP profile for the RAM but I've ran memtest all day without any errors so I don't think that's it. I'm kind of at a loss as everything seems to be working 90% of the time until it suddenly isn't.

I'd appreciate any insight you guys could provide and happy to share more info if needed.
 
You can test it with one stick of ram at a time to rule out memory problems, and only use your boot drive while testing. But I'll bet it's fine.

There have been a lot of reports on the current gen of high-end video cards having problems with older PSUs. I know that was a pretty good PSU when it was released, but it might still be on the edge of too old to reliably power your GPU.
 
Try it without XMP enabled for a couple days to rule out memory as the cause as it's the simplest way to begin to troubleshoot the issue.
 
I was considering it hopefully being a PSU thing and am in touch with Corsair. I had tried disabling XMP a while back and while I didn't see the issue for the week or so I ran that way but I also didn't see it for the most recent 2 weeks running with XMP enabled and playing games like Cyberpunk regularly, until recently that is.
 
The first thing I do with crashing issues (with a wced rig)is pull the psu and swap in my backup(you seem to have eliminated the memory). Then comes the pita testing...gpu, mb and cpu. Testing the gpu would normally be the easiest but wcing it nixes that. Unless you're using qdcs? If your using soft tubing and have left enough slack you may be able to pull the gpu and definitely the cpu for testing. It's a pita but you can do it without draining your loop with the help of long zip ties, that is if you have something to swap in to test with.
 
Unfortunately it's all hard tubed and swapping out the gpu would be a royal pita, same with the cpu. I also don't have anything to swap in anyway. I am going to try to get myself another power supply and see if that helps. It's the only thing I can swap out that won't require a drain / rebuild.

I did have the cable management pretty tight and I loosened it all up and the problem seems to be happening less but I don't really have anything to back that up other than possible coincidence.
 
I'll do that as soon as I can order a new one and update my post after testing. Thanks guys!
 
Axman said:
You can test it with one stick of ram at a time to rule out memory problems, and only use your boot drive while testing. But I'll bet it's fine.

There have been a lot of reports on the current gen of high-end video cards having problems with older PSUs. I know that was a pretty good PSU when it was released, but it might still be on the edge of too old to reliably power your GPU.
Click to expand...


Any particular PSU you'd recommend for a setup like this?
 
I don't nerd out over PSUs, I just buy SeaSonic. I'm definitely not the expert on PSUs around here, I just keep seeing threads about PSUs causing crashes with high-end video cards.
 
Yep, seasonic focus gold, plat or prime, Corsair rmx, hx, ax, EVGA g2/3 or newer g6, g7, bequiet! Straight or Dark power units are all solid to mention a few. I would shoot for a 1000w unit to try and avoid the transient spike issues but you could easily get by with an 850w.
 
  • Like
Reactions: Meeho
like this
You must log in or register to reply here.
Back
Top