Random shutdowns without warning - No BSODs, PSU lockout, completely reinstalled system?
Random shutdowns without warning - No BSODs, PSU lockout, completely reinstalled system?
Hey everyone,
Pardon if this is the wrong place to put this but here it goes!
I'm experiencing a weird and persistent issue with my desktop PC, it randomly shuts off as if someone pulled the plug, just instantly powers off.
After that, I am unable to restart until I
power cycle the PSU
. Then it boots just fine, runs for a while and eventually shuts down again. Sometimes during a game, loading of a game (hates Cities Skylines 2 menu?), watching a movie or sometimes even just staring at the desktop without much more than Spotify and Discord running. So, no consistent load pattern...
System specs:
Ryzen 7 5800X3D
EVGA RTX 3070 XC3 ULTRA* not the FTW
ASUS ROG STRIX x470-F GAMING
4x8GB 3600 MHz CL18 (changing these around made no difference, nor did running with two sticks (neither pair))
2 M.2 drives (No signs of wear, SMART OK)
4TB HDD (mostly idle / spun down) SMART OK
Corsair Vengeance 750M PSU
Newly installed Windows 11 Pro, fully updated and reflashed BIOS
I've tried:
Full Windows reinstall
Formatted ALL drives
Updated BIOS
Cleared CMOS
No OC or DOCP
Plugged directly into the wall + different breaker
Reseated ALL power cables + GPU
Checked all fans and temperatures and everything is fine! Even PSU fan is spinning...
Stresstest via OCCT while logging sensors via HWiNFO
CPU peaks at 146 W or 125 W by my PBO settings
GPU draws 250+ W without the slot being ¿overwhelmed? and the two 8-pin having the draw evenly distributed
No shutdown occurred during these tests, this time. However after running one stresstest it did crash 5 minutes later, on the desktop
Changed the PSU from Multi-rail to Single-rail
Can't see anything burnt on any part (except PSU, difficult to check that one) or smell anything funny.
It's been running fine for roughly 1,5 years so I haven't brought myself to reseat the CPU, can't believe that would be the culprit.
Through the logged data I found the 12V rail moving +/- 1.4%, 5V moving at most 0.11V and 3.3V staying between 3.335-3.357V.
Same goes for the voltage reported by the PCIe slot with a variance of 2.4%
Event Viewer shows the generic Event 41 and 6008 stating "
The system had rebooted without cleanly shutting down first.
" and "
The previous shutdown at xx:xx was unexpected.
" Minidumps are never generated either, the shutdown is just too abrupt.
Anyone with experience of these PSU lockout shutdowns? I'm stuck and would appreciate any feedback and ideas as my natural next step is to try and start swapping part by part, but that isn't exactly easy without parts on hand.
Welcome to the forums, newcomer!
PSU details:
Changed the PSU from Multi-rail to Single-rail. Could you clarify what this means?
After making that change, I can't restart until I power cycle the PSU.
Do you mean turning the rocker switch on the PSU or the wall outlet from On to Off and then back again?
You might want to consider getting a more reliable 850WPSU and check if the problem continues. The RTX 3000 series had a high transient load spike issue;
View: https://www.youtube.com/watch?v=wnRyyCsuHFQ
Thank you!
The PSU is quite old, having been changed in January 2020 from Multi-rail to Single-rail.
It was obtained from the Corsairs website.
It seems like a simple switch inside that adjusts the rail choice—higher tolerance for spikes with one rail or more sensitive OCP with multiple rails, though I'm not sure. It didn't make much of a difference either way.
I had to turn off the back rocker (or unplug it), hold the power button until all lights turned off (LEDs on motherboard and Ethernet lights), then flip the rocker back on and power up the system.
Could I have just been surfing for a year or two before the PSU wore out from the stress? The system would shut down right when starting Cities Skylines 2 (it really seems to dislike that game?) or staying idle on the desktop, but that doesn't seem to fit. It's a tricky situation.