PowerSpec G901 RTX 3080 Video Card Shutting Down?

Pmoss

I have a PowerSpec G901: POWERSPEC **G901 5900X/32/2/3080/PG

As of about 2 weeks ago the video card has started randomly shutting down while in use. This is the troubleshooting I've done so far:

I was using 2 monitors, one on DisplayPort and one on HDMI both from GPU.

Rebooted into safe mode and used DDU to remove all video drivers. Reinstalled most recent Nvidia drivers after reboot.

Ran single monitor on DisplayPort, Ran Single Monitor on HDMI

Monitored temps, GPU does not get above 51C at any time when shutdown occurs. Shutdown has even occurred when temp was under 50C and it hadn't even triggered GPU fans to turn on.

Disabled Gsync on both monitors.

The computer has been working fine for 8 months and all of a sudden this started happening.

There is nothing in Event Viewer for Windows as the PC itself does not shut down the lights on the GPU turn off and obviously that means no video output to monitor.

I've ordered new VESA DisplayPort cables but due to the fact it's also happening on HDMI I'm not positive that's the right solution.

I don't think it's power draw, as it happens seemingly randomly and there's no heat temp increase on the GPU. Yesterday after 2 hours of use and then today after 30 minutes doing the same thing. (Web Browsing, some light gaming, a 20+ year old game no less, so nothing AAA or stressful on the GPU)

I haven't ever opened the computer case so there's nothing that has changed internally.

Any ideas? Suggestions?

Thanks

Edit: Also tried bypassing Surge Protector and plugging directly into wall outlet and the GPU still shut down. Also played music to verify it's not the computer shutting down only the GPU. Music continued to play after video card shutdown, so it's not a system crash, just a GPU issue.

PowerSpec_MikeW

@Pmoss

You've done a pretty thorough job and seem to have isolated this to the card itself. One other thing I'd suggest trying the next time this happens. Press Control+Shift+Windows+B. System should play a tone and reset the video driver. See if that has any effect.

You said the LED's on the card actually turn off when this happens? If you're comfortable, you can also try moving the GPU to another x16 slot on the board.

Pmoss

@PowerSpec_MikeW

Thanks for the response!  I've actually tried the ctrl-shift-win B, and it beeps, but display does not come up.
The GPU is shutting itself down, the lights on the card shut off completely, but the computer is still running.  If I'm playing music when it happens, the music continues with no video.

Rebooting fixes it temporarily, but sometime within 2-3 hours it shuts down, and at temps as low as 37C.  It actually shut down right after I posted here, with nothing running except one tab of Chrome.

I will try another X16 slot and let you know if it helps.  I'm getting to the point where I'm going to have to assume it's the GPU, and setup a repair appointment to diagnose and confirm.

Again, thanks for your help!

PowerSpec_MikeW

@Pmoss

Testing another x16 is the last thing to try and it's very unlikely the board an issue here. It's one last thing to rule out. Please do let us know if you see any changes after moving the card.

Beyond that, you're correct. Next step is to bring it in and we'll take care of it under the manufacturers warranty.

Pmoss

@PowerSpec_MikeW So I went to move the card to the other X16 slot, and when I was removing the card I realized it hadn't been seated properly and wasn't "latched in".  I reseated it and double checked all the power cables, and it's now working correctly!
While trying the other X16 slot wasn't the exact answer it did lead to solving the issue, so thank you!

PowerSpec_MikeW

@Pmoss

Good catch. Intermittent signaling issues can cause all kinds peculiar symptoms that are hard to diagnose.

Pmoss

@PowerSpec_MikeW Unfortunately it ended up only working for 4-5 days and then started crashing again.  I brought it in for repair and now I'm waiting.  

As of now they can't replicate the crash, I'm not sure where we go from here?  Will they continue trying?  It's a 5hr roundtrip drive for me to drop off/pick up, and I know there's something wrong, but as you said, intermittent issues can be difficult to diagnose.

Will I just be told there's nothing wrong and to come retrieve the PC?  Or will they continue to try to replicate the crash? 

Thanks for your help!

PowerSpec_MikeW

@Pmoss

Give them some specific games to run to recreate the crash. Personally I'd look in the Windows\Minidump for crash logs. Those can be analyzed and we can make an educated guess on what the problem is potentially. They're not always helpful.