Display suddenly cuts out, fans go onto full RPM, all lights on

[Cl0verr]

Recently, (within the past week or so), my PC has suddenly began to seemingly "power off", in where the display suddenly loses signal, fans begin to blast, and RGB on my keyboard suddenly freezes. When this happens, however, I can still hear audio from my PC, and seemingly have a limited level of interaction with what I was doing. It usually happens when I boot up a game and play for around 3 minutes. On MSI Afterburner, GPU frequency seems to be completely normal, memory diagnostic shows that the memory seems to be fine, chkdsk doesn't really seem to have shown any errors either. Temps for both my CPU and GPU are both at most hovering around 50-60 degrees, so I don't think overheating is the issue, and all drivers are up to date. Here's a video of the issue (and no, I do not have my PSU like that normally, I was checking to see if there was a ton of dust under it or something):

My PC specs are:
CPU: AMD Ryzen 7 2700X
GPU: Nvidia Geforce RTX 2080 Super
PSU: Coolermaster V850
Motherboard: B450 GAMING PLUS (MS-7B86)
RAM: 16 gigs DDR4 (4x4, running at lowest possible frequency cuz when I first changed the ram I was stupid and mismatched)
Storage:
Boot Drive/SSD: KINGSTON SUV400S37120G
Drive 2: WD WD20EZAZ]
Drive 3: WD WD10EZEX

My PC at its current state (storage is original from around 2016, everything else is replaced) is around 4.5 years old save the GPU, which is around 3.5 years old.

Extra info:

It has bsod'd only once, with the Kernelbase.dll failure.
It is not consistent - I have played some games for hours and it hasn't done anything, then suddenly booting up a light game like Minecraft can cause it.

 

[sLowEnd]

Are you able to get another graphics card to test with?

 

[Cl0verr]

Are you able to get another graphics card to test with?

Not currently, no

 

[sLowEnd]

Not currently, no

Did Event Viewer log any critical errors?

 

[Cl0verr]

Did Event Viewer log any critical errors?

Not very fluent in looking through event viewer, to be honest, so I'm not sure

 

[sLowEnd]

Not very fluent in looking through event viewer, to be honest, so I'm not sure

What does yours list?

 

[Cl0verr]

View attachment 287151

What does yours list?

Now that I'm looking at the past 24 hour ones, the "Dwminit" one is the only one that has had warnings only within the past week.

 

[Selaya]

faulty gpu

 

[Thimblewad]

Are your PSU cables okay? This started happening to me when my PCI-E power cables burned through and the PSU kicked in some sort of protection. The PC was still running but no video output.

 

[sLowEnd]

View attachment 287154

Now that I'm looking at the past 24 hour ones, the "Dwminit" one is the only one that has had warnings only within the past week.

Your issue sounds very similar to this

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

I FIGURED IT OUT!!! I have been dealing with this for over 9 months with a BRAND NEW system I built following all of Linus Tech Tips videos. The system would just randomly crash from loading a game, all the way to being in the middle of a game and it crashes. I wasn't even playing graphically int...

linustechtips.com

Some of the people in that thread have found success from repasting their cards, so it's probably worth giving it a shot if you're comfortable with that.

 

[Cl0verr]

Are your PSU cables okay? This started happening to me when my PCI-E power cables burned through and the PSU kicked in some sort of protection. The PC was still running but no video output.
View attachment 287162View attachment 287163

I'll check them out on my end and replace them to see if it works. If that doesn't work I'll attempt to repaste the card, since the warranty is likely long-gone anyways. I'll update if it works or not - thanks for the help so far, everyone

 

[Thimblewad]

I'll check them out on my end and replace them to see if it works. If that doesn't work I'll attempt to repaste the card, since the warranty is likely long-gone anyways. I'll update if it works or not - thanks for the help so far, everyone

Since your temps are okay, I doubt repasting the card will help. To me it seems like a PSU problem. Try to get a friends PSU to test. Don't buy anything new just yet.

 

[sLowEnd]

Since your temps are okay, I doubt repasting the card will help. To me it seems like a PSU problem. Try to get a friends PSU to test. Don't buy anything new just yet.

That's my initial thought too, but apparently some people in the LTT thread had "fine" temperatures too before repasting.

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

If there's an issue with hotspots, repasting could alleviate the issue. Paste isn't very expensive, so I figure it may be worth a shot.

 

[igralec84]

My 4090 has been doing this (display off, fans 100%, drivers crashed but still in windows just no GPU probably) with the Cablemod cable, stopped on the OEM one, so might be a power issue to the GPU.

 

[Thimblewad]

My 4090 has been doing this (display off, fans 100%, drivers crashed but still in windows just no GPU probably) with the Cablemod cable, stopped on the OEM one, so might be a power issue to the GPU.

Exactly same here when I burned through my cable. Took me a while to notice, thankfully nothing happened to my old Vega.

 

[Cl0verr]

My 4090 has been doing this (display off, fans 100%, drivers crashed but still in windows just no GPU probably) with the Cablemod cable, stopped on the OEM one, so might be a power issue to the GPU.

I'm beginning to think this is the case, but after changing both the cable and the position of the power cable into the PSU, it still happens - so I'll be attempting to get a friend's PSU to test this, and I'll also replace the thermal paste to see if that solves the issue as well.

 

[Selaya]

it's almost guaranteed to be a faulty gpu and not a psu issue

there's an entry for an nvlddmkm crash in the eventvwr, and the other behavior observed (display cutting out while gpu fans go 100% rpm) is consistent w/ what you'd observe from a gpu driver crash which is usually caused by a faulty card (or bad oc, but idt the op's ocing?)

 

[Cl0verr]

it's almost guaranteed to be a faulty gpu and not a psu issue

there's an entry for an nvlddmkm crash in the eventvwr, and the other behavior observed (display cutting out while gpu fans go 100% rpm) is consistent w/ what you'd observe from a gpu driver crash which is usually caused by a faulty card (or bad oc, but idt the op's ocing?)

I am not OCing, and the nvlddmkm crash has only happened once in the past week, and does not happen when observing the problem, and testing it consistently.

That's my initial thought too, but apparently some people in the LTT thread had "fine" temperatures too before repasting.

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

BLACK SCREEN GPU FANS AT FULL SPEED CRASH [FIXED]

linustechtips.com

If there's an issue with hotspots, repasting could alleviate the issue. Paste isn't very expensive, so I figure it may be worth a shot.

Sorry for Double-posting, but I decided to run a quick test with all benchmarks up and wanted all of your opinions on it. It seems the temperature goes from around 49 degrees to 85 or so degrees in less than a minute. Along with this, the GPU does suddenly get around 223 watts of power at the same time. Here's the video: https://www.youtube.com/shorts/1lDAU-bANFo

I believe it might be the thermal paste problem said then, is the temperature spike meant to be that drastic within such a short period of time?

 

[sLowEnd]

I am not OCing, and the nvlddmkm crash has only happened once in the past week, and does not happen when observing the problem, and testing it consistently.


Sorry for Double-posting, but I decided to run a quick test with all benchmarks up and wanted all of your opinions on it. It seems the temperature goes from around 49 degrees to 85 or so degrees in less than a minute. Along with this, the GPU does suddenly get around 223 watts of power at the same time. Here's the video: https://www.youtube.com/shorts/1lDAU-bANFo

I believe it might be the thermal paste problem said then, is the temperature spike meant to be that drastic within such a short period of time?

85c is still safe for a GPU, but your hotspot temperature might be different. Are you able to monitor it with GPU-Z?

 

[Selaya]

oh, nvm then

yeah in that case it could be a faulty psu or something ig

 

[Cl0verr]

85c is still safe for a GPU, but your hotspot temperature might be different. Are you able to monitor it with GPU-Z?

Just tested. The hotspot temperature went from 66 to 99C in 30 seconds, and then when it reached 99.9 it immediately shut down.

Does this narrow it down to being a thermal paste issue?

To add to this - it seems the PSU is not the issue, as the power draw is completely nominal in the log.

 

[Thimblewad]

Just tested. The hotspot temperature went from 66 to 99C in 30 seconds, and then when it reached 99.9 it immediately shut down. View attachment 287165
Does this narrow it down to being a thermal paste issue?

To add to this - it seems the PSU is not the issue, as the power draw is completely nominal in the log.

Are your GPU fans operating correctly? Usually GPUs shutdown ad about 105 or 110°C, but yes, having a hotspot temp of over 90°C is usually not okay. It could be a thermal paste issue, could be bad seated cooler but nonetheless, try a different PSU.

 

[Cl0verr]

Are your GPU fans operating correctly? Usually GPUs shutdown ad about 105 or 110°C, but yes, having a hotspot temp of over 90°C is usually not okay. It could be a thermal paste issue, could be bad seated cooler but nonetheless, try a different PSU.

The fans seem to be working completely fine - I took apart the heatsink while troubleshooting and did a cleaning of dust just incase, and overall fans are still working fine. It might be partly a badly-seated cooler as well

 

[sLowEnd]

Just tested. The hotspot temperature went from 66 to 99C in 30 seconds, and then when it reached 99.9 it immediately shut down. View attachment 287165
Does this narrow it down to being a thermal paste issue?

To add to this - it seems the PSU is not the issue, as the power draw is completely nominal in the log.

Yikes. That's very hot very fast.

 

[mclaren85]

Typical GPU power cable problem. Can also be a sudden transient loss of power