pc randomly black screen and gpu fans on full speed *SOLVED*

[Ratchet&Clank]

I had a faulty RAM-Slot on my board, so I removed the RAM from the Slot, I thought this would fix my issue but it didn't.

My PC is giving me a black screen (doing specific things increase the chance for it to happen)
I had it like 2 times in idling and often when doing specific things. For Example when I get energy spikes from compiling shaders or when rendering fog/poison clouds ingame (low performance consumption peaking) the chance is rather high this happens. I checked Core temp from GPU and CPU both looks good, GPU 50° and CPU 60°.

I also had after all the crashes microstutters then I had to turn off the pc and make sure the whole mainboard had no energy stored which lets me still suspect it's some cache issue but I just can't locate my problem.

I also think it could be my power supply but it worked since 2018 without issues until 1 year ago, this is when it started and now it's literally present every day. I just hope it's not my GPU

Specs:

Power Supply: corsair CS450M
Mainboard: MSI Z390-A PRO
GPU: MSI RTX 2080
CPU: I7 8700 3,2ghz
RAM: 4x DDR4 8GB G.Skill 3200C(2 removed atm still have the issues)
disk: SAMSUNG 850EVO 250GB & 500GB, 2x 860 QVD 1TB
Win 10 Pro 22H2

[Phénomènes Mystiques]

I had this happen after an April .Net Framework update and I saw another one last week so it could be that optional update but I am not sure but also the past three Win10 updates have been broken (also not sure if this corresponds to your issue) - here are two of the three broken Win10 Updates:

https://www.windowslatest.com/2024/11/03/microsoft-confirms-issues-in-windows-10-kb5045594-kb5044273-mandatory-update

Remembering back to a time long ago, I believe my BIOS was outdated/broken/at fault and did the same PC screen blackout and GPU fan/case fan rev up but there was also someone long ago who told me that is was their RAM which was at fault - I am only a messenger.

[Ratchet&Clank]

Originally posted by WinterSolstice:

I had this happen after an April .Net Framework update and I saw another one last week so it could be that optional update but I am not :100percent: sure but also the past three Win10 updates have been broken (also not sure if this corresponds to your issue) - here are two of the three broken Win10 Updates: :csd2smile:

https://www.windowslatest.com/2024/11/03/microsoft-confirms-issues-in-windows-10-kb5045594-kb5044273-mandatory-update

I doubt it's and OS related issue and those people said programs and stuff stopped running but my system is crashing to black screen

[Missing Spartan]

OP get a new power supply! your running your entire system on a 450 watt psu, they degrade over the years. that system should atleast be on a 650W psu

[Ratchet&Clank]

I still had a 500 psu in reserve, it's working for now but I need to test it for at least a week to be sure, I opened the 450w psu and checked the capacitors and other parts by view but couldn't see any real degradation or bloating, I hope it's really just the psu we'll see until next week if it works out.

[_I_]

corsair CS450M

this is way too small for a 2080
nvidia recommends 650+w for it

the 2080 has a pci-e 6pin and 6+2, the cs450m only has 1x 6/6+2 connector

[Ratchet&Clank]

Originally posted by _I_:

corsair CS450M

this is way too small for a 2080
nvidia recommends 650+w for it

the 2080 has a pci-e 6pin and 6+2, the cs450m only has 1x 6/6+2 connector

I was supplying the other over a power adapter making 2x4pins to the 8pin connector

[_I_]

if the psu does not have the connectors needed, its too small
never use adapters like that, and dont use chained connectors for pci-e

[Illusion of Progress]

Agree with the others, with an issue like this, the PSU is usually one of the first things to test because you need to ensure you have good power first and foremost. if it was a good PSU and you had reason to believe the PSU wasn't causing it, then because of the symptoms, I'd say maybe skip that and the next assumption would be the GPU. But if it's a not-so-great 450W unit, I'd start there.

Shader compilation is usually heavy on the CPU. What's likely is that the PSU wattage capacity isn't coping well with the spiking CPU use plus GPU load. I'm surprised it's not shutting down though. Usually, the loss of video signal and the fan at 100% is the GPU in "panic mode". Could be bad GPU itself, driver behavior, or stemming from the aforementioned power issue. this is why it's important to start with ensuring you have good, and enough, power.

What happens if you run Prime95 (small FFTs) with either Fur Mark or an OCCT GPU test at the same time? Does it also shut down or cause the GPU to do the same thing then?

If we reach a point that the PSU and GPU are not the issue, I'd say maybe motherboard (maybe VRMs giving out) since you say it happens specifically with shader compilation, or CPU itself.

Worth testing RAM too though. It's not something I'd suspect near the top of the list, but it's easy to test and rule out.

[ˢᵈˣ FatCat]

your GPU is the Costume Card from MSI? Like Strix from Asus,
If it is, they usually dont have power limit, which could draw power more, so it can boost its clock which give your PSU struggling,

Any games, or specifiy game?
It's really hard to decide about this black screen crash, if it was PSU not enough juice, the pc will totally shutdown, like blackout says Random tech support from seasonic.
And, the GPU Pcie slot power will blinking Red

[Ratchet&Clank]

Originally posted by ˢᵈˣ FatCat:

your GPU is the Costume Card from MSI? Like Strix from Asus,
If it is, they usually dont have power limit, which could draw power more, so it can boost its clock which give your PSU struggling,

Any games, or specifiy game?
It's really hard to decide about this black screen crash, if it was PSU not enough juice, the pc will totally shutdown, like blackout says Random tech support from seasonic.
And, the GPU Pcie slot power will blinking Red

I had this happen in 7 days to die when a zombie cop was spitting on me with a roughly 30% chance especially when I was looking at point blank of the spit which is some fog rendering or whatever this is called. And I had it now on stalker 2 with compiling shaders triggering it more likely than anything else. Also had one crash idling on desktop between all the crashes when trying to compile stalker 2 shaders. After PSU change I have nothing so far but it's just day 2 now so we have to wait ...

[Ratchet&Clank]

After 4 days of switching from 450psu to 500 and no more crashes I'd say the psu gave in, which is still odd since I tested my system on full power benchmark years ago and it showed the power consumption peaking at 300W so I expected the 450 psu to do fine. It could be that the 2x4 pin adapters I put together for the GPU weren't dimensioned to support the GPU with that much wattage, whatever the reasons are ... all I can say is, usually try the PSU first because they just seem to suck balls if you bought something cheap like me.

[ˢᵈˣ FatCat]

I used to think like that, oh mine peaked 300watt, so i think 650 watt is just do fine, work flawlessly for 3 days, but then black screen + my PCie power blinking Red, after that i changed to 850watt, crashed still lol but different reason, RAM was the culprit, BUT after i upgraded my cpu cooler , even using 3200mhz RAm no more crash,

Guess auto OC using stock cooler was bad

[Bad 💀 Motha]

Black Screen w/ GPU Fans @ 100% basically means the GPU Driver crashed and couldn't recover.

Wipe everything for all GPUs and Displays out using the DDU app; boot into safe mode when you actually do any cleaning with it.

Then reboot and install very latest non-Beta GPU Driver.

If still a problem, try using something like MSI Afterburner and under-volt and/or down-clock the GPU to gain stability.

[Phénomènes Mystiques]

DDU gets people in trouble, you should not recommend this app because I had someone mess up their drivers with this thing - DDU is NOT safe for regular consumers to use.

[Ratchet&Clank]

I want to give an update that my PC is still running flawless since my exchange from old 450W PSU to 500W PSU, it really seems like the PSU was the culprit all along(my first black screen was more than 1 year ago). The only assumption that makes sense for me is that some parts of the PSU got dry or stopped working correctly giving a crappy output and thus crashing the GPU under stress. My PC has a high uptime and I had the PSU running my system since 2018. My Problem is solved and can be closed. Always check the ♥♥♥♥♥♥♥♥ PSU first ... this guy is nuts.