Black screen during gaming - 2080 Ti

[Prozn]

Hi there,

I am hoping someone has experienced and fixed this, or can give me some pointers on how to troubleshoot this issue further.

Specs: 8700K, Asus Strix RTX 2080 Ti, 16gb g.skill 3200 CL14 RAM, maximus x hero, EVGA SuperNova G3 850W.

tldr:

• Screen is turning black and receiving no signal after gaming/benchmarking for a while. System is not shutting off, can still hear discord voice chat.
• Power cycling the PC brings it back to life (as long as I let it sit turned off for 10 seconds or so).
• Have DDU cleaned and reinstalled nvidia drivers and nuked all my Asus software.
• Card is on a custom loop, die temps seem fine (60C when stress testing), although a bit warmer now ambient temps have increased.

Story:

Randomly whilst gaming or running heaven benchmark my screen is going black, monitor not receiving a signal. This typically happens after the card has had time to warm up a bit. I can still hear discord chat in the background, but game sounds stop. This happens at stock and overclocked clocks. I then have to hard power cycle the PC.

If I turn the PC back on straight away the screen remains black (don't even see the post screen), but the motherboard goes through all the usual post codes on the LCD display. If I wait a few seconds and then power back on, the PC will boot into Windows fine.

Event viewer has been showing a crash related to AsusGPUFanHelper around the time of the black screen. However I have now uninstalled all the ASUS software and still get the problem. There are also warning level entries about windows detecting a problem with hardware.

I have DDUed my nvidia drivers in safe mode, and reinstalled the latest nvidia driver. But still get the black screen.

At this point I'm not sure what else to try. My card is water cooled on a custom loop - the die is hitting 60C under heaven benchmark load, 50C under normal gaming. Water temp 38-40C when saturated.

The possibilities I can think of are:

• Memory chips overheating - this seems the most feasible, but I'm not sure how I would check? Would the chips overheating cause the card to shut off like this?
• Standard RTX hardware issues - I am not getting any artefacts, game freezes, or anything else that would suggest a GPU failure. A possibility though...
  
• Nvidia driver bug - but this seems unlikely.
• Insufficient power available - I have a 2 year old 850W EVGA G3 PSU, supplying power to the card via two separate cables. Draw at the wall for the PC is about 500W. So I don't think this is the issue.

Any suggestions or tips are welcome. Thank you.

 

[Prozn]

Some more googling has shown other possible fixes/issues:

• Corsair iCue / Link software, especially when used alongside HWInfo64 has been shown to cause black screens. I use both iCue and HWInfo.
• Disabling Intel VT-D can fix the issue.
• Installing nvidia driver 419.67 can fix the issue.

 

[PhoenixSP]

Just to check is the audio from the monitor and which monitor is it ? A cheap potential fix would be to try different cables ? Maybe try underclocking the card to stock values for the 2080ti and seeing if there is still a problem.

 

[Prozn]

Audio is coming via a USB headset.

• I uninstalled all iCue and monitoring software, still get the crash.
• I ran heaven benchmark from a boot with my water at ambient. The screen went black about 30 minutes in, when water temp hit 40C, core 52C, VRM 60C.
• I ran heaven benchmark again, with HWInfo64 loggin enabled. This time it lasted 20 mintues. Logging stopped as soon as the screen went black, which rules out a dodgy cable.

I played CS:GO all night without any crashes with FPS limted to 200. With FPS unlimited I get black screens.

Tomorrow I will try rolling back to 419.67 drivers, but based on limiting my FPS (thus keeping memory load and temps down), it seems to me there is a problem with the card under prolonged load. Either overheating of the memory or some other component, or a faulty memory module.

My GPU VRM is hitting 60C when the black screen occurs, but I don't think that is too hot for the VRM?

This is all at stock clocks.

 

[Funky-Melon]

It's more than likely a card issue.

I had this on a GTX 970 before I gave up and replaced it.

It would be fine during desktop use, fine on older less demanding games, would never crashed, but on more demanding games the screen would just go black and nothing would recover it unless I reboot the machine.

Sometimes it died 10 minutes in, sometimes an hour in but it happened enough to be a problem.

I even gamed with the fans on 100% in case it was sensitive to heat but it did the same.

Time to RMA and hope that they can replicate the problem....

 

[Prozn]

Mine seems a little different, it only ever seems to happen when my watercooling loop reaches a certain temperature, although that is partly wishing thinking hoping to avoid an ASUS RMA. I can't really be without my PC for that long...

 

[Funky-Melon]

I've had some pretty hot processors before and they haven't caused that issue

 

[Kelt]

Mine seems a little different, it only ever seems to happen when my watercooling loop reaches a certain temperature, although that is partly wishing thinking hoping to avoid an ASUS RMA. I can't really be without my PC for that long...

Asus RMA?

May whatever lord you believe in be with you.

 

[Prozn]

Didn't want this £1200 GPU anyway...

 

[Prozn]

• Underclocked memory, still black screen after 30 minutes
• Underclocked core and memory, ran heaven for an hour no problem
• Overclocked memory with core still underclocked, ran heaven for an hour no problem

So that seems to rule out dodgy or overheating memory. It seems odd that after running stable at 2090mhz since launch the card would suddenly not even be able to maintain stock boost clock. At this point I'm suspecting maybe a PSU issue... my PSU fan was sounding like a jet engine during the benchmarking implying it was fairly heavily loaded/hot.

Still all wishful thinking hoping to avoid ASUS RMA

 

[jupiter_man]

would it be the HDMI / display port cable problem?
I used to have a 2 seconds black screen then everything back to normal, then I bought a expensive display port cable and then I rarely have this issue again.

 

[shankly1985]

I have the same motherboard and I know they has been a few bios update releases. Maybe update the bios if you not already done so.

 

[azazl187]

You should RMA the card or try it in a completely different rig

 

[Chrisc]

Try a different PSU, I had a similar issue with a crap Corsair RM1000, the screen would go black and I could still here people on discord for roughly 10seconds. I could also get half hour or so out of superposition stress test.

 

[Prozn]

So a little update: I have replaced my EVGA G3 850W with a Corsair HX1200i (I went super overpowered so that the fan didnt turn on) and so far I've had no black screens. A little bit confusing as the G3 is supposedly one of the best PSUs?!

My apartment is about 5 degrees cooler at the moment however, so it hasn't completely ruled out memory overheating I guess.

would it be the HDMI / display port cable problem?
I used to have a 2 seconds black screen then everything back to normal, then I bought a expensive display port cable and then I rarely have this issue again.

I'm fairly confident I have isolated the problem away from a dodgy cable now. It was my first thought too.

I have the same motherboard and I know they has been a few bios update releases. Maybe update the bios if you not already done so.

BIOS is up to date already (thanks for the microcode patches!), so think I'm in the clear there.

You should RMA the card or try it in a completely different rig

My last resort

 

[hyperseven]

What monitor?

 

[azazl187]

So a little update: I have replaced my EVGA G3 850W with a Corsair HX1200i (I went super overpowered so that the fan didnt turn on) and so far I've had no black screens. A little bit confusing as the G3 is supposedly one of the best PSUs?!

My apartment is about 5 degrees cooler at the moment however, so it hasn't completely ruled out memory overheating I guess.



I'm fairly confident I have isolated the problem away from a dodgy cable now. It was my first thought too.



BIOS is up to date already (thanks for the microcode patches!), so think I'm in the clear there.


My last resort

Sounds like it's sorted to me

 

[Prozn]

What monitor?

Alienware AW3418DW 34" Ultrawide

Sounds like it's sorted to me

Just happened again this morning. Oddly, it happened exactly when I opened the front panel of my case while under full system load, which will have cause a little jolt to the components. This happened once before and I assumed it was coincidence. I have double checked and all my cables are seated firmly

Everything was still running in the background, could hear the game music, whatsapp web messages coming through etc, I could even remote into the PC but the remote session was black too.

Event Viewer shows:

Faulting application name: dwm.exe, version: 10.0.17763.1, time stamp: 0xe52aabf3
Faulting module name: udwm.dll, version: 10.0.17763.404, time stamp: 0xc4744ab5
Exception code: 0xc0000005
Fault offset: 0x0000000000024149
Faulting process ID: 0x1224
Faulting application start time: 0x01d516d18cc80dd9
Faulting application path: C:\WINDOWS\System32\dwm.exe
Faulting module path: C:\WINDOWS\SYSTEM32\udwm.dll

At this point I'm thinking it has to be a hardware issue causing desktop window manager to crash...

 

[VonScar]

I had issues with my old card, the MSI GTX 970, it was so heavy and was dropping, you had to fiddle with card to ensure it was fully in the PCI-e slot.

Perhaps you should try another PCI-e slot, if you have one or re-seat the card and see if there is any warping of the card itself?

 

[scubes]

My arous did the black screen thing sent it back for rma got a zotac stellar preformer. Cool as well.