Hello Everyone,
 
I am having the black screen 100% fan issue with my 3090 FTW3 Ultra (24G-P5-3987-KR). It has happened a few times while playing Call of Duty Cold War multiplayer over the past few months. I didn't think it was a major issue so I ignored it. When I started playing the New World beta it happens every time I play. Usually after about 15 minutes of playing.
 
After it happened the first time in New World I saw all the news about it and uninstalled the game. I tried it again today to see if the game update helped at all. I set the graphics settings to High and FPS limit to 60FPS for the play session. Took about 15 minutes for the black screen issue to occur. HWMonitor temps looked fine before the issue happened. I was playing with my case side panel off. I was not in a populated area.
 
Luckily, the card still works after I do a hard reboot. I am very concerned though since this is happening with multiple games.
 
System Information:

• I have zero overclocking set on my CPU or GPU
• Latest Nvidia drivers
• Latest AMD chipset drivers
• Windows fully up to date
  • 21h1
• No RGB or Fan Controller software running
• CPU: Ryzen 5900x
  • DEEPCOOL Castle 360EX
• Motherboard: Gigabyte X570 Aorus Master
  • Latest BIOS
• Video Card: 3090 FTW3 Ultra (24G-P5-3987-KR)
  • Firmware: 2.01.21
  • Re-Bar enabled
  • All stock settings
• Power Supply: EVGA SuperNOVA 750 G2 220-G2-0750-XR 80+ GOLD 750W
• Ram: G.Skill Trident Z Neo Series 32GB (2 x 16GB) 288-Pin SDRAM PC4-28800 DDR4 3600 CL16-19-19-39 1.35V Dual Channel Desktop Memory Model F4-3600C16D-32GTZNC
  • Running in XMP mode
• Monitor: Acer Predator XB1 XB271HU bmiprz (2k monitor)
  • Gsync enabled
• Computer case: LianLi O11 Dynamic
• Ambient Room temperature: 68 Degrees F

 

oceansong34
Hello Everyone,
 
I am having the black screen 100% fan issue with my 3090 FTW3 Ultra (24G-P5-3987-KR). It has happened a few times while playing Call of Duty Cold War multiplayer over the past few months. I didn't think it was a major issue so I ignored it. When I started playing the New World beta it happens every time I play. Usually after about 15 minutes of playing.
 
After it happened the first time in New World I saw all the news about it and uninstalled the game. I tried it again today to see if the game update helped at all. I set the graphics settings to High and FPS limit to 60FPS for the play session. Took about 15 minutes for the black screen issue to occur. HWMonitor temps looked fine before the issue happened. I was playing with my case side panel off. I was not in a populated area.
 
Luckily, the card still works after I do a hard reboot. I am very concerned though since this is happening with multiple games.
 
System Information:

• I have zero overclocking set on my CPU or GPU
• Latest Nvidia drivers
• Latest AMD chipset drivers
• Windows fully up to date
  • 21h1
• No RGB or Fan Controller software running
• CPU: Ryzen 5900x
  • DEEPCOOL Castle 360EX
• Motherboard: Gigabyte X570 Aorus Master
  • Latest BIOS
• Video Card: 3090 FTW3 Ultra (24G-P5-3987-KR)
  • Firmware: 2.01.21
  • Re-Bar enabled
  • All stock settings
• Power Supply: EVGA SuperNOVA 750 G2 220-G2-0750-XR 80+ GOLD 750W
• Ram: G.Skill Trident Z Neo Series 32GB (2 x 16GB) 288-Pin SDRAM PC4-28800 DDR4 3600 CL16-19-19-39 1.35V Dual Channel Desktop Memory Model F4-3600C16D-32GTZNC
  • Running in XMP mode
• Monitor: Acer Predator XB1 XB271HU bmiprz (2k monitor)
  • Gsync enabled
• Computer case: LianLi O11 Dynamic
• Ambient Room temperature: 68 Degrees F

 

Which reversion number is your card?

Hi,

Same thing happened to me.

Here is my specs.

Gaming PC:

- Case Segotep Phoenix ATX - Exhaust fan 2 x LL140 Corsair RGB fans
- AMD Ryzen 9 5900X CPU
- Corsair H150i 360mm Rad - 3 x LL120 Corsiar RGB fans
- MSI MEG X570 Ace Gaming Motherboard
- 64gb - 4 x 16Gb Corsair Vengance RGB Pro 3200Mhz
- EVGA GeForce RTX 3090 FTW3
- Samsung 970 Evo M.2 EVO Plus 1Tb (OS + apps)
- HP EX900 M.2 500gb (games)
- Samsung 860 Evo SSD 500gb (misc)
- Corsair HX1000i fully modular PSU
- 2k 3440x1440 gsync monitor


Capped my fps to 60 in nvidia control panel.
Hwsinfo open with precision x1 open ICX enabled on second monitor

Installed New World tonight.

Played for about 1 hour. Temps were about 65 on ICX

2 screens went black, still hearing game sound and gpu ramped up to 100% fan. As soon as it happened, hard shutdown the pc. No smell , all good.
Waited for about 5 minutes. Rebooted and posted and in Windows 10.

Played again for about 1h30 and same issue happened.

This time, when gpu ramped up, sound of like... clicking and weird sound happened before I was able to hard shutdown.

Rebooted after 5 minutes all good.

Didnt play since them.

I had no issues with any games prior to New World.

I am now super worried that I actually damaged the gpu.

What you think?

oceansong34

kevinc313
PSU is hopelessly undergunned.  G2 is a Super Flower Leadex I, good platform, but is running at ~600w estimated with a 420w 3090, 5900X on up to 165fps game play load maybe 100w+, plus mb and components.  When the 3090 throws crazy loads at the PSU it doesn't get a good result back and eventually trips the card's protection. 
 
I'd recommend a 1000w minimum in a Leadex platform PSU.  People have been reporting good results lately with the Corsair RM850x. Use OEM cables with no extensions or daisy chains.
 
Try turning down card power limit to 70% (300w) and capping 100 FPS to limit cpu load, see it that stops the shut off.
 
I'd also be curious about the core and vram junction temps in the O11, as it is know to have marginal air flow across the back of the card and limited space for venting the top of the card.

Thanks for the replies.
I believe it is the Rev 1.0. Didn't see it on the card anywhere, but it has the black bezels, not red.

My third RMAed card is Rev 0.1 with black bezels. You can check max readout of PCIe Slot Power in GPU-Z or HWinfo to identify revision number, max < 67W is revision 1.0.

oceansong34
 
I was curious about my PSU wattage. There was one thread which I can't find now that had mine listed as "known good/working". XStarwake above appears to have the same issue as me with a beafier PSU.
 
EVGA support contacted me and asked to monitor 12v rail. During a Time Spy benchmark no fluctuation and steady at 12.312v.

Yeah, I also had a problem with a 1300w Seasonic Prime and 3080 Ti FTW3, but it was actually the PSU shutting off to protect itself, not the card. 
 
Turn down your power / FPS and test.

GameArk
I have the same issues, however, I can play a little longer like 40 to 60 min thn black screen. The only way to get back hard restart. Please fix it. I really like the game New World. 

Set an FPS cap and turn down the GPU power.

kevinc313

XStarWake
What you think?

I wouldn't push your luck, but I'd use HWiNFO with Riva Tuner OSD on a 100ms polling period to monitor current and MAX of GPU Use, Power, Core temp, VRAM JUNCTION TEMP, hotspot temp, fan speed, FPS, vcore and anything else you feel like.  Then MINIMUM and current of the 8-pin, PCIe and rail voltage.  Reset after starting game.

That is hard to understand for me. 
 
I run HWiNFO to check temps and all fine as VRAM junction temp, Hotsop temp. Always under 65celcius.
i use XP1 ICX to monitor Fan speed, vcore and such. 
I dont know how to check the MINIMUM and current of the 8-pin, PCIe and rail voltage and 100ms polling period to monitor current and MAX of GPU Use, Power, Core temp. This is out of my confort zone as I dont understand at all. 
 
thanks for your reply tho. I beleive, since I brought the card in dec. 2020, its revision 0.1.
 
I played, since dec. 2020, all the new games and never did something like this happened. 

XStarWake

kevinc313

XStarWake
What you think?

I wouldn't push your luck, but I'd use HWiNFO with Riva Tuner OSD on a 100ms polling period to monitor current and MAX of GPU Use, Power, Core temp, VRAM JUNCTION TEMP, hotspot temp, fan speed, FPS, vcore and anything else you feel like.  Then MINIMUM and current of the 8-pin, PCIe and rail voltage.  Reset after starting game.

That is hard to understand for me. 
 
I run HWiNFO to check temps and all fine as VRAM junction temp, Hotsop temp. Always under 65celcius.
i use XP1 ICX to monitor Fan speed, vcore and such. 
I dont know how to check the MINIMUM and current of the 8-pin, PCIe and rail voltage and 100ms polling period to monitor current and MAX of GPU Use, Power, Core temp. This is out of my confort zone as I dont understand at all. 
 
thanks for your reply tho. I beleive, since I brought the card in dec. 2020, its revision 0.1.
 
I played, since dec. 2020, all the new games and never did something like this happened. 

Time to start learning.  
 
There's no way your hot spot and Vram junction are under 65C with a gaming load.
 

yangning8109
 
My third RMAed card is Rev 0.1 with black bezels. You can check max readout of PCIe Slot Power in GPU-Z or HWinfo to identify revision number, max < 67W is revision 1.0.

Looks like I have a 0.1 Rev Card. PCIe Max Slot Power was above 67W. I got it last November and SN starts with 2014.

Was reading other articles about the issue and wanted to note two other observations that I also noticed on my end.

• New World FPS during loading screens went over the 60FPS limit. I had the game's FPS monitor on and remembered it being over 200FPS during a load screen.
• The last time I played before the black screen, I had changed my Quality settings to High (down from Very High). Wasn't long after that the black screen occurred. I know that I did the same thing the first time I played the beta, but can't remember how long after change the black screen occurred. I wasn't hyper aware of the issue the first go around. 

KeithW19

tyranus7
At this point, I'm like 80% convinced that Rev. 0.1 has been the one that has been dying. Probably because the PCIe port over power issue. 10 A fuse there exploding as framechasers said:
 
https://www.youtube.com/watch?v=ZlzhaYosci0
 

I'm wondering if he is only half right. I think his reasoning for the failures is correct but he relates it to the uncapped menu fps which I dont believe is the cause. I'm thinking its the fan issue where for some reason it gets asked to spin at 200,000 rpm and that sudden power surge is what causes the fuse to blow

Fan can't do that.  It's PWM on a fixed 12V. There's no exceeding 100% PWM.

KeithW19 article from PC Gamer 

Come on man.  They are idiots and don't understand what Igor is saying, and his theory is dubious at best.  There have been multiple instances of mis-reporting of telemetry in GPU-Z or HWiNFO with this gen.  Try plugging a fan into straight 12V and see what happens.  OK?  Nowhere on the card is there above 12V, it is all stepped down from the incoming 12VDC. 
 
Now, could there be something disastrously wrong with the PWB signal generation?  Maybe.

KeithW19
Good point. It probably is wise to ignore plausible explanations because of your bias against the source.
Fact of the matter is that if that guys theory was correct then Amazon's "Fix" would have prevented further failures and it hasnt. So something else is going on. I had my menu capped at 60 fps (by Amazon) and had my in-game capped at 140 fps and this issue affected my card and killed it. So either capping fps is not the fix and the card still runs amok in the background or there is something else causing a fuse to break.
I'm curious where this fuse is and if I can check mine to see if its blown

Sorry man, you're out of your depth here.  No such thing as a PWM fan overspeed, on 12V.  Totally IMPLAUSIBLE explanation.
 
The vast majority of the PC press is filled with idiot journos who can't get a real job, has nothing to do with me being prejudiced against the source.

KeithW19
Not really. I'm just not closing off possibilities. What is happening is clearly not expected behavior so something is not working as it should. The fan thing may just be a symptom of the bigger problem (and likely is) but with all that has been done with fps caps, etc. cards should not be failing anymore and they are. I've been googling trying to see if there is a way to check this fuse but I cant find anything on it. It may not be something visible anyway but I'm very curious

Like I said, there very well could be a catastrophic failure of the PWM generation, or of the fan controller system.  But it IS LITERALLY IMPOSSIBLE for the fans to exceed their rated speed on that card unless it gets struck by lightening. 
 
For PCGamer to be saying they are overspeeding means they are idiots and don't understand a fundamental concept of PC setup, they should not be writing technical articles.
 
 

Just to be clear when you say rated speed did you mean voltage? I dont think anyone is saying the fans are hitting 200,000 rpms (or even going past their rated speed). What they are saying is that its being requested. The fan could never go that actual speed but that wouldnt stop it from asking for power to do so or like I said... thats a symptom of a bigger issue where the fan *thinks* it needs to go that fast due to some other hardware issue.
 
My main concern is about the replacement card from EVGA. Are all replacement cards not having issues with the game (that would suggest rev 0.1 vs rev 1.0 is the fix). I also firmly believe the issue is known but nobody is saying anything. Lots of speculation (including myself obviously) but no hard facts