Black screen, fans at full speed, new GPU installed. Feeling really down. Need help.
Black screen, fans at full speed, new GPU installed. Feeling really down. Need help.
Hello! I'm feeling really frustrated. I don't know what to do. My computer keeps rebooting too often. Sometimes the screen turns off and the fans spin at full speed for about 5 to 7 seconds, during which I can hear sounds from games and Discord, and others can hear me. Then the sound disappears and the screen goes black with fans running at reduced power (RGB lights on GPU and other parts are lit).
This happens randomly, while browsing the web or playing games—reboots seem more likely. The only solution is to press the hard reset button. It might take a couple of days, or it could happen 10 to 15 times in one evening.
It started after I swapped my GPU from the 3080 model to the 4090 Gigabyte Gaming OC. Temperatures are unusually low on both CPU and GPU (including VRAM).
Current system specifications:
- Motherboard: MB-Gigabyte Z490 Aorus Master
- GPU: 4090 Gigabyte Gaming OC
- CPU: Intel i9 10900K
- RAM: Ballistix 3600 64GB/4, 3600MHz
- SSD: 1 XPG GAMMIX PCIe 2TB NVMe, Samsung 970 EVO Plus, Crucial P5 2TB NVMe
- HDD: Toshiba 7200RPM, 4TB
- AIO: Arctic Liquid Freezer II 420
- Power Supply: EVGA Supernova G Plus 1300W
- Case: Fractal Design Meshify 2 XL
- Monitor: LG 34GN850-B
What I attempted to fix the issue:
1. Clean install of drivers via NVIDIA Experience (no improvement)
2. Used DDU with boot flash (no success)
3. Ran Syscheck (everything appeared normal)
4. Performed full memory test (took over a day, passed)
5. Replaced all PC components (including MOBO)
6. Double-checked every cable connection five times
7. Updated BIOS on both GPU and motherboard twice
8. Installed older drivers for the GPU
9. Updated all system drivers
10. Tried dummy Windows 10 Pro and Windows 11 Pro installations
11. Changed power modes in Windows settings (saving, performance, custom)
12. Disabled XMP profile on RAM
13. Switched to different power modes in Windows settings
14. Turned off XMP profile for stock frequency on RAM
15. Disabled auto CPU frequency adjustment in BIOS
16. Changed from DP cable to HDMI cable
17. Purchased a surge protector
18. Bought an APC LE1200 Line-R 1200VA Automatic Voltage Regulator
19. Tested various power adapters
20. Disabled hardware acceleration for GPU scheduling
21. Changed GPU mode from OC to Silent
22. Switched PSU from Seasonic Prime 1000W to Corsair 1000W
23. Updated PSU from Corsair 1000W to EVGA 1300W G Plus
24. Reviewed event logs (clean, except kernel errors from incomplete reboot)
25. Replaced CPU cooler from Noctua dh-15 to Arctic Liquid Freezer II 420
26. Ensured the PC build was correct and not incomplete
27. Tested with a different PC build
I concluded that my GPU is likely faulty because everything worked fine on the previous model (3080). I took it to a store, where they tested it and confirmed it was stable for 48 hours. I brought it back, and after 2 days everything functioned normally again.
I tested my rig with both CPU and GPU disabled, running stress tests, but found no problems.
My PC is my only comfort right now. I’m struggling emotionally. Please help me—I’m not a tech expert, just a healthcare worker. My knowledge is limited to these basic steps. I really need guidance.
Furmark places strong emphasis on the GPU, resulting in minimal shifts in power states. Have you performed this check overnight? This isn't an ideal scenario for assessing stability in low-power conditions. The PC shop should replicate your home usage—browsing, gaming, etc.—to verify performance. Try Furmark as they did and observe any crashes.
In NvidiaCP, adjust the power plan to maximum performance and monitor the results.
Check with Gigabyte; a firmware update might be necessary for the card. A replacement could be justified if needed.
Your issues started with this card, so it's unlikely to be unrelated. Your troubleshooting efforts have been commendable, but consider that the problem may stem from driver or hardware power concerns.
Hello! I used Furmark and 3d mark along with an OCCT test. I experienced reboots in both programs, but they occurred in the same way in Furmark and 3dmark at different times. Either there were no reboots or frequent reboots at different intervals.
I also installed a firmware update from the Gigabyte official website, which introduced two new BIOS versions, and I installed both.
Would probably stop persuing a fix for this card and try replace it. As you said, problems started immediately after installing it. Good chance there is something wrong with it, like many others.
View: https://www.reddit.com/r/nvidia/comments...ith_video/?
Writing to inform you about the identical problem I’m experiencing. Began a few days ago after updating to the latest drivers, and it occurs randomly while browsing the web or playing games as well. My crashes vary slightly; all screens turn black, but audio remains. All Corsair lights become pure white, and I can’t see any changes on the GPU. Sometimes it takes 10-20 minutes for everything to reset perfectly, exactly where I left off. Usually, it ends with me having to completely restart it to fix the issue. I can usually recreate it by adjusting display settings—like toggling HDR on and off repeatedly, switching in and out of exclusive full-screen apps, and other similar actions.
Interestingly, both of us have the same GPU (3080, mine EVGA, yours ASUS TUF) and the same CPU. I’m running out of ideas. It seems likely there’s a physical power cord problem with the 12VHPR connector, but it might be worth reaching out to the manufacturer now. Since we’re from different brands, I’m pretty sure it’s not a flaw specific to one brand.
I am genuinely sorry this has happened to you. Truly sorry. I understand your perspective. Maybe you're right—it could be the fault of someone else. I didn't attempt to use another one. Two days ago, I initiated a return. I hope they will replace it. If they do (perhaps in January) and the problem is resolved, I'll let you know here. If it continues, I'll try another cable, probably. If you act faster than me, please message me as well.
Okay, wanted to give ample time to make sure that it was solved. After sending my post and reading some other accounts on reddit, I investigated the power cable. My case (Lian Li 011 dynamic XL) is a really tight fit with the card, so the cables kind of press against the side glass. I went and took off the side panel, gave the cables some breathing room by making the angle less sharp (since the glass isn't there) and I have not had a single crash since then. That was on Friday night. Nothing has changed software wise, same drivers, same windows installation, all of it. The only thing that has changed was just me angling the cables differently. Did not even reseat. Just gave it a firm push into the card to make sure the connector was fully into the card. My side panel is still off and i'm waiting to see if i get another crash. If I do, I will update you here. Hopefully this solves your issue, I know firsthand how frustrating this is especially with such an expensive investment. Maybe try giving it a shot while you wait for RMA response.
Have a happy new year, and I really hope this solves your issue. Sorry that your going through this as well.
I also faced the same problems with a Corsair Crystal 570x and a 4080. I attempted several of the same steps OP suggested, but nothing worked. Read this, removed the side panel, and it's been a day without any black screens now. I think I'll wait for Cable Mod's new right angle adapter, which should let me reattach the side panel.
I'm glad I could assist you. If this comes up again in the future and the power cables touch your side glass or panel, I strongly suggest removing the side panel to give the cables space. I've been avoiding a crash for almost a week now since doing this. The standard NVIDA cable is too large for my Lian li o11 dynamic XL and Asus TUF 4090. Tomorrow I'll bring in a cable mod with a 3x8 pin that I hope will work if I leave the side panel on, because it's much thinner. This is just a temporary fix while the 90 angle adapter is expected to be delayed for several months. I'll post updates once I see results.
Also, anyone looking for a cable mod 12VHPR to fix this? I advise acting fast—cable mods are being upgraded at their main facility, which is delaying all orders until February-March. They still have some in stock, and I managed to get one before it sold out. There are other brands for cable replacements, but I personally trust the ones I mentioned most.