Shows black, RGB fans stop working but motors keep spinning, need to switch power supply – after installing the 7800xt.
Shows black, RGB fans stop working but motors keep spinning, need to switch power supply – after installing the 7800xt.
I recently upgraded to a 7800xt from a 2070, added 16GB of RAM (same sticks, properly seated). I ran DDU in safe mode, connected a Thermaltake 850W Grandpower PSU, and used two PCIe cables instead of one long daisy chain. Still, my system freezes into a black screen during intense games like Escape from Tarkov and Arma 3. Sometimes it runs for hours without a crash, but other times it crashes repeatedly in just ten minutes, always in the same area of the game. It seems like rendering is the culprit. Each crash causes my display to go dark, the tower fans stop, but the RAM fans keep spinning. The power button won’t shut it down; it only turns off via the PSU switch. After this incident, my CPU debug light activated. I don’t think it’s a CPU problem—my 5800x has worked fine for two to three years without issues. This only started after I changed the 2070 to a 7800xt. I followed proper installation, but I did undervolt the GPU using AMD software and it still crashes. I’ve stress-tested the GPU with FurMark, and here’s some odd data: https://imgur.com/a/AfWCFne. Every metric plummets to near zero every seven seconds. What’s going on? It definitely doesn’t seem normal. I’m sure it’s related to power delivery, but I’m unsure if it’s the PSU or the GPU. I have no clear solution left. Any advice would be greatly appreciated!
Specs: 5800x 7800xt Thermaltake Toughpower Grand RGB 850W Gold Fully Modular (6 years old) B450 Tomahawk Max (4 years old) BIOS version 7C02v3A 4x 8gb G-Skill DDR4 3600 Windows 10 version 10.0.19045
I'm also facing issues with my new 7800XT. After rebooting, the BIOS sometimes fails to recognize it, but once it boots up, everything works properly. I've read many reports about other users experiencing similar problems with their 7800XTs not functioning correctly. It seems the motherboard isn't detecting the card properly, and the only consistent solution is to gently massage it before powering it on. I'm curious if these issues might be connected to the chiplet design, with variations in die thickness affecting cooler contact. Would you mind testing my theory and giving me a try by lightly squeezing the card to see if it makes a difference?
I experienced the same problem on my machine too. It eventually became my power supply unit. I plan to run a stress test using Cinebench and possibly a stress test through the Radeon software at the same time (ensure it's installed and remove any Nvidia drivers). If that doesn’t cause it to shut down, it might be the PSU. I was wondering why it would fail only under certain conditions—especially during specific game levels. When a PSU deteriorates, it can sometimes function but will fail at high loads, like during intense gameplay. This seems like the most probable explanation based on what you shared. Before replacing the PSU, run a memory test: https://www.memtest86.com/ and verify CPU temperatures, as those are the most likely causes of a total shutdown.
I swapped out my power supply today—it was outdated. My CPU temperatures stayed stable under any load, thanks to the strong cooler. The GPU kept stuttering all day, and I spent a lot of time testing fixes. After taking out the new RAM I got for Christmas, the issue persisted. I booted into safe mode and reinstalled the latest drivers, but it didn’t help. I reinstalled my old RTX 2070 with the correct drivers, and the problem disappeared. I removed it again, updated the BIOS (which was only a couple versions old), and fresh AMD drivers, but the issue returned. I ran Unigine Heaven under stress and FurMark afterward, and everything seemed normal. Then I started playing Escape from Tarkov, and the stuttering returned. I suspect the problem isn’t just from Tarkov—it also occurs in Rainbow 6 Siege during practice runs. Oddly enough, using FurMark seems to be the culprit. It began after I tried it for the first time yesterday. I’ve heard it’s bad for GPUs, so I plan to clean drivers again and share results using only Unigine Heaven 4.0.
It seems the black screen issue might be linked to GPU instability or driver problems. Some users reported erratic clock speeds on the 6600, which could explain the problem. The concern about GPU defects is valid, though it’s hard to pinpoint without more details.