Signals drop once monitors are used for GPU-heavy games for roughly 30 to 45 minutes.
Signals drop once monitors are used for GPU-heavy games for roughly 30 to 45 minutes.
I recently changed from a RTX 3060 to an RX6800XT, which caused some early problems where the system would shut down completely. At that time, I was using a 650W MSI power supply, so I thought the issue was with the PSU drawing too much power. I replaced it with an 850W Gold-rated unit from FSP. After the PSU change, another problem appeared: my monitors would lose signal while the PC stayed on, fans spun, RGB stayed active, and all peripherals except the monitor remained powered. I found discussions about RX6800XT issues, mentioning a Kernal 411 problem and some users needing to return their cards for repairs. I saw others report RMA requests due to error codes, but I simply swapped the card for a 3070Ti without resolving the signal loss. The issue resurfaced even after using new power supply and slightly used GPUs, requiring a hard reset for the monitors to regain signal. I’ve tried DDU drivers for both AMD and Nvidia, updated BIOS, but the problem persists. It only occurs during intense graphical games like Black Myth Wukong or Cyberpunk 2077, not during normal play or when GPU usage is low. Here are my current PC specs:
Specs:
- Ryzen 5 5600X RTX 3070Ti Gaming Trio
- 8GB MSI B450 Gaming Plus Max
- Corsair Vengeance 2133 MHz
- 32GB (2x16GB) FSP Vita GM
- 850W Gold-rated PSU
Use Nvidia and AMD together simultaneously. Start DDU in safe mode. In the settings, enable "prevent windows update from installing drivers". Follow the instructions: "clean and don't shutdown first and then clean and shutdown for installing new GPU". For the 6800xt, try Adrenalin 24.12.1. Ensure BIOS is set to enable rebar. Verify fast boot is configured the same way in both BIOS and Windows—either enabled or disabled—to avoid crashes and GPU removal.
Rebar was turned on throughout this process, so I believe we can exclude it from the possible reasons for the signal loss. I've performed a complete graphics driver reset, which should help. It's possible a leftover driver from the initial uninstallation is still causing the problem. Since my PC was functioning normally before the GPU change, I'm not sure what might be triggering it now. The only recent additions are the GPU and PSU, so I'll try playing Black Myth Wukong to see if the issue recurs.
It seems the PC encountered an issue similar to others, though not under load this time. I was in a loading screen in Forza Horizon 5 when it happened. Checking the GPU device manager logs didn’t reveal any error codes. I’m confused—especially with my 6800XT it showed an error, but with the RTX 3070 Ti I’m seeing nothing. Could be a faulty PSU?
I don't believe it's due to a bad PSU, it might be the PCIe rail damaged but that's not typical in my opinion. If the PCIe section of the PSU is faulty, it would likely cut power to the GPU completely. Since you had the same problem with a previous PSU, I’d rule that out. Have you changed the PCIe slot for the GPU just to try another one, in case it's the motherboard? I don't think testing it would be worth it.
Open Event Viewer, select the system view, and search for critical errors.
You opened the PC and saw your PSU's intake fan pointing up into the case. After flipping it over, you've been testing Cyberpunk with perfect ray tracing for an hour and a half. It's surprising you didn't consider the PSU's orientation—it was drawing in too much heat, causing it to reset itself. You've been dealing with this problem for three days, and fixing it took just three minutes with a screwdriver. Turning the component 180 degrees solved the issue. Thanks for all the support!
Consider additional tests before drawing any conclusions. Swapping the PSU in this manner usually poses no significant problems. I spent several years working at a PC shop, where we frequently adjusted systems this way to stop users from obstructing the PSU intake with dust and pet hair.