My monitors randomly go black while gaming, watching videos, or even browsing. The computer itself stays powered on and does not restart, but I have to reboot it manually to restore the display.
My system has an Asus Phoenix RTX 3060 12GB, Ryzen 7 5800X, ASRock X570 Velocita motherboard, 32GB of Corsair Vengeance RGB Pro memory, and a Cooler Master V750 Gold V2 power supply. I use two Samsung Odyssey G5 monitors over DisplayPort, with an HDMI connection to a TV that is usually turned off.
I have already reinstalled Windows, replaced the monitor cables, tried both GPU power cables included with the power supply, and tested both 6+2-pin connectors on the daisy-chained GPU cable. A repair shop tested my graphics card in another computer and tested a different graphics card in my system, but no crashes occurred during their testing. Once I brought my computer home, the problem returned.
The issue also seems sensitive to movement. When I barely touched the unused hanging connector on the GPU power cable, both monitors immediately lost signal while the PC remained on. However, the problem can also happen when I am not touching the case, usually somewhere between 30 minutes and two hours into use. What should I check next, and could the GPU power socket, cable terminal, or modular PSU connection be damaged?
2 Answers
Since the display cables and hardware tests have not isolated the cause, try narrowing down the power and display paths one at a time. Connect the PC and monitors directly to a wall outlet instead of a power strip, test different outputs on the graphics card, and temporarily disconnect the TV's HDMI cable. Also check whether the problem appears in Event Viewer after rebooting, although Kernel-Power 41 generally only indicates that Windows did not shut down cleanly rather than identifying the original fault.
A clean boot and timing log may help, but the random intervals make this less likely to be a software schedule issue. If the display cuts out again, see whether audio continues and whether the keyboard still responds. That can help distinguish a graphics-driver or GPU power failure from a complete system lockup.
The fact that lightly moving the unused power-cable pigtail immediately caused a signal loss is a strong reason to inspect the GPU power connection. With the PC powered off and unplugged, use a flashlight to check the GPU socket and both cable connectors for a pushed-back terminal, looseness, discoloration, or burn marks. Make sure the connector is fully seated and that the cable is supported so it is not pulling sideways on the graphics card.
Since you already tried the other PCIe slot and another cable, the fault could still be a loose connector, the GPU socket itself, or a problem with the PSU-side connection. Do not keep flexing the cable while the system is running. A damaged terminal may be replaceable, but repairing a GPU socket or PSU modular connector is best left to a qualified technician; never mix modular cables from different power-supply models.
I tried the second PCIe slot and the issue remained. If a pin in the GPU socket is pushed back or the socket is damaged, can it be repaired? The strange part is that the display can also cut out when I am not touching the case or cable.
Would trying another modular PCIe port on the PSU side be a reasonable test? The current cable has been connected for about three years. I will inspect the pins, but I want to avoid making the situation worse.
The only event I see is Kernel-Power 41, which I suspected was just recorded because I had to restart manually. I already changed the power strip and tried a clean boot, but the failures still happen randomly—sometimes after an hour of gaming and sometimes after only 30 minutes.