My monitors randomly go black while gaming, watching videos, or even browsing, but the computer itself stays powered on and does not restart. I'm using 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. Both Samsung Odyssey G5 monitors are connected through DisplayPort, with an HDMI cable also connected to a TV that is usually turned off.
I have already reinstalled Windows, replaced the monitor cables, tried another GPU power cable, and tested the hardware at a repair shop. My GPU worked in another system, and another GPU worked in my system during their testing, but the issue returned after I brought the PC home. The black-screen crashes can happen after anywhere from 30 minutes to two hours, with no consistent pattern.
I recently noticed that barely touching the unused dangling connector on the GPU power cable caused both monitors to lose signal immediately while the PC remained on. The cable uses two 6+2-pin connectors. I tried both connectors and also another PSU cable, but the problem remained. I also tried the second PCIe slot. Event Viewer only shows Kernel-Power 41 after the incident. What should I check next, and could the GPU power socket, cable, or PSU connection have a loose or damaged pin?
2 Answers
The fact that moving the dangling part of the GPU power lead immediately kills the display is a strong clue. It could be tugging on the plugged-in connector and briefly breaking contact, or there could be a loose or damaged terminal in the GPU socket, cable, or modular PSU socket. Inspect both ends with a flashlight for a pushed-back, burnt, or discolored pin, and make sure the connector is fully seated. Support the cable so it cannot pull sideways on the graphics card. Do not mix modular cables from another power supply, even if the plugs fit. Since the issue also happened in the second PCIe slot, the slot itself is less likely to be the cause. If the socket or pin is damaged, have the card or PSU repaired or replaced rather than trying to bend contacts while the system is powered.
It is worth isolating the power and display paths, even though the cable movement makes the GPU power connection suspicious. Connect the PC and monitors directly to a wall outlet temporarily, try different outputs on the graphics card, and test with only one monitor connected. Check Event Viewer after a failure, but Kernel-Power 41 generally only means Windows noticed an unexpected loss of normal operation; it does not identify the original fault. A clean boot and timing the failures may help, but the inconsistent intervals do not rule out an intermittent hardware connection. If possible, test with a known-good, correctly matched PSU or have the GPU power connector and PSU modular socket inspected under load.
I already changed the power strip and tried a different wall outlet, and the clean-boot test did not prevent the issue. The failures still happen at random times, sometimes after about 30 minutes and sometimes after an hour or more.
I tried the second PCIe slot and the problem still occurred. If a pin in the GPU socket is pushed back or damaged, is that something a repair shop can replace? I’m also wondering whether using a different modular socket on the PSU would be a sensible test.