I built this PC about a year ago, and it worked normally for the first couple of months. Since then, it has either rebooted unexpectedly or completely frozen. During a freeze, the mouse may still move and audio from a video can continue briefly, but eventually the only way to recover is holding down the power button.
I initially suspected the XFX RX 580 8GB because some crashes happened during demanding games. I replaced the 750W power supply, tested the Kingston DDR5 memory with MemTest86 without errors, monitored temperatures and voltages, and tried both newer and older chipset and graphics drivers. None of that solved the issue.
The random restarts stopped after removing the graphics card and using the Ryzen 5 7600's integrated graphics, but the same type of freezing still happens occasionally. The freezes are now difficult to reproduce, while the crashes were more likely to occur under load with the RX 580 installed.
Event Viewer only shows Kernel-Power 41, and stress tests have not reliably triggered the problem. Could the motherboard or CPU be faulty, or are there other tests and settings I should check first?
Specs:
CPU: Ryzen 5 7600
GPU: XFX RX 580 8GB
RAM: Kingston Fury Beast DDR5-5200 32GB (2x16GB), CL40
Motherboard: MSI B650M Gaming Plus WiFi
SSD: Forgeon Nimbus Pro 1TB PCIe 4.0 NVMe
PSU: MSI MAG A750GL PCIe 5, 750W, 80 Plus Gold
3 Answers
Kernel-Power 41 only tells you that Windows did not shut down cleanly; it is usually a result of the lockup or forced power-off, not the underlying cause. Check Reliability Monitor as well as Event Viewer, and look for display-driver, WHEA, storage, or chipset errors around the time of each incident. BIOS and AMD chipset updates are also worth installing.
Since the RX 580 made the problem easier to trigger during games, test the system without it for a while, then test with another known-good graphics card if possible. The card is an older model, so a hardware fault or driver interaction is plausible.
Run the memory at completely stock settings while troubleshooting. If DDR5-5200 is being enabled through EXPO or another memory profile, disable it and use the motherboard's default JEDEC speed. A memory test passing does not prove that an overclocked memory configuration is stable in every workload. Leave it at stock for a week or two and see whether the freezes return.
Also check for a firmware update for the NVMe drive and make sure the motherboard BIOS and chipset drivers are current.
When you replaced the power supply, replace every modular cable too. Modular cables are not guaranteed to be wired the same way between PSU models, even when the connectors fit. Reusing the old cables can cause instability or hardware damage. After confirming the cabling is correct, reseat the RAM, GPU, and NVMe drive and inspect the motherboard power connections.
If the issue continues at stock memory settings with the GPU removed, test with a different boot drive or run a temporary clean operating-system installation. That helps separate a storage or software problem from a motherboard or CPU fault.
I did check Event Viewer, but the only recurring entry is Kernel-Power 41. Stress testing has not consistently reproduced the freeze; with the RX 580 installed, demanding games were the closest thing to a reliable trigger.