For the past three or four days, my PC has been completely shutting off and restarting instantly at random. It happens both while gaming and during light browsing or idle use, so it does not seem tied to system load. The only recent change was using a Windows in-place upgrade to repair possible system corruption.
Temperatures, PSU readings, and SSD health look normal. I tried reducing the RAM speed, changing my undervolt settings, and running a 10-minute OCCT Power AVX2 test, which completed without errors. Disconnecting the case reset-switch header also did not stop the restarts. Event Viewer shows Kernel-Power Event ID 41, but no WHEA errors or other obvious clues.
My undervolt profiles have been stable for around four and a half years, although I have temporarily reset the BIOS and used more conservative settings without any improvement. Could Windows corruption cause this kind of instant restart, or should I focus on the motherboard, PSU, RAM, or another hardware issue?
System specifications: Asus ROG Strix B550-F motherboard, Ryzen 7 5800X, Aorus RTX 3080 Ti, Aorus Gen4 SSD, 32 GB Trident Z Neo DDR4-3600 CL18, Corsair RM850x 3.0 PSU, and an Arctic Liquid Freezer III AIO.
2 Answers
Start by loading completely default BIOS settings, including disabling the undervolt, memory profile, and any other tuning. Even if the settings have been stable for years, a component can become less tolerant over time, and idle crashes can still be caused by voltage or power-state transitions. If the system remains stable at defaults, re-enable changes one at a time.
Also check Windows Reliability Monitor rather than relying only on Event Viewer. Press Win+R, run `perfmon /rel`, and inspect the exact times of the red error markers and unexpected shutdowns. It often gives a clearer timeline of crashes, updates, and startup failures. Event ID 41 only confirms that Windows did not shut down cleanly; it does not identify the faulty component.
A Windows repair or corrupted system file can cause crashes, but an instant power-off followed by an immediate restart is more commonly associated with hardware, firmware, or power-management problems. The in-place upgrade makes software worth checking, but it does not rule out a failing motherboard, intermittent PSU connection, RAM instability, or a GPU power issue.
Try running fully at BIOS defaults, reseat the RAM and GPU, check both ends of every PSU cable, and test with one memory module at a time if possible. If Reliability Monitor shows no useful software failure, a clean Windows installation or a separate test drive can help distinguish an operating-system problem from hardware.
I reset the BIOS and temporarily made the undervolt settings even more conservative, but the restarts continued. I’ll check Reliability Monitor for anything that lines up with the crashes.