My roughly one-year-old PC has started hard-crashing at random. The display suddenly loses signal, the GPU fans ramp up, and the system becomes completely unresponsive until I force it off. There is no blue screen, and it can happen while gaming, sitting idle, at the Windows login screen, or when waking from sleep. The crashes tend to come in clusters: I may go days or weeks without one, then experience several close together.
The system uses a Ryzen 7 7800X3D, ASUS B650M-AYW WIFI motherboard, Gigabyte RTX 5070 Eagle OC 12GB, 16GB of Kingston DDR5-6000 running at default JEDEC 4800 settings, and an MSI MAG A850GL 850W Gold PSU. It runs Windows 11 Home 25H2.
Event Viewer only records Kernel-Power Event ID 41, with BugcheckCode 0 and no other useful parameters. I reset and updated the BIOS, disabled PBO, confirmed there is no manual voltage offset or EXPO profile enabled, and checked that the SoC voltage is still on Auto. HWiNFO logging showed the 12V rail holding around 12.172V and temperatures rising normally, around 45–48°C, before one crash. I also performed a clean Windows installation, removed Armoury Crate, and reseated the RAM and GPU power cables. The Minidump folder is empty. What hardware or settings should I test next?
3 Answers
Event 41 only means the system was turned off unexpectedly; it does not identify the failed component. Stable temperatures and a steady sensor reading do not completely rule out the PSU, GPU, motherboard, or a transient power problem, especially because software logging may stop before the failure. Try running the system with another known-good PSU if possible, verify every modular cable is the one supplied with that PSU, and test the GPU separately using a different cable or another graphics card if available.
Test the memory independently even though EXPO is disabled. Run one DIMM at a time, try each motherboard slot recommended by the manual, and use a bootable memory test for several passes. Also test with the GPU removed if the processor’s integrated graphics can provide display output. Those tests can help separate RAM, the graphics card, and the motherboard from the CPU.
This looks more like a hardware-level lockup than a normal Windows crash. Since there is no bugcheck and the fans remain at full speed, I would start by testing CPU stability even at stock settings. A processor or motherboard power-delivery issue can appear as black screens and hard freezes without producing a dump file. Try temporarily reducing CPU boost or applying a small, conservative power limit, then run a CPU stability test. If the crashes stop, that points toward the CPU, motherboard, or firmware rather than Windows.
That matches what I’m seeing: the GPU fans definitely ramp up and stay there, while the screen has no signal. I’ll test with reduced boost and power limits rather than assuming the PSU is at fault.
A clean Windows installation and an empty Minidump folder are expected with this kind of complete lockup. The absence of a dump does not mean the hardware is fine; it means Windows never got a chance to record a normal crash.