My PC has worked reliably for a couple of years, but it recently started shutting down completely while gaming—even during something lightweight like Roblox. The system has a ROG Strix B550-A motherboard, Ryzen 5 5600X, 32 GB Corsair Dominator DDR4-3600, EVGA RTX 3080, and a 750W Gold-rated ROG Strix power supply. CPU and GPU temperatures and usage look normal, and I've previously played demanding games without problems. The C: drive is also nearly full, with only about 10 GB remaining. When the shutdown happens, the display goes black and the power button does nothing. The PC only starts again after I switch the PSU off for several seconds and turn it back on. Loading into Roblox then causes another shutdown within 10–20 seconds. Could the PSU be failing, or might another component, airflow issue, outlet, or cable be responsible?
4 Answers
Try narrowing it down by reducing the load. If the PC stays on with the RTX 3080 removed or with a low-power graphics configuration, that makes the PSU, GPU power cables, or the graphics card more suspicious. Make sure the GPU uses properly connected separate PCIe power cables where applicable rather than an overloaded daisy-chain lead. Also inspect the case power and reset switches, since a faulty switch can occasionally cause strange behavior.
The behavior strongly points to a power-delivery problem, especially because the system won’t respond until the PSU switch is cycled. A failing or overheating PSU can trip its protection circuitry under GPU load even when the CPU and GPU temperatures look normal. Check that the PSU vents and fan are clean and unobstructed, reseat the motherboard and GPU power cables, and if possible test with a known-good, adequately rated PSU. If the unit is still under warranty, an RMA or replacement would be sensible.
A completely instant shutdown without a blue screen is different from a normal Windows crash. Your nearly full system drive is worth addressing—free up substantially more space—but that alone usually wouldn’t cause the PSU-style lockout you described. Event Viewer may only show an unexpected shutdown because Windows cannot record the underlying hardware failure. Also try connecting the PC directly to a wall outlet instead of an extension, or test the extension with another device.
Give the whole system a careful cleaning, including the GPU cooler, CPU cooler, PSU intake, and filters. The PSU can overheat even if the CPU and GPU sensors report safe temperatures. Check the SSD’s SMART health data as well, and consider whether the problem began after a BIOS update or hardware change. If cleaning, cable checks, and a direct-wall test do not help, testing the PSU with a known-good replacement is probably the fastest way to confirm the fault.
I’m using an NVMe drive and haven’t updated the BIOS automatically, so I’ll check whether the timing matches a manual update. I’ll also clean the PC, free up space, and test it without the extension before replacing anything.