My desktop shuts off instantly after roughly 25–30 minutes. There is no blue screen, visible freeze, or obvious temperature spike, and Windows later records Kernel-Power Event ID 41. The system has a Gigabyte GA-H81M-DS2 motherboard, an ASUS GTX 560 Ti DirectCU II TOP 2GB, a Thermaltake Litepower LTP-550AL2NK 550W power supply, DDR3 memory currently reduced to 4GB, and SATA SSD/HDD storage. Nothing is overclocked, and the BIOS dates from 2015.
I performed a fresh Windows installation, installed current drivers, tried different storage devices, tested the CPU and memory with OCCT, and monitored temperatures. The tests eventually produced the same shutdown without reporting errors. It also happens outside Windows: the computer powers off while sitting idle in the BIOS after about 30 minutes. I tried different wall outlets and even tested generator power, where it ran for approximately 30 minutes during one attempt. I have not updated the BIOS yet.
Since the failure occurs in the BIOS, I suspect the power supply, motherboard or VRM, graphics card, or unstable electrical power rather than Windows or drivers. What should I test next to identify the faulty component before replacing parts?
2 Answers
A graphics card can cause a sudden power loss if it has an electrical fault or makes the PSU protection trip, but the fact that the PC also turns off while idle in the BIOS makes the PSU or motherboard more suspicious. If possible, borrow a known-good PSU first; that is a better test than buying a GPU. Inspect and reseat the 24-pin motherboard connector, CPU power connector, and graphics-card power leads, and try running with only one memory module installed. Since the shutdown occurs outside Windows, focus on hardware, power delivery, and the timed behavior rather than drivers or storage.
Because it shuts down while sitting in the BIOS, crash dumps and Windows troubleshooting probably will not help. Start with the power path: test with a known-good, quality PSU, preferably at another location or through a UPS with only the computer connected. The Thermaltake Litepower is a likely suspect despite its wattage rating. Also disconnect every extra storage device and leave only the boot device connected.
Check the motherboard closely for damaged or discolored VRMs, bulging capacitors, warped areas, or scorch marks. If the board exposes VRM or northbridge sensor readings, monitor them with OCCT and stop the test before the usual failure time. A BIOS update may be worth trying, but it is less likely to fix an abrupt, highly repeatable power loss than a failing PSU or motherboard.
The motherboard is a Gigabyte GA-H81M-DS2. Unfortunately, the onboard VGA output does not work, so I cannot simply remove the GTX 560 Ti and test using the integrated graphics. I am trying to understand whether a bad graphics card could really shut down the entire computer before buying replacement parts.