Gaming PC Random Restart Repair: Finding a Power Supply Fault Under Load

We checked Windows logs, temperatures, GPU power connections, memory stability, BIOS defaults, and separate component loads. A combined CPU and GPU load reproduced an immediate voltage drop that did not occur with a known-good bench power supply.

We changed one variable at a time and recorded the result. This matters because two failures can look identical to the owner. A replacement part may hide a symptom temporarily without correcting the cause. Our diagnostic goal was to identify the failed system, confirm that the remaining parts were healthy, and give the customer a clear repair option.

Before opening the device, we also noted its physical condition and protected accessible data whenever the situation allowed. If a drive is unstable or liquid is present, the safe order of operations changes; preventing additional damage takes priority over forcing the computer to start.

We replaced the undersized aging unit with a quality power supply appropriate for the installed graphics card, used dedicated PCIe power leads, checked every modular connection, and corrected cable routing around the intake.

We inspected every connection disturbed during the repair and avoided unnecessary changes to unrelated components. Firmware or software settings were changed only when they supported the verified hardware condition. The computer was then reassembled far enough for controlled testing before final closure.

The system completed repeated combined loads, long gaming sessions, cold starts, and restart cycles. No power event or unexpected reboot returned.

A computer turning on once is not a completed repair. We attempted to reproduce the original condition, compared the before-and-after behavior, and checked nearby functions that could have been affected by disassembly. Only after those checks passed was the repair considered ready for customer review.