Failure phenomenon: Computer failure, the phenomenon of using the average every 20 minutes will crash once, reboot after a few minutes will again panic.
Failure Analysis and processing: The initial estimate is the CPU temperature in the chassis caused by the panic, in the BIOS to check the temperature of the CPU, found that the temperature is only 33 ℃. Later found that the computer boot in the BIOS to check the temperature is only 31 ℃, the power-on use 1 hours later, the temperature only rose 2 ℃, at that time at the temperature of 35 ℃ or so. It appears that the measured CPU temperature is not accurate. Open the chassis found the fan on the heatsink because of the accumulation of dust too much, has been turned, and then replaced the CPU fan, then the boot, the computer ran for several hours of the game did not occur the crash. It was later found that the temperature probe on the motherboard was glued to the heatsink to measure the CPU temperature. And now this probe does not have close contact with the heat sink, separate a very large distance, heat sink heat can not directly to the temperature probe, measured temperature natural error is very large. When the CPU fan is replaced, the probe and heatsink are securely fastened together so that after 20 minutes on the boot, the temperature measured in the BIOS is 45 ℃, after which everything is normal.