When a Computer Keeps Turning off

To find out what is causing a computer to shut down or crash unexpectedly, use built-in Windows diagnostic tools and third-party monitoring utilities to check for software errors, hardware failures, or overheating. 

• Windows Event Viewer: Open the Start menu, type Event Viewer, navigate to Windows Logs > System, and look for critical errors marked by a red "X" around the time of the shutdown.
• Reliability Monitor: Press Windows Key + R, type perfmon /rel (or search for Reliability Monitor in the Start menu) to see a chronological timeline of application failures, Windows updates, and unexpected shutdowns.
• Blue Screen View: If the PC experiences a Blue Screen of Death (BSOD), download a specialized utility like Blue ScreenView to inspect mini-dump files and identify the exact driver or module that triggered the crash.

• Hardware Sensors: Use free monitoring software like HWMonitor, Core Temp, or MSI Afterburner to check CPU and GPU temperatures while the system is idle and under load.
• Overheating Thresholds: If temperatures exceed 80°C to 90°C+ during basic tasks or gaming, the system may be thermally shutting down to prevent damage. Clean out dust from fans and heatsinks, or check if the CPU cooler needs new thermal paste.

• Windows Memory Diagnostic: Press the Windows Key, type Windows Memory Diagnostic, and choose to restart and run the test to check for faulty RAM sticks.
• MemTest86: For a deeper diagnosis, run a bootable USB tool like MemTest86 to isolate memory errors.

• System File Checker: Open Command Prompt or PowerShell as an administrator and run sfc /scannow to repair corrupted Windows system files.
• Driver Updates: Check the Device Manager for warning icons, and update your graphics card, chipset, and BIOS drivers to the latest versions from the manufacturer.
• Check Disk: Run chkdsk /r /x in an elevated terminal to repair file system errors or bad sectors on your storage drive.

• Stress Tests: Run benchmark tools like FurMark for the GPU or CPU stress tests to see if heavy loads trigger an immediate shutdown (which often points to a failing power supply unit or unstable overclock).
• Power Supply (PSU): If the computer cuts out completely power-wise (like pulling the plug) without a blue screen or error log, the power supply or its 12V rails may be failing under load.

To confirm if a computer reset fixed the issue, you need to simulate heavy usage. A software reset fixes driver or OS corruption, but it won't fix hardware issues like a failing power supply or overheating.

Here is how you can stress-test the machine to see if the problem returns:

1. Run Hardware Stress Tests

Since you don't know what triggers the shutdown, force the hardware to run at maximum capacity to see if it triggers a crash.

• CPU Stress Test: Download Prime95 or Cinebench and run a stability test for 30–60 minutes. If the PC cuts power, the CPU is overheating or drawing unstable power.

• GPU Stress Test: Run FurMark or 3DMark. This pushes the graphics card to its limit, drawing maximum wattage from the power supply unit (PSU).

• Combined Test: Run Prime95 and FurMark at the same time. If the PC shuts off instantly, the PSU is likely faulty and cannot handle a full power load.

2. Monitor Temperatures During the Tests

Keep a monitoring tool open while running the tests above.

• Tool: Use HWMonitor or HWiNFO64 to watch temperatures.

• What to look for: If the CPU or GPU climbs past 90°C–95°C right before a crash, the issue is thermal throttling/overheating, meaning the fans or thermal paste are failing.

3. Let it Idle and Loop Video

Sometimes crashes happen due to low-power state instability or RAM issues.

• Leave the computer running a looping 4K video on YouTube or a benchmark loop overnight.

• If it is still running in the morning, the system is likely stable for basic office workflows.

4. Check Event Viewer Again

After running these tests (whether it crashes or survives):

• Open Event Viewer (eventvwr.msc).

• Go to Windows Logs > System.

• Filter for Event ID 41 (Kernel-Power). If you see this error timestamped during your testing, the PC is still unexpectedly losing power, proving the reset did not fix the hardware issue.