How to Fix Windows Restarting Randomly and Find the Cause

To fix a Windows PC that restarts randomly, first disable automatic restart so you can read any blue-screen stop code. Then use Reliability Monitor, Event Viewer, and minidump files to determine whether the cause is a crash, power loss, overheating, hardware failure, or a planned update. Back up important files, reverse recent software changes, test Safe Mode, check heat and connections, and run Windows and memory checks before replacing hardware.

First, find out whether Windows crashed or simply lost power

A sudden restart can produce a blue screen, cut power without warning, or follow a planned update. If the exact stop-code text is unknown, use this symptom description:

The PC goes black and restarts without warning, or shows a blue screen too briefly to read.

Disable automatic restart before the next failure. This requires Administrator permission and does not delete your files.

Control Panel > System and Security > System > Advanced system settings > Advanced > Startup and Recovery > Settings

Clear Automatically restart, leave Write debugging information set to Small memory dump, and confirm the dump folder is %SystemRoot%\Minidump. Select OK, then OK again.

If the next failure shows a blue screen, record the stop code and any driver name. WHEA_UNCORRECTABLE_ERROR points toward a hardware or low-level driver problem, but it does not identify the failed component. A message naming a display or storage driver is useful evidence, not proof that the hardware is healthy.

A black screen followed by an immediate reboot, with no stop code and no dump file, fits power loss, a hard reset, firmware trouble, or a crash too severe to write a dump. Record whether the screen goes black instantly, whether fans stop, whether the PC powers back on by itself, and whether the failure occurs only during gaming or another demanding task.

A restart after Windows displays Installing updates or a similar message is deliberate. Windows may also restart outside active hours after an update. A restart at the same time every day can indicate a scheduled task, update policy, or third-party utility.

Which Windows versions and restart patterns point to each cause

These checks apply to supported desktop and laptop editions of Windows 10 and Windows 11. You normally need an Administrator account for dump settings, driver changes, repair commands, and some Event Viewer details. Reliability Monitor can be opened by a standard user, although Administrator approval may be requested for related actions. Save work and back up important files before repeated testing.

Before stressing the PC, do two safe checks: note the exact restart time and remove any recently added peripheral or accessory. Do not open a powered-on machine.

Open Reliability Monitor first. Press Windows, type reliability, and open View reliability history. Select the day of the restart and inspect Critical events and Windows failures. Windows was not properly shut down marks an abnormal shutdown but does not explain it. Windows stopped working with a bug check and stop code supports the conclusion that Windows crashed.

Event Viewer provides the lower-level record:

Event Viewer > Windows Logs > System

Inspect events around the exact time:

Event 41 appears after power is cut, the reset button is used, or a severe crash prevents Windows from recording a clean shutdown. Compare it with BugCheck, WHEA, temperature readings, and physical symptoms.

Look for minidumps in %SystemRoot%\Minidump. A new .dmp file after each blue screen confirms that Windows captured crash data. Microsoft debugging tools can analyze these files; without them, the stop code, timestamp, and repeated driver names still provide useful direction. No dump, no BugCheck event, and a sudden loss of power shift attention toward power delivery, overheating protection, firmware, or hardware.

Work through the fixes in the safest order

Make one change at a time and record the result. Do not reset Windows, delete drivers manually, overclock, or run a stress test on a machine that may be overheating.

1. Undo the most recent driver, utility, or Windows change. In Settings > Windows Update > Update history, identify updates installed just before the first restart. For a recent display, chipset, storage, or network driver, open the device's Properties > Driver in Device Manager and select Roll Back Driver when available. For a problematic quality update, use Settings > Windows Update > Update history > Uninstall updates. A restart after rollback is expected. If failures stop, keep the stable driver and avoid reinstalling the same faulty package until a newer version is available.
2. Test Windows in Safe Mode. Hold Shift while selecting Restart, then choose Troubleshoot > Advanced options > Startup Settings > Restart, followed by 4 for Safe Mode or 5 for Safe Mode with Networking. Safe Mode loads a limited driver set. Stability there makes a startup program, third-party driver, or graphics component more likely. Restarts in Safe Mode make software loaded only during normal startup less likely.
3. Disconnect nonessential external devices. Remove docks, USB storage, hubs, capture hardware, and recently added peripherals. Test with only the keyboard, mouse, display, and required network connection. Reconnect one item at a time after stability returns.
4. Check temperatures without opening a powered-on machine. Use the PC maker's diagnostic tool or a reputable hardware monitor during the workload that normally triggers the restart. Sudden fan noise, severe performance throttling, blocked vents, or a laptop hot enough to shut down point toward cooling trouble. Shut the PC down before cleaning vents; do not use it on bedding or block a laptop's intake.
5. Repair Windows system files from an elevated Command Prompt. Administrator permission is required.

DISM.exe /Online /Cleanup-Image /RestoreHealth
sfc /scannow

DISM should complete successfully. sfc should report either that it found no integrity violations or that it found and repaired files. If either command reports that it could not complete, record the exact message; Windows is not confirmed repaired.

6. Test memory. Press Windows, type Windows Memory Diagnostic, choose Restart now and check for problems, and let the test finish. Any reported memory error is significant. Power down, reseat removable RAM only if you are comfortable doing so, and test modules individually according to the computer or motherboard manual. Do not continue gaming or editing important files with confirmed memory errors.
7. Inspect storage, power, and battery symptoms. Back up first if a drive clicks, disappears, reports repeated controller errors, becomes unusually slow, or causes file corruption. An instant power loss under gaming load can indicate a failing or undersized power supply, a loose motherboard or graphics power cable, or a failing graphics card. A laptop that restarts when the charger moves, runs only on AC power, or has a swollen battery needs service. Do not puncture, squeeze, or continue charging a swollen battery. Repeated WHEA errors, burning smells, visible damage, arcing, or immediate shutdowns justify professional repair before further testing.

How to confirm the random restarts are actually gone

A single successful boot is not a fix. Reproduce the activity that caused the restart, but stop if the case becomes abnormally hot, a burning smell appears, a drive clicks, or the system shows repeated hardware errors. Back up files and arrange repair instead of continuing to stress the machine when those signs appear.

Use the same workload in stages: normal desktop work first, then the application or game that usually triggered the failure. Do not change several drivers or hardware parts at once. Keep automatic restart disabled until the machine has remained stable long enough to catch another stop code if one occurs.

After each test, review Reliability Monitor and the System log for entries at the test time. A successful fix should produce no new BugCheck event, no new minidump, and no repeating WHEA errors. An old Kernel-Power event can remain, so compare timestamps rather than treating its presence as a new failure.

If the PC restarts again, compare the new event with the old one. The same stop code or driver points back to that fault. A restart with only Kernel-Power and no dump points back toward power, heat, firmware, or a hard reset. Restore your normal workload gradually and keep recent backups until the pattern stays clean.

Prevent the problem from returning and know when to replace hardware

Once the system is stable, install drivers and firmware from the computer or motherboard manufacturer's support page, or use Windows Update when appropriate. Prefer one known-compatible release over automatic driver-updater utilities. Keep Windows updated, but review update history if restarts begin immediately after a particular update. Do not use overclocking or aggressive firmware profiles while diagnosing unexplained shutdowns.

Keep vents clear, clean dust with the machine powered off, and monitor temperatures during demanding work. Ensure desktop power cables are fully seated at the wall, power supply, motherboard, and graphics card. Avoid overloaded or damaged extension equipment. A laptop charger should match the manufacturer's required specification. A battery that is swollen, unusually hot, leaking, or lifting the case should not be used.

Set a regular backup routine before the next failure. Important files should exist on a separate drive or trusted cloud service, and a backup is useful only if you can access or restore it. If storage is degrading, copy irreplaceable files before running repair scans or causing more restarts.

Replace or service hardware when symptoms persist after clean software and Safe Mode testing. Instant shutdowns under load, repeated WHEA errors, memory-test failures, a drive that clicks or vanishes, a swollen battery, burning smell, visible scorch marks, or a system that will not stay powered long enough to back up all justify professional repair. A failing power supply, motherboard, graphics card, or battery can damage data or other components; stop experimenting and arrange repair, starting with a backup if the machine remains safe to operate.

Conclusion

Disable automatic restart, record the next stop code and event pattern, and work through recent software changes, Safe Mode, cooling, connections, system files, and memory checks. If the PC shows heat, battery, storage, electrical, or repeated hardware-failure warnings, protect your files and stop using it until it is serviced.

Frequently Asked Questions

Why does Event 41 mean my power supply is bad?

It does not. Kernel-Power event 41 only says Windows did not shut down cleanly. A failed power supply is one possibility, alongside a hard reset, overheating, firmware trouble, or a crash that prevented Windows from writing a dump. Compare it with BugCheck, WHEA, temperature readings, and physical symptoms.

What if Windows restarts before I can read the blue screen?

Disable `Automatically restart` under `Control Panel > System and Security > System > Advanced system settings > Advanced > Startup and Recovery > Settings`. Windows should remain on the stop screen so you can record the code. Keep minidump creation enabled for later crash analysis.

Should I reset Windows to fix random restarts?

Not first. A reset can remove software causes while leaving a failing power supply, RAM, drive, battery, or overheating problem unresolved, and it can complicate data recovery. Collect crash and hardware evidence and back up files before considering a reset.