Rightek Rescue · Guides · Recurring blue screens

Recurring blue screens (random BSODs)? Here's how to find the cause

When Windows blue-screens, it writes a small crash dump naming what failed. Random, repeating BSODs almost always trace back to one specific driver (very often a graphics driver) — or, less often, failing hardware. The trick is reading the dumps to find the culprit instead of guessing.

Updated 23 August 2026 · written and tested by Rightek IT Services

What you're seeing

Blue screens at random times, or a loop of them at startup
Stop codes like SYSTEM_SERVICE_EXCEPTION, SYSTEM_THREAD_EXCEPTION_NOT_HANDLED, DPC_WATCHDOG_VIOLATION, KERNEL_POWER 41
They may have started after a driver or Windows update, or a new device

Why it happens

A buggy or mismatched driver (graphics: amdkmdag.sys, nvlddmkm.sys, igdkmd*.sys; also network, audio, or vendor utilities)
A bad Windows/driver update
Hardware: failing RAM, an overheating GPU, or a dying drive (KERNEL_POWER 41 is often power/heat)

Fix it yourself — the manual steps

Note the stop code on the blue screen and search it — it narrows the cause.
Boot into Safe Mode and roll back or reinstall the most recently updated driver (start with graphics).
Run Windows Memory Diagnostic to rule out bad RAM.
Reading the actual crash dumps yourself needs the Windows debugger and some know-how — most people can't tell which module is really at fault.

Reading the crash without a debugger

Every blue screen writes a small dump file to C:\Windows\Minidump, and it names the driver involved. Properly analysing it needs Microsoft's debugging tools, but there are two routes that need nothing extra:

Reliability Monitor. Search for it from the Start menu. It gives a timeline of crashes against installs and updates, which is often enough on its own — if the crashes start on the day a driver updated, you have your answer without opening a dump at all.
Event Viewer. Windows Logs → System, and look for Event ID 41 (the unexpected shutdown) and the BugCheck entry just before it, which records the stop code and parameters.

Make sure the dumps are actually being written first: System → About → Advanced system settings → Startup and Recovery, and set "Write debugging information" to Small memory dump. If it is set to None, every crash so far has left no evidence.

Ruling out a driver properly

Rolling back the graphics driver is the right first move, because graphics drivers cause more of these than anything else. Do it thoroughly:

In Safe Mode, remove the display driver completely rather than installing over the top — a partial mix of old and new files is its own cause of crashes.
Install the version from the PC or GPU manufacturer, not the newest one available. Newest is not most stable, and on laptops the vendor build often contains power-management fixes the generic one lacks.
If the crashes started immediately after a Windows update rather than a driver one, uninstall that update and pause updates for a fortnight to confirm.

Driver Verifier: useful, and easy to trap yourself with

Windows has a tool that stress-tests drivers and forces an immediate crash naming the culprit, started with verifier. It genuinely works and it is how the offending driver gets identified when nothing else will.

The trap is that it makes the machine crash more, deliberately, and if the driver at fault loads at boot you can end up in a crash loop that you cannot switch it off from. Only turn it on if you are confident you can reach Safe Mode, and know the escape: verifier /reset from a Recovery Environment Command Prompt. If that sounds like a bad afternoon, it is a reasonable thing to skip.

When it is not a driver at all

Varying stop codes with no common thread point away from software:

Memory. Run MemTest86 from a USB stick for several passes. Windows' built-in tester is quicker but misses subtle faults. If you have multiple sticks, testing one at a time isolates it for free.
Heat. Crashes only under load, with temperatures at the limit, mean cooling. Dust and dried thermal paste, usually.
Storage. A failing drive produces crashes that look like anything. Check the disk's health.
An uncorrectable hardware fault, which has its own stop code — see WHEA_UNCORRECTABLE_ERROR.

⚡ The one-click way: Rightek Rescue names the culprit

Rescue reads the crash dumps and, when they point at one driver, names it and safely disables just that one driver (reversibly — it renames the file and can be undone) so Windows can boot on its built-in one; you reinstall the vendor's latest afterward. It never disables storage or boot-critical drivers, and if the evidence points at failing hardware (RAM/heat/disk) or is inconclusive, it says so instead of pretending a software tweak will fix it. On any working PC: download Rescue, choose Crash Medic or Fix My PC. Not sure what's wrong? It tells you free first.

Get Rightek Rescue → See it work

FAQ

Will I lose my files?

No. Disabling a faulty driver doesn't touch your personal files. Rescue's driver fix is reversible and never overwrites your data.

How do I know if it's a driver or hardware?

The crash dumps tell you - if they consistently blame one driver, it's software; if they're scattered or point at memory/power, it's likely hardware. Rescue reads the dumps and gives you a straight answer.

My screen is at low resolution after the fix - is that normal?

Yes. If Rescue disabled a faulty graphics driver, Windows falls back to a basic display driver until you reinstall the vendor's latest one - which is the safe way to get you booting again.

Which Windows versions are supported?

Windows 10 and 11. Rescue runs from a USB, so it can read the crash dumps even when Windows won't start.

Where are the crash dumps, and how do I read them?

In C:\Windows\Minidump. Reading them properly needs Microsoft's debugging tools, but Reliability Monitor and Event Viewer give you most of the same information with no setup — particularly a timeline showing whether the crashes began with a specific install or update.

Should I install the newest graphics driver?

Not automatically. Newest is not the same as most stable, and on laptops the version from the PC manufacturer often carries power-management fixes the generic one does not. Also remove the old driver rather than installing over it, since a mixture of old and new files causes crashes of its own.

Is Driver Verifier safe to use?

It is effective and it is easy to trap yourself with. It deliberately makes the machine crash more in order to identify the faulty driver, and if that driver loads at boot you can end up in a loop you cannot switch it off from. Only use it if you can reliably reach Safe Mode, and know that verifier /reset undoes it.

Different stop codes every time — what does that mean?

Usually hardware rather than software. A single faulty driver tends to produce the same code repeatedly. Memory faults, overheating and a failing drive all corrupt whatever happens to be running at the time, so the reported code varies from crash to crash.