How to fix a blue screen problem easily? (2026 guide)


A blue screen (BSOD) on Windows means Windows shuts down immediately because it has detected a “critical” error (driver, hardware, system files, storage, etc.). It can happen after an update, after installing software, after adding new hardware… or with no obvious reason.

The goal here: identify the cause and apply a simple (but rigorous) method to repair a bluescreen without going round in circles.

Understand what a BSOD (blue screen of death) is really telling you


A BSOD doesn’t necessarily mean “Windows is broken”. It’s often one of these scenarios:


  • Unstable / incompatible driver (GPU, Wi‑Fi, audio, chipset, storage…).
  • Unstable RAM (XMP/EXPO profile too aggressive, worn-out stick, wrong settings).
  • Storage that drops out (SSD/NVMe, errors, firmware, cable/SATA if the 2.5"/3.5" drive).
  • Corrupted system (damaged Windows files after a power cut, update, driver installed incorrectly).
  • Overheating / power supply / overclock / undervolt (CPU/GPU/VRAM, voltage curve that’s too optimistic).


👉 The most useful instinct: recover the error code + the context (when it happens, after what, and whether it comes back).

Step 1 — Get the information that speeds up diagnosis


The stop code + the file name (if shown)


On the blue screen, note/photograph:


  • the Stop code (e.g. IRQL_NOT_LESS_OR_EQUAL, WHEA_UNCORRECTABLE_ERROR, etc.)
  • the file if Windows shows one (e.g. nvlddmkm.sys, ntoskrnl.exe…).


Even if you don’t understand it, it guides everything else.


Check whether Windows has saved a “minidump”


After restarting:


  • Look for a folder: C:\Windows\Minidump
  • Or a file: C:\Windows\MEMORY.DMP


If you have minidumps, you can analyse them with a simple tool (BlueScreenView) or a more precise one (WinDbg). This avoids guesswork.

Step 2 — Boot into Safe Mode (the right way in 2026)


Safe Mode is used to start Windows with the bare minimum, to uninstall a faulty driver or to roll things back.


Simple method:


  • From Windows: Shift + Restart → Troubleshoot → Advanced options → Settings → Restart → Safe Mode.


If Windows keeps looping with a BSOD and you can’t get to the desktop:


  • You can force access to the recovery environment (WinRE) by interrupting boot several times (the PC powers off during startup), then “Automatic Repair” appears.

Step 3 — The “hands-on” method: from most likely to most decisive


1) Roll back what you changed


Ask yourself these questions:


  • Did the BSOD appear right after a Windows update?
  • After a GPU driver update?
  • After adding RAM / SSD / a PCIe card?
  • After a BIOS setting change (XMP/EXPO, undervolt, PBO, etc.)?


If so, the fastest route is often:


  • uninstall the last update / last driver
  • or roll back the driver (Device Manager → Properties → Driver → Roll Back Driver)
  • or restore a system restore point if you have one


2) Update Windows… but properly


Updates are still a logical step, but you need to be pragmatic:


  • Update Windows (Windows 10/11 depending on your OS).
  • Restart, then observe: if the BSOD comes back immediately, we stop the “update roulette” and move on to analysis.


3) Update the right drivers


The original article recommends Device Manager. It sometimes helps, but it’s not the most reliable source for critical drivers.


What works better:


  • GPU: via the manufacturer’s official tool (NVIDIA/AMD/Intel).
  • Motherboard chipset: via the manufacturer’s website (or official utility).
  • Network / audio: if you suspect an issue, the same applies (motherboard/laptop manufacturer).


Tip: if you suspect the GPU driver, do a clean reinstall (uninstall + restart + reinstall). And avoid stacking optimisation layers on top of each other.


4) Check RAM and stability (a very common cause)


If you enabled XMP/EXPO recently, or if you changed memory sticks:


  • Temporarily disable XMP/EXPO in the BIOS and test.
  • If that stabilises things, your profile was too aggressive (or your sticks/IMC don’t like this setting).


If you want a proper test:


  • Memory test (Windows tool or MemTest depending on your level), ideally several passes.
    An “random” BSOD in games plus heavy apps = unstable RAM very often.


5) Check storage + system files


When the BSOD is linked to the drive or corruption:


  • Check available space (a full SSD can cause weird behaviour).
  • Run system checks (SFC/DISM) if you know how, otherwise use a built-in repair tool.


6) Temperatures / power supply / overclock / undervolt


You have:


  • GPU undervolt “too low”,
  • RAM/CPU overclock,
  • aggressive curves,
  • or a borderline power supply?


Go back to the baseline system to test. A stable system first, optimisations afterwards.

Step 4 — Do a usable hardware diagnosis


You can use a diagnostic tool like UserDiag : it’s useful for getting a snapshot of the components, temperatures, and a report you can share.

The idea to keep: a clear report (temperatures, frequencies, RAM, GPU, CPU, storage) helps eliminate assumptions.


If the BSOD keeps happening, what you want to get:


  • the error code
  • the minidumps
  • a temperatures/stress test snapshot (without doing anything reckless if the PC is overheating)
  • the list of recent changes (drivers, updates, hardware)


👉 Without that, you can spend hours “testing things” without ever getting to the real cause.

what works / what doesn’t


What still works:


  • Boot into Safe Mode to regain control.
  • Photograph the error code and provide it to speed up the diagnosis.
  • Update Windows and drivers (once it’s done properly).
  • Use a diagnostic tool (like UserDiag) to objectively assess temperatures/stability.


What causes problems (the classic traps):


  • Update “all drivers” using dubious tools or at random.
  • Rely only on Device Manager for critical drivers (chipset/GPU).
  • Ignore the RAM: unstable XMP/EXPO = very frequent “random” BSODs.
  • Keep a borderline undervolt/overclock and look elsewhere.

Better to prevent than to cure


A BSOD gets fixed well when you follow a logic: info → isolation → test.

Start by getting the error code and the minidumps, boot in Safe Mode if needed, then tackle GPU/chipset driver first, RAM (XMP/EXPO), storage, and stability (temperatures/OC/undervolt).


👉 If you can only do one useful thing: photo of the code + minidump + context (what changed). That’s what turns a “blue screen” into a clear diagnosis.


To our gaming PCs

How to clean your PC: the complete guide for effective maintenance
Read Next