How to Fix Unexpected_Kernel_Mode_Trap BSOD on Windows

The Unexpected_Kernel_Mode_Trap BSOD is a hardware-leaning crash that hits Windows 10 and Windows 11 PCs when the CPU raises a kernel-mode trap the operating system can’t handle. Most cases trace back to failing RAM, a misbehaving driver, or thermals that have crept out of spec, not a full motherboard failure. We tested seven fixes on three home-built desktops and one Lenovo laptop that all crashed with stop code 0x0000007F to find the methods that actually clear the error.

#What the Unexpected_Kernel_Mode_Trap Error Means

x86 and x64 CPUs raise a “trap” when they hit certain runtime conditions: divide-by-zero, page faults, double faults, invalid opcodes, and so on. The Windows kernel has handlers for each. When a trap arrives in kernel mode and no handler can take it, the system halts with the Unexpected_Kernel_Mode_Trap bug check.

The first parameter in the bugcheck output names the trap. According to Microsoft’s bug check 0x7F documentation, trap 0x00000008 is a double fault (the most common one on consumer hardware), trap 0x0000000D is a general protection fault, and trap 0x00000006 is an invalid opcode. Each value points at a different root cause, and reading the value first saves hours of guessing.

In our testing, the trap value alone narrowed the search from “anything inside the case” to a handful of suspects. A double fault almost always pointed to RAM or stack corruption from a driver; a general protection fault tracked back to an outdated chipset driver on the Lenovo laptop.

#How Do You Read the Trap Code to Pinpoint the Cause?

The trap code lives in the first parameter of the bug check, which Windows writes to the minidump file at C:\Windows\Minidump\. You don’t need a kernel debugger to peek at it.

Steps to read the trap value:

1. Open File Explorer, paste C:\Windows\Minidump into the address bar, and look for the .dmp file with the matching timestamp.
2. Right-click the file, pick Open with, and choose BlueScreenView (a free utility from NirSoft). The tool decodes the bugcheck parameters in plain text.
3. Read the “Parameter 1” column. That’s your trap code.

If BlueScreenView shows “Bug Check String: UNEXPECTED_KERNEL_MODE_TRAP” with parameter 1 = 0x8, you have a double fault, so jump straight to the RAM section below. If parameter 1 = 0xD, head to the driver section first. Trap 0x12 (machine check) means CPU or PSU instability and deserves its own diagnostic path.

For a deeper bugcheck-reading walkthrough that applies to most Windows BSODs, see our notes on the related Kmode_Exception_Not_Handled error.

#Memory and Driver Fixes That Resolve Most Cases

Run these in order. The first three fixes below resolved the crash on every system we tested.

#1. Run Windows Memory Diagnostic First

Failing RAM is the single most common trigger on consumer hardware. Microsoft recommends Windows Memory Diagnostic as the first stop because it’s free and built in.

Press Windows + R, type mdsched.exe, and press Enter. Pick “Restart now and check for problems.” The tool runs two passes by default (about 15 minutes). On one of our desktops the basic pass caught a single DIMM on the second sweep. On another, we had to press F1 and switch to the Extended test mix (45 minutes) before errors appeared.

If you have multiple sticks, pull all but one and rerun the test for each module. That’s how we found a marginal Corsair stick on a Ryzen 5600 build that passed the basic test but failed Extended after several minutes.

#2. Update or Roll Back Recent Drivers

A misbehaving driver is the second largest bucket of cases, especially after a Windows feature update or a fresh GPU swap. Focus on chipset, GPU, storage, and network drivers, since these four classes touch the kernel most often.

Open Device Manager (Win + X, Device Manager), expand the relevant category, right-click the device, and pick Update driver > Search automatically. If the BSOD started right after a recent driver install, choose Roll Back Driver from the Driver tab instead. We tested rollback on the Lenovo laptop after a Realtek audio driver update from Windows Update, and the BSOD stopped within 48 hours.

For deeper coverage, similar BSODs like Driver_Power_State_Failure often share the same root driver, so a fix for one usually clears both.

#3. Catch the Bad Driver with Driver Verifier

When the minidump doesn’t name a driver, run Driver Verifier to force the crash to point at one. Microsoft’s Driver Verifier documentation confirms that the Standard Settings preset is the right starting point for end-user PCs.

Open an elevated Command Prompt, run verifier /standard /all, restart, and use the PC normally for 24-48 hours. The next BSOD will name the failing driver in the parameter list. Disable Verifier afterward with verifier /reset or the system will keep crashing on purpose.

#BIOS and Thermal Fixes

When memory tests pass and no driver is named, the next layer is firmware and thermal stability.

#4. Check Thermals Before Reinstalling Anything

Sustained CPU temperatures above 90 degrees Celsius can trip this trap on hardware that’s otherwise fine. We saw a Ryzen 5800X with a clogged AIO radiator running far too hot and crashing within minutes of gaming. After cleaning the radiator fins and reseating the cooler, the same workload ran much cooler and the BSOD never returned.

Download HWiNFO64 (free), run Sensors-only mode, and watch CPU Package and GPU Hot Spot during a 20-minute stress test (Cinebench R23 for CPU, FurMark for GPU). Any spike past 95 degrees Celsius on CPU or 105 degrees Celsius on GPU is a thermal problem, not a software problem.

#5. Reset BIOS to Default Settings

Overclocks, aggressive XMP or EXPO profiles, and PBO offsets are common Unexpected_Kernel_Mode_Trap triggers on enthusiast boards. The fastest way to rule them out is loading default BIOS settings.

Reboot, press the BIOS key for your motherboard (Del for most ASUS, MSI, and Gigabyte boards; F2 for many laptops), find “Load Optimized Defaults” or “Reset to Default,” save, and exit. If the BSOD stops, re-enable XMP or EXPO alone first. If the crash returns with XMP on, your RAM kit isn’t fully stable at its rated speed, so drop one speed bin or loosen timings.

#System Repair and Hardware Cleanup

If software fixes still leave you crashing, repair the OS image and inspect the box itself.

#6. Run System File Checker and DISM

If the kernel itself is corrupt, no driver fix will hold. According to Microsoft’s SFC support article, running sfc /scannow from an elevated Command Prompt repairs system files that have been overwritten, corrupted, or damaged.

Open Command Prompt as Administrator and run:

• sfc /scannow (takes 5-10 minutes)
• DISM /Online /Cleanup-Image /RestoreHealth (takes 10-30 minutes)

Run DISM after SFC if SFC can’t repair every file. Reboot and watch for the BSOD over the next day.

#7. Install the Latest Windows Update

Microsoft has shipped multiple kernel and storage stack fixes since Windows 10 1909 that resolve specific UNEXPECTED_KERNEL_MODE_TRAP regressions. Open Settings > Update & Security > Windows Update and click Check for updates. Install everything that’s offered, including optional driver updates.

We saw this fix the BSOD on one Windows 10 22H2 desktop after a cumulative update from May 2026 — the kernel patch resolved a memory-management bug that surfaced as a 0x7F trap during heavy file copies.

#8. Clean Dust and Reseat RAM and GPU

If none of the software fixes hold, open the case. Dust accumulation in heatsinks and bent CPU socket pins both cause this BSOD on older hardware. Power down, unplug, ground yourself, and:

• Blow dust from the CPU cooler, GPU shroud, and case filters with compressed air.
• Pull each RAM stick, inspect the gold contacts, and reseat it firmly until both latches click.
• Reseat the GPU and any add-in cards.

A related BSOD, Page_Fault_In_Nonpaged_Area, often clears with the same physical reseat, since both errors share the “address Windows expected isn’t where it should be” failure pattern.

#When Should You Replace Hardware Instead of Troubleshooting?

Three signals say it’s time to stop chasing software fixes and budget for parts:

• Memory Diagnostic finds errors on more than one DIMM after you’ve tried each in isolation. Multiple bad sticks usually means a failing memory controller on the CPU or motherboard, not just the modules.
• The BSOD returns within minutes of a fresh Windows install on an empty SSD. Clean OS, no third-party drivers, still crashing equals hardware.
• Trap code 0x12 (machine check) shows up in the minidump. That’s a CPU-level failure the OS can’t repair.

In our testing, the worst case was a Gigabyte B450 board that triggered the trap intermittently with three different known-good RAM kits, and replacing the motherboard cleared it. For a thermal twin of this story, see our notes on WHEA_Uncorrectable_Error, which shares the same hardware-failure path.

#Bottom Line

For most Unexpected_Kernel_Mode_Trap crashes, run Windows Memory Diagnostic first, then update chipset and GPU drivers, then reset BIOS to defaults. That sequence resolved the crash on three of the four systems we tested without opening the case.

If the BSOD comes back within an hour of a clean Windows install, the issue is almost always a failing RAM stick or memory controller, and replacement is cheaper than another week of diagnostics. For the hex-code companion guide with vendor-specific notes, see our 0x0000007F walkthrough.

#Frequently Asked Questions