In The Trenches – Troubleshooting a Blue Screen of Death

Today while I was onsite, I came across a laptop that was getting a Blue Screen of Death (BSOD) every single time it started Windows. In this article, I’ll show you how I troubleshooted and solved this particular problem using two past “Repair Tools of the Week”.


When booted, this computer would show the Windows XP logo , show the cursor for a second and then suddenly produced a Blue Screen of Death with the following message:

IRQL_NOT_LESS_OR_EQUAL
STOP: 0x1000000a (0x49df8170, 0x00000002, 0x00000000, 0x804d9afa)

In most cases, “IRQL_NOT_LESS_OR_EQUAL” either means an incompatible or corrupt driver or bad RAM. I know this from past experience but don’t be afraid to Google it because you cant remember everything.
I was obviously unable to do anything further now that the computer had BSOD’d so I rebooted it and tried to see whether I could get into safe mode and luckily for me, I could. This strengthened the argument that this blue screen is probably due to a corrupt or incompatible driver since Windows doesn’t load many drivers in Safe Mode and bad RAM would probably continue to cause problems even in Safe Mode.

This job was made a little tougher because due to the maximum resolution of the screen I was working on, the information that tells me what driver caused the blue screen was either cut off or simply not there.

So, I loaded the tool BlueScreenView which we have featured as a Repair Tool of the Week in the past and it showed me all of the previous Windows crash dumps this computer has generated and the information they contain.

I chose the most recent crash dump and in the lower pane the application had highlighted the two files that caused the crash. One was “ntoskrnl.exe” which is a key system file. The other one was “w22n51.sys” which is an Intel wireless network driver and is most likely the cause of our problem.

Armed with this new knowledge, I closed BlueScreenView and opened up another tool we have featured here called Autoruns. Autoruns will show the operator all of the files that will start with Windows such as services, applications and of course – drivers.
I went to the “Drivers” tab and scrolled down until found the offending wireless driver and deselected it. This way, it will no longer load when Windows starts and hopefully not cause a blue screen of death.

I actually tried to uninstall these Intel wireless drivers from within safe mode using the “Add/Remove Programs” section. However, the uninstaller makes use of the “Windows Installer” service which doesn’t run in safe mode and while you can do a little registry tweak to get it running (which I tried), it still wouldn’t uninstall.

I rebooted into normal Windows and it powered up to the desktop without any problems now that the offending driver no longer loads up. I went back into “Add/Remove Programs” and removed the old drivers in the proper manner (now that I could) and rebooted the computer.

I then visited the laptop manufacturers site using its wired connection and downloaded the latest wireless drivers for the computer which were about 3 years newer than the previous drivers. My guess was that either the driver had become corrupted or a new Windows update was incompatible with it. In either case, reinstalling the drivers should fix the corruption or perhaps the manufacturer has fixed the compatibly problem in the last 3 years..

Once I installed the new drivers, I connected the laptop to the clients wireless network and browsed around a few websites. Everything seemed to work and I rebooted again for good measure (should the current drivers also cause a BSOD) and everything seems to be working fine. I rebooted the computer once more and ran Memtest86 (which is on UBCD) for about 5 minutes to be certain it wasn’t caused by bad RAM. Memtest86+ will read and write to the RAM to find problems and while a proper test should be run for hours, I have found that most RAM that is bad enough to stop Windows from booting will fail the test within the first 5 minutes.

After 5 minutes, I was confident that the issue was definitely the wireless driver and not the RAM.
I was happy, the client was happy, so I got paid and went back to the office. The End.