Discussion on:

10 tips for coping with blue screen errors

I thought the BSOD went away with Windows 98. I guess I may have rarely seen it with WIndows XP, but never since Windows 7 came out. I certainly don't miss them. Windows 7 on a 64-bit pc is pretty darn solid from my experience. I certainly get clitches and issues, but I almost never need a hard reboot anymore. Programs may lock up, but the OS rarely does. Still, the tips are helpful to file away, just in case.

I use a program called "Who Crashed." It's free, and when software causes a Blue Screen,
it identifies the program. Very handy.

First, I've had one BSOD in maybe 5 years - unlike the days of Win 9x/ME.
#6 and #8 are generally together. Test the memory and if an issue replaced.
PSUs rarely would cause BSOD. If anything they would just shut down the computer. Always buy a PSU with more wattage than you need and buy something from a brand that you heard of.

I would say that more than 50% of BSOD's are from the Users not rebooting their computer for a long time. There are a lot of Users that think when they "Log off" that they are actually rebooting their computer. I always advise people to reboot their computer at least one time a week.

I have had a few issues with old video drivers and newer browsers/websites that stream video causing BSOD's also.

Very rarely does reseating hardware cause a BSOD. Hardware not properly seated will cause a PC to not boot and give a beep code.

Running the software on another computer usually doesn't work because chances are... you might not have another PC set up exactly the same just laying around.

Reproducing the error will not solve the problem but give you a better understanding of what is causing it.

If you open up a case and it is full of dust... I would definately clean it out with canned air and a hand-held vacuum. The processor might be overheating from the dust, but usually in this case, the computer will just immediatly power down when it gets too hot. You can also tell if a computer is overheating by listening to the fan speed. When it speeds up too fast it means that it is starting to get hot.

Don't remove your memory right away. If you have more than one stick, try removing and re-arranging the sticks. This will assist to pinpoint which stick is causing the error.

You can also run HD tests, boot tests, and memory tests from the BIOS.

If all else fails... I would strongly suggest calling your manufacturer ASAP to take advantage of any type of warranty. Most companies usually have free tech support as well.

You could also say goodby to Microsoft and say hello to either your choice of flavor of Linux, or go to a Mac OS-based system. Your stability issues will go away. However, if you choose to fire Microsoft, be aware you may have to budget for psychological/psychiatric counseling after suffering from depression from not seeing any more blue screens.

Okay, my comments may sound cynical and some corporate/government accounts do not--or cannot--have this option. But for those that have the leeway, you may wish to examine this route.

one of the major causes I have found is forced M$ updates which fail to fully install.
The first time I encountered this was on one of my own computers running an OEM install of XP.
Some update came though, but failed to completely load, causing the BSOD, I managed to force the system to uninstall about 10 updates until I got rid of the BSOD ... but then suddenly the registration timer pops ups stating I have 14 days to activate the OS !
Next re-start of the computer the timer maxed out and locked the computer. After trying almost everything possible to re-activate, I gave up and contacted M$ support ...I was on the line for over Two and a Half hours trying every combination of codes the tech could dig up, and in between, they kept asking if this OS had been installed on another computer previously, to which the reply was it was a factory install on that machine [ a Toshiba Notebook ] ... we eventually hit on an activation code which worked - and of course I wrote it down for future reference - which I did have to dig up another time ... ... ...

MS updates hose .NET so bad the drivers can't operate with the applications and just give up. It never fails for me that either recovering from a backup image or restoring back and re-installing all updates and holding off with .NET until the last, solves the problem. This is primarily on Vista/Win7 systems, but I'm working on an XP system that pretty much did the same thing. The intel built-in mobo video chip just could not take anything above .NET v 1.1. Finally had to buy an AGP card to quickly solve the problem. The OEM updates were no help at all.

I know your answer to this response is going to be "I tried that" but this has worked for me every time. When updates have failed for me, I noticed that some are required to install before others. If you select to install ALL of the updates, they might not update in the correct order. All you do is manually select and install one or a few at a time. If an update fails then don't select the update(s) until you are done installing the rest of the updates that work. Then after everything else is done... try the update(s) that failed at the end and I bet it works.

Before I start guessing and start ripping apart the hardware, I would turn to my trusted friend Windbg(debugging tools for windows), 8 out of 10 times I was able to resolve BSOD using Windbg.

BSOD, (******** overload) is a common problem which seems to be hitting a peak period just now in other parts of industry. Prevalent at first in the tech world, it has now spilled out and spread about like wild fire. The political world is becoming almost over saturated in BSOD. Doctors are concerned. more later...

If the frequency of occurrence of BSOD goes higher then it may cause a permanent loss of data. Whether its Windows or a Mac but, a clone of hard drive is always a great rescue in severe data loss scenario.

The pro techs are gonna know a lot more than I, but generally I've found it's one of four things:

1. "What's changed?" . . . usually the BSOD happens after someone installs new software or hardware.

2. Hardware . . . skunky video cards, network cards, sound cards, memory, whatever, who knows. Power supplies for sure.

3. Corrupt drivers and/or system files. A total PITA, but tracing those culprits is a long and thankless task (as we all know).

4. Motherboard . . . gotta single this one out as particularly nasty and pesky because when all else fails, the 'board is pretty well the only guilty party left once everything else is ruled out.

Step -1: Disable "Automatically restart on system errors" so you can actually SEE the BSoD text
Step 0: Digital camera, flash off, macro on; take pics of screen and content

Then I'd research the error text and context as per Step 0 (search-accurate text is easier if clients send pics rather than type out what they remember) on a working PC.

I wouldn't "try to reproduce the error", because every BSoD causes a bad exit, and therefore potentially loses files due to interrupted writes and file system damage. If those searches don't find anything specific, my generic steps would be:

Hard drive out of PC, into known-good PC
- Bart boot, HD Tune, SMAT detail
- if even slightly bad, file copy off, then BING and image C:, then surface test
- ChkDsk /F, check free space
- formal malware scans, etc.
Rest of PC
- eyeball capacitors, fans, etc.
- 24hrs+ in MemTest with boot disc ejected to spot restarts

Then when hardware is known to be good...
- Safe Cmd OK?
- Safe Cmd OK?
- MSConfig disable all startups and non-MS services, offline
- if OK, add back one by one on "test to break" basis
- if still OK, add back peripherals and external storage, one by one
- if still OK, check firewall is on, then go online
- if still OK, return to client for evaluation

Some further notes...

2) Logs to check include antivirus cleaning, ChkDsk/AutoChk "fixing", updates, "registry cleaners" (if any, I'd avoid), reliability monitor, Programs and Features sorted by installation date, and intra-application updates. I don't routinely update BIOS or other firmware and drivers, as this code should be far enough from the edge to avoid exploits, and yet is close enough to the bare metal to cause BSoDs and STOP crashes.

3) The article you linked, shows a different kind of blue screens compared to what we see in consumerland - we don't get nearly as much obvious detail. A reference I use is at Aumha.org, i.e. http://www.aumha.org/a/stop.php There is a risk to searching for error codes and file names on the Internet, as malware is often offered via 'fits-all' SEO pages that generally go "To Fix {your.searched.error}, Download This!"

4) You have Vista to thank for Reliability Monitor! That was the version of Windows that first gave this to us, as well as the first of an ongoing set of file operation dialog improvements. It amuses me to see Reliability Monitor highlghted as a "new feature" of Windows 7, and again with Windows 8 - though 8 does enhance program compatability and file operations dialogs.

5) Overheating isn't always a processor thing; in addition to CPU and GPU, PSU and hard drives can overheat (bad sectors), as can the motherboard. So I add a PSU-driven front case fan to blow over the hard drive, and motherboard-driven fan to dump air out the rest of the case. Bad capacitors can cause localized power glitches fast enough to BSoD rather than reset the system, and that too may be related to heat.

6) Because I routinely split hard drive and rest-of-PC to work on separately, I've had long durations of MemTest to compare with the testing period before first detected error. I've evolved "long enough" test periods from 6, 8, 12, 18 to 24 hours, as thus far I have had only one PC show the first error after 24 hours (just over 25 hours it was) while too many doing so after the shorter "good enough" test periods. Note that MemTest may not test everything, e.g. graphics or DMA access, full memory range, all processor cores etc. so read the ReadMes. I'm less keen on Microsoft's tester, as it attempts to load Windows to show results (!).

It's not only RAM test errors you are looking for; spontaneous resets, lock-ups, graphic trash etc. are also signs the hardware is unfit for OSs that write to storage. A reset can be missed if it restarts the memory testing, so I put in a different boot disc after MemTest starts.

8) Bear in mind that every time you touch things, wiggle in slots, etc. you run the risk of generating new errors through now-bad contacts or static damage, obliging the "best practice" practitioner to re-do the 24 hours MemTest.

Finally, I wonder if SFC checks the internals of files, rather than simply asking for version info, and whether AutoChk/ChkDsk communicates with SFC etc. when it "fixes" file system errors in ways that are sure to corrupt the affected files (crosslinks, truncations, etc.).

I had an issue where after upgrading to Windows 8, my PC started crashing.

I've now tracked it down to an overheating Northbridge chip. This despite the PC being an unmodified recent Dell desktop, less than 2 years old, and didn't crash in Windows 7! I've bought a small fan and fitted it to the top of the Northbridge's heatsink, and the problem is, I think, solved!

I used a utility called 'speedfan' to help me understand what was happening. Why is temperature monitoring not built into Windows?

I've been successful in using the minidump file that is created when the computer blue screens. It requires a download of the WinDbg program to decipher it. It is almost as cryptic as the blue screen message, but it has pointed me in a more direct path to determine what the cause was.

Make sure to set the minidump file where you will remember it (it also shows the location on the blue screen). Then open that file in WinDdb.

This is one of the best utilities I have come accross for troubleshooting BSODs. Whoever wrote this app, I give my sincere thanks. http://www.nirsoft.net/utils/blue_screen_view.html

So where does find the Reliability Monitor?

Make a photo of BSOD display!