trying to diagnose repeated BSODs...help appreciated!

[niccolo]

I have an old T42 that's still going fairly strong (with some help from the excellent James Arndt). But I recently installed a new, larger hard drive, and I get BSODs sporadically. I've formatted and reinstalled Windows XP and all my programs, installed relevant Windows updates (e.g. service pack 3), run thorough anti-virus checks, but it still happens sporadically. I only have very basic programs installed, Office and STATA statistical software. The BSODs don't seem to be associated with a particular program, although it does seem to happen more frequently when the computer is coming out of sleep mode (although it also happens at other times). The error message I get is copied below, I'm afraid I'm not sophisticated enough to interpret it. Can anyone shed any light? Appreciated!

STOP: 0x0000008E (OxC0000005, 0xBF80A61E, 0xBA2129D8, 0x00000000
win32k.sys - Address Bf80A61E base at BF800000 DateStamp 4a85564f

[Harryc]

Try reseating your RAM. Try one module at a time. I deleted your other thread. We do not allow cross (duplicate) posting on the forums.

[niccolo]

James Arndt did install additional RAM along with the new hard drive, and that's when the problem began. But the BSODs happen infrequently enough, about twice a week, that it would be hard to diagnose whether this was the underlying problem. I suppose I should remove the new RAM module and leave it out for a week or so and see whether the computer BSODs? Is there any way to diagnose this absent removing the RAM?

[Kyocera]

You say it's intermittant, can you look into your admin tools> event viewer and possibly associate a specific time/event with the BSOD.

[niccolo]

So the error message I posted above doesn't provide helpful time/event information?

Assuming not, how do I glean this from the "event viewer"? I'm afraid I'm not a terribly sophisticated user...appreciate your help!

[Kyocera]

Go to admin tools, event viewer, in the left pane you'll see selections for different types of "events" in vista it's broken down pretty specifically. Select a category in the left pane and the events will populate in the larger frame. Scroll through and look for red and yellow exclamations. Try to match up an event with a timeframe, the times and dates are listed, then you'll see an event ID.

This is just one additional way of attacking problems, this has been very helpful for me in the past. So much so that I try to check my watch or the windows clock when working on a pc that is having weird issues.

[frankiepankie]

Try running Memtest (bootable CD) for a night, and look if it finds any errors.

[niccolo]

Appreciate the helpful comments.

The computer hasn't BSODed in the past two weeks, so I haven't had an opportunity to use the event viewer. There were a fair number of past red and yellow events flagged, but they didn't seem to match up with the BSODs--that is, at least the most recent BSOD didn't seem to have anything associated with it in the event viewer, and the other events didn't sound like BSOD events in the descriptions.

I have downloaded Memtest and will report results as soon as I run it.

A faulty connection to memory seems all the more plausible because the BSODs seem to be a bit more likely when I am using the computer on my lap, rather than on a desk.

[SafeHarbor]

See if you can find a memory.dmp file in C:\Windows. If so, you can download the tools from Microsoft to analyze the dump file and possibly find the cause of your BSODS. When I tried it last August, I found one left over from the previous October on my HP desktop.

Here's a walk-through I wrote that should help.

http://www.brighthub.com/computing/wind ... 29860.aspx

[niccolo]

I installed the debugging tools, but can't find the "memory.dmp" file. It's not located in the Windows directory, and a file search for that title yields nothing. The computer last BSODed two weeks ago, is it possible the file was not preserved, or not created at that time?

Memtest gave me a clean bill of health on multiple scan passes, but that's not surprising if the problem is a sporadic bad connection to the memory, rather than than internal to the memory. Still, good to rule the latter out.

Appreciate all the help--this forum is great.

[RealBlackStuff]

These files normally go to \Windows\Minidump
Also, go to Control Panel/System/Advanced tab, and check the settings under Startup and Recovery.
See what it says under: Write debugging information.
You may have set this to NO dump.

[niccolo]

I don't seem to have a Windows\Minidump folder (or file).

In Control Panel\Systems\Advanced, System failure is set to "Write an event to the system log," Write debugging information is set to "Small memory dump (64 KB)," Small dump directory is set to "%SystemRoot%\Minidump" but I'm not sure what the % syntax refers to, and there is no SystemRoot folder in either C: or C:\Windows.

Appreciate all the help.

[niccolo]

P.S. I'm running Windows XP, in case that's relevant.

[RealBlackStuff]

"%SystemRoot%\Minidump" translates to "Windows\Minidump"
%SystemRoot% could be Win, WINNT, Windows, whatever the ROOT (base-directory) of your system is named, when it was installed.

[niccolo]

In C:\Windows, there is no Minidump folder--when sorted alphabetically, it goes from Microsoft.NET to msagent. There is also no file named minidump. Searching for all files named "*.dmp" yields nothing. Searching for "memory.dmp" also yields nothing. Searching for "memory" yields some files, including a file named "memory" located in the debugging tools for Windows folder I just downloaded, of type "H File." So it appears my computer was not writing dump files when it BSODed, although it was set to do so. Very odd. It still hasn't BSODed again, either--going on three weeks, whereas in the past it would do so every few days. Maybe it's getting bashful about all this BSODing now that I'm investing so much effort trying to fix it?

[RealBlackStuff]

Also check in the same System failure page, that Automatic reboot is UNchecked.
If you have no more BSODs, start counting your blessings...

[niccolo]

Alright, have unchecked "Automatic restart."

Should I perhaps create a "minidump" folder? In other words, right now the system is set to write to the minidump folder, but there is no such folder, so perhaps this means it can't write anywhere and that's why I'm not finding the dump files?

[RealBlackStuff]

Not to worry. The system will create that subdirectory when required.

[niccolo]

The computer didn't BSOD for seven weeks, so I was hopeful the problem had somehow been resolved. But it just did it again. I've noticed recently that my desktop background image has been gone after I start the computer back up from sleep. I just did that, and noticed the image was back again, which was odd--and right then the computer BSODed, and now that I've restarted, the image is gone again. Does this suggest some kind of hard drive corruption? Do the error messages I posted previously not help diagnose this? Are there other places on the computer I can get information to help diagnose this? Under administrative tools/event viewer, the crashes don't seem to be creating any error or warning events. And there is no "minidump" folder, and searching for "memory" or "memory.dmp" or "*.dmp" doesn't seem to yield relevant files.

[rkawakami]

I would re-visit an overnight session using memtest86+ and also run a disk diagnostic; just be sure that you have the entire system backed up in case something goes wrong with the disk test. Since you have mentioned that the BSOD seems to occur when the system is coming out of sleep mode, at the same time as you are running those tests, I would also subject the system to a "colder-than-normal" environment. Don't go as far as putting the laptop inside the refrigerator but if you could run the tests with the T42 inside the coldest room in your house then that might confirm or deny my suspicion that you are looking at a temperature-related issue. Before running either memory or disk tests, allow the system an hour or so to stabilize at "room" temperature. Turn the system on and as soon as possible, start the diagnostics. Also, when returning the system back to regular room temperatures, be mindful of any condensation which may occur; again, give the system some time to stabilize at the warmer environment before turning it on. This shouldn't be too much of a problem if you keep the temperature differential to no more than 15 or 20F.

That the BSODs have appeared at around the same time you have changed both the memory and hard drive systems you could also replace either one of those things and see if the problem goes away. The easiest, and perhaps the cheapest, would be to swap out ALL memory modules with different ones. After installing the "new" memory and before putting the system back into use, run the memtest86+ tests for at least a couple of hours (3 or 4 complete passes through the suite of patterns) to confirm that you have no "hard" failures.

As far as the disk diagnostics go, you can either run the Fixed Disk test in PC Doctor for DOS:

http://www-307.ibm.com/pc/support/site. ... 56222.html

or see if the manufacturer of your disk drive has their own diagnostic. Seagate drives have SeaTools, Hitachi drives uses DFT, Fujitsu has FJDT, etc. Google "hard drive diagnostics <insert drive manufacturer name>". Those tools should have a basic sector read verify test to check that no errors occur when reading all of the sectors on the disk.

Intermittent errors are the hardest things to track down and can drive you crazy.

[niccolo]

I downloaded the Western Digital Data LifeGuard Diagnostics, and ran both the Quick Test and Extended Test. Both reported a simple "PASS." However, when I click on the SMART Disk Information, the entries seem to suggest major problems, assuming I am reading them correctly. For example, "Raw Error Read Rate" gets a green check mark, but the values are ID:1, Value:200, Threshold:51, Worst:200, Warrant:1. Or the "Re-allocated Sector Count" gets a green check mark (everything does), but the values are ID:5, Value:200, Threshold:140, Worst:200, Warrant:1. I have no idea how to interpret this data, and google isn't helping. But it appears to me that I am getting worst scores across multiple categories, right? Does any of the above shed any light on the BSODs?

[niccolo]

I think I was misinterpreting the drive data reported in the prior post, i.e. "worst" is not the worst possible result, but rather the worst result the system has obtained during a scan. The only thing that matters is whether the result is above the threshold, and all of them are--the drive gets a clean bill of health, hence all the green check marks.

I re-ran Memtest overnight, clean bill of health as well.

I haven't experimented with temperature, I'm afraid there's not a lot of variation in my apartment. But if this were temperature related, what would that point to? Hard drive issues? Memory issues? Motherboard issues? There's a slight correlation between using the laptop on my laptop and BSODs, although only slight, i.e. they also happen when the laptop is being used on the desk. So my best guess is a faulty connection to memory, but one that only very sporadically disconnects. But given the BDODs on the desk, not sure that finds much support.

[rkawakami]

Coming from a background of memory test, I'd suspect the modules to be temperature sensitive, but really, anything that has electronics on it would be possible sources of the problem. My suggestion to run the system at lower than normal temperatures is essentially to try to change the failure mode from intermittent to consistent. That would make troubleshooting the error is little bit easier.

I don't know enough about the SMART information you are reading off the drive to say what that data really means.

[niccolo]

In the quest to reproduce the BSODs and narrow things down...tried moving/flexing the laptop, and also tried tapping/etc. on the RAM directly (albeit gently), nothing.

Will try to see if there's a temperature correlation. The problem, if you want to call it that, is that it BSODs so infrequently. But perhaps cool temperatures will exacerbate the problem.

Perhaps a registry cleaner is up next...recommendations on whether and if so, which one?