Nonplussed - Windows 7 random BSODS

[Carol Haynes]

Totally confused ... asked to fix a Dell computer which is constantly crashing with BSODS and pretty randomly.

It is a Dell desktop running Windows 7 x64.

Tried all the usual things and gave up and reinstalled the factory image. Everything seemed OK but next day the BSODS started again.

Had the computer in my house:

1) Reinstalled factory image again
2) Removed crapware/ Norton etc.
3) Updated from Windows update and all drivers/BIOS/software updates recommended by Dell
4) Installed MSE  and other common tools
5) Ran stress tests and memory tests for 48 hours.
6) Produced Image Backup of system

No logged errors and no BSODS throughout.

Took it back to owner and BSODS start again almost immediately.

Removed everything except screen keyboard and mouse and BSOS continue to appear at random.

Anyone got any ideas?

The commonest STOP codes are 0x1e, 0x7d and 0x3b but they all seem to be random. Most hint at driver issues but they are factory drivers and don't seem to cause issues in my house. All very odd!

[Renegade]

Dell? Look into graphic cards and 64-bit compatibility. Just an observation from personal experience.

[Carol Haynes]

Worked fine for a couple years?

[Renegade]

Well, with the Dell that I have I found that the 32-bit system is fine, but the 64-bit OS just blows on it. It's one avenue to look into.

My biggest problem was with working with a monitor attached to the laptop - constant BSODs. It's ok without a monitor attached, but... well, it might help as a clue to whatever problem you're seeing. Not sure.

[Ath]

Check the power-source (low voltage? ), and possibly new office-equipment sending out 'bad vibes' over AC-power.

[mouser]

I always suspect memory chips.
I would swap out all but one, see if problem remains.  If so, replace it and try once more (use a dif slot too).

[IainB]

If it was previously fine in the user's site, and if it worked fine and passed all CPU/RAM/disk tests whilst sited in CH's house environment, but now promptly BSODs again on return to user's site, then I would suggest an analysis of the user's environment. Something may have been changed in the user environment.

In particular, I would suggest consideration be given to the possibility of "unclean" mains power supply fluctuations, nearby strong or high frequency radio transmissions, and proximity EMI (electro-magnetic induction). Also check that the mains has a PME (positive multiple earth).

EMI is pretty rare. I have on only one occasion come across a definite case of EMI in computing, and that was in a mainframe computer room environment, where the cause was one data-carrying cable lying on top of another, and we solved the problem by running the uppermost cable along a 3-inch high metal bridge, over the lower cable.

Also, don't overlook the possibility of a poorly mounted chip, or loose cable connection, or dry joint on the circuitboard (the bad connection may have been accidentally temporarily reconnected by moving to CH's house).

[worstje]

What other hardware does the user have at home? I recall reading about a Microsoft kernel/driver engineer who at one point had similarly crazy problems, and he found out that his external harddrive had a bug in its driver that caused such BSODs. (He also managed to get the bug pushed to the vendor and fixed, so that specific one shouldn't be an issue anymore).

Regardless, I can well imagine it being some sort of weird interplay. One wonky USB port, one wonky device, or maybe just some mutual sneaky dislike somehow. It wouldn't be the first time the problem was sneaky like that.

[40hz]

Long shot - but since you seem to be there already...

Check the power supply. I've had many weird lockups and hangs under Windows (that I thought were caused by software because of the stop codes) turn out to be attributable to a desktop Dell power supply doing intermittent 'weird things' with the electricity.

Had one client who had really severe voltage drops in his house most times the AC came on. That would cause spontaneous reboots. In our shop it would be fine. It was only when he took it back home it acted up.

He told us he had a UPS so it didn't occur to us he really meant he had a surge suppressor. Eventually it damaged the power supply since too low a voltage can cause transformers to run really hot.

Luck.

[Deozaan]

Krishean found out that connecting his Nexus 4 to his PC and rebooting the phone would reliably cause the PC to BSOD. Maybe it's something similar.

[Vurbal]

If it was previously fine in the user's site, and if it worked fine and passed all CPU/RAM/disk tests whilst sited in CH's house environment, but now promptly BSODs again on return to user's site, then I would suggest an analysis of the user's environment. Something may have been changed in the user environment.

In particular, I would suggest consideration be given to the possibility of "unclean" mains power supply fluctuations, nearby strong or high frequency radio transmissions, and proximity EMI (electro-magnetic induction). Also check that the mains has a PME (positive multiple earth).

Dirty power would definitely be my first (or next) suspicion based on the OP. That's common here in Iowa, particularly in rural areas and in older houses which are rarely grounded properly (or at all) regardless of what electrical upgrades have been performed over the years. It can also be caused by something as basic as a faulty power cord or surge protector and it could even be isolated to a single circuit within the building.

Questionable quality power supplies, or even power supplies which aren't designed to deliver significantly more wattage than the computer requires are more likely to be affected by this. Assuming the is the original OEM power supply both probably apply. A cheap ATX power supply tester might show the problem, but it's equally (or even more) likely it would only be detectable under a normal operating load.

Even with the right equipment tracking down power issues is something of a black art. Without that equipment even an expert is stuck with educated guesses combined and swapping or bypassing components.

[Stoic Joker]

Check the power-source (low voltage? ), and possibly new office-equipment sending out 'bad vibes' over AC-power.

+1 - Just had one awhile back that would reboot every time the client printed...or a fax came in. Turns out they had a multifunction laser printer plugged into the UPS.

Edgy PSU also sounds plausible.

[mouser]

+1 to the idea of making sure the bsods aren't happening when some device (external hd, etc.) is plugged in.

[TaoPhoenix]

What other hardware does the user have at home? I recall reading about a Microsoft kernel/driver engineer who at one point had similarly crazy problems, and he found out that his external harddrive had a bug in its driver that caused such BSODs. (He also managed to get the bug pushed to the vendor and fixed, so that specific one shouldn't be an issue anymore).

Regardless, I can well imagine it being some sort of weird interplay. One wonky USB port, one wonky device, or maybe just some mutual sneaky dislike somehow. It wouldn't be the first time the problem was sneaky like that.

I'm going to borrow this comment.

I'm fighting my feeble memory and losing, but I had a case at my old job sorta like this.

Blah Blah, comp works, then all of a sudden it started bsod'ing, so it wasn't power supplies etc. I swear it was a *software program* that did it, and (don't hurt me if I mangle this vague reason) but it was something like the program interacted in a weird edge case with the driver, so then it became a "driver problem" causing the bsod. And if I had to pick a driver, I'd say it was maybe the video driver dealing with dual screens and maybe a program that didn't have that properly supported.

[Vurbal]

^ IME video and USB are the most common culprits of mystery driver problems. Video because so many system (software) components interact with it in a multitude of different ways and USB typically because Windows gets confused when a single USB port is used for too many different devices - particularly when there's storage involved.

[Carol Haynes]

Thanks guys ... currently the system is stripped down to monitor, keyboard and mouse ... otherwise nothing plugged in and all components are Dell factory components.

Must admit I thought about the PSU and also if it was something environmental. I suppose something could have shaken loose when they drove it home too.

[xtabber]

If you have a voltage regulated UPS available, I would try plugging the system into that at the client's location and let it run for a while.  If that makes the problem go away, you almost certainly have a marginal PSU that is cutting out on voltage dips.