Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

My computer is suffering from seemingly random crashes, that get progressively worse the longer a computer session runs.

Initially, the system will start, appear to run fine for some time, then a BSOD will appear out of nowhere with a random error. Most often, this error message contain something related to 'USB', for example BUGCODE_USB_DRIVER (but it is not consistent). If I reboot after this initial crash, the crashes become more frequent, and I get less and less further into the boot process. [Only when I take an extended break from a computing session does the lifetime seem to improve again.]

i.e. My most recent session went something like:

  1. Run for 45 mins in Windows 7 64bit: BSOD with 'BUGCODE_USB_DRIVER'
  2. Reboot; get to login screen; BSOD with 'BAD_POOL_HEADER'
  3. Reboot; get to Windows loading screen; image appears to freeze on screen, computer becomes non-responsive.
  4. Reboot; get to a BIOS/POST screen (prior to being able to enter setup); image appears to freeze, computer becomes non-responsive.

I had not changed any hardware or major software configuration before these crashes begun to happen (they appeared to start 'out of the blue')

I have stripped down my hardware to the bare minimum: motherboard, CPU, Memory, and replaced my fancy ATI video card with an old video card (which has been proven to be stable), and 1 hard drive (with OS), to no effect. My Windows software has the latest updates, as well as my hardware drivers.

I had originally suspected it as a memory problem, so replaced all my DRAM, but the problem still persists (unlikely to have received two bad batches in a row...)

Current specs: AMD Phenom II X6 1090T ASUS Crosshair IV Formula 4x4GB G.SKILL Ripjaw 1333 [This is the new memory]

...any suggestions? I'm all ears.

share|improve this question

migrated from serverfault.com Aug 17 '11 at 19:01

This question came from our site for professional system and network administrators.

3  
The fact that letting it rest for a period changes the behavior indicates that it is heat related. Note that I say related not caused. It could be something as minuscule as a bad solder joint on some surface mount component. These sorts of problems are notoriously difficult to diagnose. Usually it comes down to replacing every piece of hardware one piece at a time until the problem goes away. –  EBGreen Aug 17 '11 at 19:20
    
Yargghhh... it's the sad but obvious case I didn't want to admit, then. =\ –  David Elner Aug 18 '11 at 18:26

3 Answers 3

Start by checking temperatures of all major components. Hangs that come at higher frequency when device is "warm" would point into some component overheating. Open side panel, start some programs that load the system, put your finger on all heat sinks and memory modules in turns. If you can't hold your finger there at all, they are way too hot. If they feel warm, not hot, then they are OK.

The next on check list is memories. Run a memory diagnostics program for at least overnight; some tougher problems might require multiple days to be found. Even while you just swapped the memories, still do this; this is one of the easiest things to test without swapping hardware, and it might as well be memory controller on the CPU or memory lanes on motherboard that are faulty. Lowering the bus speed might help you buy more time on that front, if memories produce errors.

Does turning power off totally from mechanical power switch (on back of the power supply, usually) for 15 seconds or so help at all? BSODs at increasing frequency could theoretically be caused by some device collecting bad data to buffers or similar, to this is less likely.

All in all, it is pretty clear some part of your machine is failing. What we can do to help is to make informed guesses about what to replace next, otherwise you need to do it at random. But hard to work around replacing something.

share|improve this answer
    
Sounds like a plan, I'll try to do some more testing for overheating. I've been routinely disconnecting the power completely after troubleshooting sessions, but not in between reboots. What are your thoughts about the possibility of a corrupted BIOS? Most of the BSODs I have gotten are 'USB related'... –  David Elner Aug 18 '11 at 18:30
    
BIOS corruption should be easy to clear; unless I'm mistaken, just flashing the BIOS with the newest one should clear any corruption. –  Zds Aug 18 '11 at 20:21
    
Flashed my BIOS from an old revision (801) to the newest stable one (1902), and the crashes persist: I'm ruling out the BIOS. Also reduced the memory to a single (new) stick, to no effect. Checked all the identifiable heat sinks, and none were 'too hot' to hold a finger to. Next step... replace the motherboard? –  David Elner Aug 22 '11 at 0:07
    
Motherboard would be my take, yes. –  Zds Aug 22 '11 at 8:28

Sounds like a hardware problem, caused by heat to be precise. Did you overclock your processor or RAM, or at least meddle with these BIOS settings? Please try to reset your all BIOS settings to factory default. I concur with the previous reply, as the next thing to suspect, I believe is the RAM too. Memory diagnostics may take long time (esp if you have 4x4GB!), so try on the RAM one by one, at different RAM sockets. I had a similar problem which was rectified by unseating and reseating the RAM a few times. Worst case scenario, you could have warped the board. :(

share|improve this answer
    
Though I have some OC-tinkering friendly hardware, I haven't had the guts to actually mess with those settings: no overclocking done here. I'll wipe the BIOS tonight, and test each memory module individually. And what exactly is a warped board? Sounds bad :( –  David Elner Aug 18 '11 at 18:32

Warped board means the motherboard is deformed physically, causing connections to some of the surface mounted components to be unreliable. Google "warped motherboard"!. Usually microscopic cracks formed in the solder between these components cause unpredictable crashes and may be heat related, as the board warps more when overheating occurs. IF(!!) the board is really warped, it is usually a goner. Some people try to bake it in oven, in an attempt to melt the solder and hope to bridge these cracks (as is done in xbox motherboards), but I havent tried it before (and wouldnt advise it!). As for the memory modules, my PC exhibited the same symptoms as yours, as was easily rectified by some "jiggling" of those memory modules. I guess it was due to some dirty connectors. I hope it works for you too. Good luck!

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.