Help with sporadic BSOD

[Pyromaniac]

I haven't seen anything on here to suggest linking a post in another forum is against the rules, please correct me if I'm wrong. Basically it saves me a lot of time doing this, and I haven't had much help from there unfortunately.

http://forums.techguy.org/windows-7/...bsod-help.html

Thanks!

[jolly17]

Could be overheating, or bad RAM, or bad BIOS settings, or bad Windows install, or bad HDD. Hmmm, could be many things!

Try running Prime95 and see if that makes your PC cry. If it does, it's probably not the HDD. Most likely overheating or too much OC. memtest would test RAM, but also the memory contoller on the CPU, so that would not tell us the whole story maybe.

[Pyromaniac]

It's not OC afaik, because today I reset to 3.2GHz and it still bluescreened within an hour. I'm most way through another chkdsk /r after trying /f so i'll see what that does. I ran furmark and my temps stabilised withou any crashing, so i doubt it's the GPU overheating. I ran coretemp for a while and the max was 42, so i doubt that either, I can try setting it to log, not sure if that would work if it bluescreens. Given the nature of this (more frequent under load, but still happening even when idle). One thing that may be significant is that it seems, (I need to try a bit more to be sure), that bsod's or a black screen will occur quite soon after I attempt to backup steam files, even though the steam directory and the backup location is on a separate hdd to the OS. That makes me think that it's not the HDD, as does the fact that seatools ran a long generic test and found no errors. Could this be RAM? I ran the windows memory test ok, but that was only 2 passes. I'm trying and failing to get memtest86+ to boot from usb, I can't find any CD's at the moment. Could it be my RAM? This is relatively new, (about a year, maybe under), and it's corsair, but I guess nothing is faultless. Can motherboards overheat? It is hot here, but I have loads of case fans (on low, but still), and it feels fairly cool in there. I downloaded prime95 earlier, I'll run it now as chkdsk just finished

Thanks

[jolly17]

Maybe it's power? Looks like you have a good PSU, but what about UPS? Try another outlet?

Unless your mobo feels hot, or the back of it on the case side, then I don't think your mobo is overheating. Could still be RAM, though. Only 2 passes of memtest? Try letting it go all night. THat'll show you!

[Pyromaniac]

Well, I didn't have to wait long for that one, prime95 consistently killed my system, either by BSOD or hard reboot, in about 20 seconds. So we can rule out the HDD and presumably the PSU. The mobo heatsinks are pretty warm, I'd say about as hot as a fresh cup of tea. If you're from the US/ anywhere else, I apologise, coffee will do just fine
I really know nothing about how hot those heat sinks are supposed to be.

[jolly17]

Oh, that sucks! Have you tried the fan trick? Open the case and point a house fan at it? Maybe time to redo the thermal paste?

Probably not power. Even with a maxed CPU, gaming would still push your computer higher with the added graphics.

[sassie05]

Could it be my RAM?

Absolutely, could be Your RAM.

I'm assuming you have 2 x 2GB sticks of RAM.

Try booting using one stick at a time.

Run MemTest.

[Pyromaniac]

I'm going out today so I've left memtest running. When I get back I'll have time to try prime95 with eack stick. Are those heatsinks on the mobo even removeable?

[jolly17]

No, not really. If your mobo is really overheating then you'll want to claim the warranty or buy a new one. I mean, you CAN work on your mobo but where'd you get the parts??

[Pyromaniac]

So I ran memtest86+ for 11 hours today, and it found no errors, so I think we can rule out memory. Now I have a reliable way of causing a BSOD, I should be able to see if there's and major heating increase.

[Pyromaniac]

Well, I'm now thinking it's CPU temp related, since it always seems to crash at about the 55C mark. I underclocked to 2Ghz, and it's been running prime95 for some time now without crashing, with the temp hovering at about 54C. I'll try back at stock speeds and see if it's always the same temp.

[Pyromaniac]

Prime95 may have been a false positive. It didn't crash at stock settings, and when I set it back up to 3.7GHz with ai overclock on auto, it seemed to be pretty stable, although I stopped the test once it got to 64C. It's probably my dodgy overclocking causing voltage problems, I've never put it under such heavy load before. Also all those BSOD's were 0x124, whereas the ones before were a mix of mostly 0x50 and 0xA. Also those happened even when idle. I ran asus probe II and it gives the mobo temp as pretty constant 36-37C, although I'm sure the heat sinks are hotter.

[jolly17]

Hey do you have a UPS? One of the good ones with the electricity protection?

[FTT]

Are your fans still set to 'low'? Can you try setting them to auto or full?

I have a mobo with removable heatsink on the northbridge and I had an overheat problem with it but a very small clip-on armchair fan facing the mobo with the side panel off stopped the problem. I ordered a sm. mobo/chipset fan but haven't fitted it yet but I did reapply thermal paste but nothing improved. It's not the first time I've had to install a fan on a chipset.

Basically, you can't expect gaming intensity as well as whisper quiet at the same time and expect your system to 'keep it's cool'. You need to assist the cooling, not dampen it in, 'the heat of the moment' with fans set on low.

If you're overclocking either or both of your RAM and or your CPU, let your fans do their work. Stop holding the reigns back. Giddyup hossey!

[Pyromaniac]

Well it hasn't bluescreened today yet, so I'm hoping it was an overclocking problem which has fixed itself, but failing that I'll have to turn the fans up and see if that sorts it. The problem is that I have no controller, only 3-stage switches inside the case, and a resistor cable on one. They do get pretty loud on full, which is going to be a pain when I'm not gaming.

[Pyromaniac]

Ok, I spoke too soon, it seems ok when not under load, but it bluescreened twice when I tried to play BC2 today. The second error was something about bugcode USB driver and was 0xFE I think. Not sure if there's anything I can do about that, would some more minidumps help? I might try updating the bios. Is a clean install of windows likely to help?

[Pyromaniac]

Actually, after I updated the BIOS it seems ok, it hasn't bluescreened at all and I accidentally left prime95 running for ages and nothing bad happened, so it might be fixed.

[Paul Komski]

I might try updating the bios. Is a clean install of windows likely to help?

Both are possibilies. Most of the time BSODs are directly or indirectly software related with specific programs (including BIOS routines) and inappropriate/faulty drivers being at the heart of things most of the time. As with all sporadic problems it can be very hard to nail down a culprit.

If updating the BIOS does it for you then great. Sometimes simply resetting to defaults works, which tells you something else. On some boards a replacement is the only thing that gives one peace of mind.

[Pyromaniac]

It seems to have become less frequent, but the problem is still definately there. the latest one was 0x3b system_service_exception. I might as well upload the most up to date minidump folder. If nothing shows up from there, then I'll clean install windows and see if that fixes it.

[Pyromaniac]

Probably would have helped if I'd remembered to upload the folder

http://forums.techguy.org/attachment...1/minidump.zip

[Pyromaniac]

I had a 0x21 earlier, so I decided to re-install windows, I just had another 0x1a memory management, so it doesn't seem to be driver related. I'll try running for a while on each memory module to see if it really is bad RAM.

[Pyromaniac]

The frequency of these seems to have died down a bit since the BIOS update, but they're definitely still there even after reinstalling windows. It's difficult to say if they're actually less frequent than before though, because I've probably used this machine less than usual recently due to being on holiday. It's definitely not fixed though, can someone have a look at those minidumps in case something shows up? I'm guessing this is down to hardware though.

[mjc]

Grab a Linux LiveCD and boot to it...then find something for it to do...like compile the latest kernel (grab the kernel source, extract it to a folder, then type in 'make config' from a console, opened in that folder...accept the default config then type 'make'). Or you can, from a console, type in

echo “scale=XXXX;4*a(1)” | bc -l (for the xxxx I'd put in 50000...yes, that is 50,000...)

That will calculate the value of pi to how ever many digits you specified...50,000...that should keep it busy for a good long while. (if it doesn't run long enough, you can always up the number of digits...and it will entirely eat up at least one core while calculating...)

Depending on the machine that should keep it busy for a while. And test the RAM/CPU at the same time...

If it is getting the same errors, as before, then it is hard to say...because not one of them screams 'I'm a hardware problem'...they COULD be hardware OR drivers...OR...

And, yes, if it is driver related, it can come back AFTER a fresh install of Windows...if the same drivers that caused it in the first place have been reloaded.

[Pyromaniac]

Right I'll do that soon.

[Pyromaniac]

I'm trying that Pi thing, but how do I know if it's working? Other than the fact I get an error message when I type it in deliberately wrong, there's no way to see if it's actually working.