Lots of various BSOD, suspecting motherboard?

dany123

Honorable
Oct 5, 2013
4
0
10,510
Hello everyone, I've been having lots of problems with my pc lately so I decided to check with you guys for ideas.

(damn, this post got long:(, i tried to format a bit the important areas)
tl;dr version: Currently I can't use my 5 year old pc because of errors: from (rarely) not booting, to hang at "Initializing usb controllers" to BSODs or restarts after post, after the Windows 7 text appears, after the Windows 7 logo appears (very often), the logon screens (also very often) and from time to time it lets me log on but bsods after 1-2 minutes. I got a pair of RAM sticks from a friend, reinstalled Windows 7, booted Windows XP from cd - all nothing changed. The only possiblities left are that the CPU, GPU or motherboard are amalfunctioning and i don't know which one is the problem so i can replace it.

(Also note I've googled and read a lot of posts and i tried to answer most of the obvious questions below.)

PC Specs (bought separately 5 years ago):
CPU: Intel E8400 - never tried to overclock it, stays at 40-50*C
MoBo:MSI P43 NeoF
RAM: 2xADATA 4GB DDR2 800MHz CL5 Gaming Series Dual Channel Kit
GPU: HIS ATI Radeon HD4850 IceQ 4 512MB DDR3 256-bit - it gets pretty hot in idle at 50+*C but the fan is only at 30%, if i set it manually to 100% it goes down to ~40*C even with some load so i don't think overheating is a problem, just badly programed fan to prioritise silence
PSU: Sirtec - High Power HP-450-A12S @450W
HDD: 1xWD 640 GB, 1xSeagate 1TB (this one only a year old)

(bsods minidumps below)

The story:

So, 3-4 weeks ago for some reason (no new hardware or something suspicious) my pc started having a few blue screens right at the logon screen (as far as i remember) but after 2-3 restarts it would all go fine and work for 10+ hours with no problems. Then, I went away for 2 weeks on holiday and when I came back but there were way more blue screens at various times - even before logon screen, and while doing random things like maximizing a youtube video or right clicking on desktop.

Now, the 2nd or 3rd day when i tried to use my pc there were just too many bsods to the point that i couldn't log on after something like an hour so I decided to do some more troubleshooting: full chkdisk, the sfc /scannow, antivirus/malware scans but no luck, so I decided to exclude any software problem with a fresh reinstall of windows but no change... more and more bsods.

Currently I can't use the pc at all, I managed to get into safe mode for a few minutes to copy the minidumps to a memory stick and then it crashed (for some reason trying to load up firefox crashed the program, and after a few tries it did a bsod).
So that's the current state: random blue screens and when I get into windows random programs crash

Short infos:

After googleing some error codes from the bsods I thought the RAM was the problem so I ran 6+ hours of memtest86+ but no errors.
- Yesterday I borrowed a pair of 2x1Gb ram sticks from a friend but no change in errors (yes, i tried various combinations and all the slots) so I excluded RAM from possible culprits.
- I unplugged all non-essential things from the pc and no difference
- I thought maybe the Windows 7 or HDD or sata cable were damaged so I unplugged both of my hdds and ran the "Mini Windows XP" from Hiren's boot cd but it still crashed (i still could use it for a while though) so HDD aren't the problem either
- I cleared the dust and re-checked all the connectors but nothing changed but i noticed something yellowish on top of two of the condensers (or capacitor, or how it's called), this made me think that the motherboard has failed : https://dl.dropboxusercontent.com/u/45931250/bsods/20130928_101233.jpg
- I removed everything from the pc case and assembled the motherboard with only ram, gpu, cpu, hdd (also tried with only dvdrom) and power supply but still the same errors. So I figured out the only possibility was the CPU, motherboard or GPU - I couldn't get another gpu to test and i don't have an integrated graphics card so i can't test it (I did use Video Memory Tester but no errors were given);
- I tried testing the CPU with prime95 but it either gives me an error: "rounding was 0.5 expected less than 0.4" or it simply crashes after a few seconds/minutes - so CPU might be faulty too.

- I never overclocked anything
- the RAMs appear at their correct timings in bios (although they run at 1.8V but on them it says 1.9-2.1V)
- sometimes it doesn't even post, but after a restart or two it works
- Very rarely (i think 3 times) but maybe it helps, the pc froze while I was in bios trying to modify stuff.
- I should also mention that random applications crash while I manage to log on, like explorer, firefox, chrome, CPUz, games
- also while logged on I got a message that the display driver has stopped responding but was succesfully recovered
- tried modifying some settings in bios but I couldn't find anything that influenced the errors


I have looked at some bsods and their parameters are pretty weird:
- some drivers try to read/write to adresses like 0x0....0 and 0xFF....FF which as far as i know shouldn't happen
- one of them indicated some wrong bit in a CPU register (i don't remember exactly) which worried me it might be cpu's fault
- memory management errors everywhere
- lots of error's messages include kernel, irqls, page faults and PFN list

The conclusion:

That's all the info I can think of right now, if you need more feel free to ask me.
The reason I'm here is that if I buy a new motherboard and discover that it was my CPU or something else then I'm screwed because I have to buy a very specific one that supports DDR2 and lga 775 CPU so if then i find out that the cpu was the problem then i'd have a buy the same generation of cpu and i'd rather upgrade if i actually have to buy new parts.


bsods_5.10.2013.png


Archive with current minidump folder:
- Some of the more common BSODs are: 0x7E, 0x7F, 0x1A, 0x50, 0x0A, 0x1E, 0x7A and 0xF4 (there are more BSODs happening than the ones in the minidump folder that somehow don't get stored there, for example probably today i got 20+ BSODs but there are only 7 in the folder)
https://dl.dropboxusercontent.com/u/45931250/bsods/Minidump.rar
 

dany123

Honorable
Oct 5, 2013
4
0
10,510


I would do that but I don't have another one around and I don't think someone would lend me theirs, I do have my old pc around but I don't think it's powerful enough. Also, I don't think the PSU is the problem because I did use FurMark to stress the gpu and that should drain a lot of power but it didn't crash then. Also if it was fluctuating it would also have happened in the 6+ hours of memtest and crashed it but it didn't.
 
It's not unknown for a PSU to partially fail, I've seen it before where the unit had a good 12 volt line but the 5 volt line was dying so I really think you should invest few minuets and swap power supplies. As long as the old PC PSU has the correct motherboard leads it should power the system, even if you have to use a molex/PCI-E converter lead to run the 4850, as long as you do n't try any stress tests on the GPU a 250 watt OEM PSU will run that rig perfectly well.
Another option is to remove and reseat the CPU, the pushpins used by the stock coolers are known to give trouble over the years, perhaps one has worked loose or the CPU connections have gone bad over time.
Of course it could be a motherboard or CPU problem but unless that old PC has a LGA 775 socket motherboard in which you can try your E8400 the only troubleshooting option will be to take it to a shop.
 

dany123

Honorable
Oct 5, 2013
4
0
10,510
Ok, so I looked at the old PSU and on it said its total output is 400W and I connected it to this pc and with the help of the PCI connector you mentioned I was able to connect everything.

But the bad news is it's doing exactly the same as the old one, [strike]weirdly, in this order (meaning if it posts succesfully and i restart it it will post succesfully again)[/strike] (nevermind, i got the Init. USB controllers one again, it's randomly one of these) it either:
- doesn't boot at all (not even the post beep) - but if i remove the ram it beeps 3 times -> the code for ram problem
- tries to boot but it gets stuck on Initializing USB controllers...
- POSTs succesfully and tries to load windows but it restarts (or just freezes) right before the Windows logo appears - no matter which option I choose : safe mode, normal mode or repair windows

So this makes me really think that the GPU is somehow involved. I'll try to find another gpu or to give my gpu to a friend to test it but i'm having a hard time finding one to accept this...

I forgot to mention: I've reseated the CPU but no luck. :(



Quick update: I've tested my GPU on a friend's computer (by playing some games at high resolution and quality) and it worked perfectly so now I'm sure it's the motherboard or the cpu. I'm planning on buying a new one soon ( I'm also considering buying a used one from a friend but I'm not sure if I should because it might get broken in a few years too ).