BSOD Unknown type - Windows 7 x64

[samwise210]

Hey all, I've been doing some tests on a computer to fix a frequent BSOD. I've been trying to duplicate it to no success, so I only have logs captured from previous instances.

I hope that some data can be extracted from this to point me in the right direction.

· OS - Windows 7 x64
· Full Retail
· Age of system - Motherboard was bought in 2009 but never used. CPU was a Kijiji purchase (which I'm worried about). GPU was pulled from a working system, put in use in 2012. Ram is all new, SSD is a year and a half old.
· Age of OS installation - 8 months, though the computer has not been in use for 3.

Desktop
· CPU - i7 920 c0
· Video Card - Sapphire Radeon HD 5750
· MotherBoard - ASUS P6T Original
· Power Supply - 500w PSU, unknown brand.
· 6 DIMM Corsair XMS3 DDR3-1600 2GB sticks

(The Graphics card has been replaced since the BSODs happened. It used to be a 4870.)

 

[Jared]

Looks like a CPU or motherboard failure, given we're seeing 0x9C and 0x101 errors.
Most likely the former though, where the WHEA structure hasn't been initialised.

We'll need a Kernel memory dump.

Go the Start
Right click My Computer
Select Properties
Click Advanced system settings
Click on the Advanced tab
Select Settings under Startup and Recovery
Then under Write debugging information select Kernel memory dump.

Once a dump is created go to:
C:/Windows/memory.dmp
Copy the file to the desktop, zip it up and upload it to a file sharing site like Onedrive. After the upload is done post the download link in your next reply.

 

[samwise210]

Damn, that's what I was afraid of.

Currently there's no kernel dump available - it's set to generate one but I've been trying and failing to replicate the BSOD.

When one becomes available I will upload it.

Tentative question, though. Do you think it's likely enough that it's the CPU that I should jump on a chance to replace it for ~~cheap?

 

[Jared]

Well, because, for some reason it doesn't show up, I was looking at dump files from June 2014.
However, the latest dumps are still from December 2014, with similar bugchecks. A recent 0x101 is there, so stick to what I originally said.

Error         : Internal timer (Proc 6 Bank 5)
  Status      : 0xfe00000000800400
  Address     : 0x0000388000e28ca3
  Misc.       : 0x0000000000007fff

It seems an internal timer error is consistent on the 5th memory bank, which could either be a coincidence or a fault in that memory bank.

Whether you want to buy a new CPU now is up to you, we could wait for more dump files to find a pattern or you can risk it if you so wish.

Apart from the crashes, your CPU seems to be very old, dating back to around 2008, the Bloomfield series if I'm not mistaken, an upgrade might be the best option anyway.
If you upgrade, do remember you'll need a new motherboard to fit the new sockets, the Haswell series is good (LGA 1150 socket).
I'm cautious about your PSU though, you say you don't know the brand?
If you use a PSU which isn't from a reputable company such as corsair, you can risk losing your components in the event of a failure, usually due to the lack of safety features in the PSU, hence why a lot of them are cheap. If you upgrade, it might be worth getting a different PSU, I recommend Corsair, personally. Their RM series are good.

 

[samwise210]

I've now had a stress test going for 5 hours with no problems, and I ran memtest overnight with no problems after 40 passes. These are things that would previously have caused a BSOD.

Without a kernel dump I'm aware that discussing the actual problem is hard, but I've got a couple questions despite that.

Is it possible that the problems were either being caused by overuse of the PSU (the old grapics card drew considerably more power than the new one), or simply by the GPU itself?

While I cannot put the old gpu back in to see if problems happen again, this might explain why I cannot replicate a bsod that was incredibly common before.