CPU or MoBo fault?

crABtoad

Limp Gawd
Joined
May 15, 2012
Messages
508
Have a machine that blue screens when it's doing *anything*, it's stable if you just leave it alone.

I've reinstalled Windows, swapped out the GPU to a known good GPU, booted to Linux using a USB drive, swapped out the memory.

Only parts I haven't cycled out of the machine are the CPU, PSU, and MoBo. Is there a good way to determine if it's the CPU or MoBo on an unstable system?

I don't have a like-socketed CPU/board to test on or I would just isolate that way. It doesn't stay up long enough to run any real software on it, and honestly I'm not sure what software could check this.

Any ideas other than buy a new MoBo to continue testing?
 
My guess would be power supply. Get something decent but cheap after checking Hot Deals or Slickdeals for testing. It's always good to have a backup anyway. If that doesn't help then I would say motherboard. CPUs are rarely bad, and if so, most of the time they just don't work at all. Well, some who really push overclocks can have degradation over time, but I don't think that's your case.
 
That's interesting, I never would've suspected the power supply. I have a few spares laying around, I'm gonna kick myself if that's the root cause. It is a relatively new PSU I picked up as a replacement for the original, but that doesn't mean it didn't fail. I figured CPUs are generally "all bad" or "not bad", and that wasn't going to be the case but it was still a possibility on my mind.

Thanks for the reply yo! <3
 
Only way it is a bad CPU is if you OC'ed it to like 4.6 Ghz with a stock heat sink. :eek:

If it isn't the PS like pgaster said, then it pretty much has to be the motherboard.
 
I figured CPUs are generally "all bad" or "not bad", and that wasn't going to be the case but it was still a possibility on my mind.

CPUs can be in any state between "all bad" and "not bad" depending on the age and how hard of a life it has had.

I've had both CPUs that have been abused (overheated, overclocked to the moon) and CPUs that have been run into the ground (in a server with 100% CPU utilization 24/7/365) and they all exhibited different quirks. While most of them would boot into an OS, they'd start misbehaving once certain types of loads were performed. I've seen cache, FPU, IMC and IGP failures as well as certain instruction sets stop working correctly. Then there are the more strange failures where the CPU stops working correctly at its designated clock speed, but will work if underclocked, or even more strange, only work correctly within a narrow temperature range.
 
CPU was never OC'd, it's a 3770 non-K (for virtualization), had it on an AIO cooler since I bought it. I tested the PSU last night, it still bluescreens. Just gonna replace the MoBo next, there's only a couple parts it could be at this point. Thanks for the advice yall.

GiGaBiTe: I guess you're right about things failing in stages/parts, my experience with CPUs has been more of the "all or nothing" sort.
 
The problem with installing software to check the event logs is that the system will crash constantly when it's told to do anything. I've been doing hardware isolation to troubleshoot as a result. Next step is to buy a new motherboard, I think I'll try to find an ITX Z77 board and swap it out next. Seems that the chipset is old enough that it's difficult to find ITX variants though. Maybe time to just buy a new CPU as well.

@Nobified: Swapped out the memory awhile ago, issue persists. The RAM I tested with was known good.
 
I guess you can't check the CPU temps with SW, did you try re seating the CPU cooler with new thermal paste?
 
Try HWInfo 64/Sensors - check voltages and temps.

CPU voltage < 1.3V?
PSU - 12V, 5V and 3.3V within 0.2V of target?
CPU temp < 50C?
 
Hey, sorry for the late reply. I saw those over the weekend but I was otherwise occupied. Reseating the cooler sounds like a great idea, I'll try to get around to that if I have a quiet hour or so anytime soon. I've basically shelved this project until I can get a motherboard for it, but fixing it without parts is even better. Checking SW sensors is problematic, it's super twitchy, bluescreens basically once it does anything at all.

Thanks for the continued advice guys. I'll update this thread if the reseated cooler or new mobo fixes it. :D
 
Back
Top