Random Reboots and freezes

[cshapeshifter]

Hi there,

4.5 years ago I bought my current system:
- Gigabyte GA-P35-DS4, Intel P35
- Core 2 duo E6850
- 2x1gb crucial ballistix CL4
- Enermax Liberty ELT500AWT PSU

A few weeks ago the graphics card (an nvidia 8800 GTS OC) went bust, so I had to decide to either buy a completely new set of components or just replace the graphics card. I don't game *that* much so I decided to just go for an nvidia GTX-560 TI which is just about appropriate not to be bottlenecked by the CPU, at least at 1920x1200.

In any case, The PC now randomly crashes and freezes in any of the following manners:

- Starts and hangs at POST (e.g. after counting memory or after 'verifying DMA pool data)
- Starts and runs fine, I even played Assassins Creed 3 and Crysis 2 for a short while without a problem, ran windows update and rebooted fine. But then...
- It crashes at any random point, e.g. during POST, boot up, while in Windows or Linux at any rate between a few seconds and several minutes
- From a cold start, nothing happens, black screen. Hitting the reset button doesn't do anything
- From a cold start, black screen but the PSU is making that 'tick' noise every second as if I were pressing the reset button all the time
- The reboots are either immediate (as if pressing reset) or off/on with a powered-off pause of 2 seconds.

Unfortunately I don't have any replacement HW I could test in it, but two weeks ago when the 8800 went bust, I had another GPU (a GT-240) in it and everything ran just fine. The PSU is strong enough for the system and all the wires are plugged in correctly. I tried resetting the BIOS.

It's seriously irritating that the reboots happen at any point and that sometimes, the Box runs fine for 20 minutes even when playing games and stressing both CPU and GPU.

Any ideas how I could figure out which component is the problem other than sourcing drop-in replacements? I don't have a volt meter but I could buy one, I wanted one for a long time.

Thanks!

 

[LowTech]

Have you check the thermal compound between the cpu and heatsink? Or check the cpu temps when it crashed?

 

[cshapeshifter]

CPU and GPU temps are just fine, the CPU goes to about 80° under full load. I have an enormous heat sink (Scythe SCNJ-1100P Ninja) and silver thermal compound, I'm pretty sure it's all good. Also, it even reboots spontanously right after turning it on, when all components are cool.

 

[Lasthitlarry]

Also 80C isn't that great, so I'm guessing the CPU is overheating and shutting itself off. Consider full clean down, re-apply CPU with paste, get better fans, try with case open, etc.

Other than that you've got an old PSU and a new GPU. I would take out your GPU and boot to onboard video and go into BIOS to check for voltage fluctuations and temperatures.

 

[discerning]

I believe that the mb the OP listed does not have onboard video.

I'm going to invoke Occam's razor and say that if you had no problems with the 8800 (until it broke) and no problems with the GT-240, the problem lies with the GTX-560 TI. Doesn't sound like you have another box you can test it in. Is the mb BIOS up to date?

 

[cshapeshifter]

Thanks for all the input, but I think I've found the culprit. Most unexpected fault in history:

Faulty reset switch!

I disassembled the whole machine, layed out the components on the table and started with just psu, mobo, ram and cpu. The system was stable and I worked my way back up putting everything back in the case, including all the disk drives and fans, everything.

Then I reconnected the front switches and the reboots were back that instant!

Somestimes it really isn't what you expect, but this explains why the reboots were truely random, it didn't have anything to do with any of the components...