Good Computer goes BAD

[mhz]

With a lot of advice from this forum, I built the computer in my sig about two weeks ago. It went fairly smoothly and I was able to attain a pretty decent OC of 2800 (280 x 10) 1.58 vcore + VID. It was stable for several hours under Prime95, but the temperature (at least according to MBM5) was 60C and I couldn't see running it that way for an extended period. So I backed way off to 2600 (260 x 10) 1.36 vcore + VID (46C under load).

 

I ran it at that OC for a week or more without any problems. Then one evening while playing Far Cry (got to love free games), the computer lost power. It started at the press of the power button and after some head scratching, I started playing again. About 45 minutes later it happened again. There were some thunderstorms in the area and although I didn't notice the light flicker, my VCR (yes, I still own one) was flashing the time, so I knew that there had been at least one power fluctuation. So maybe its not my shiney new computer's fault, right?

 

The next day while I was booted into the Maxtor Powermax CD, the computer lost power again. In each case it would power back up again at the press of the power button. Later that day I was messing with IDE cables (powered down of course) and connecting a different IDE drive. When I tried to power up, it was a no go. I got the 4 LEDS lit and nothing else. To make a long story short, I ended up pulling everything off the motherboard and power supply until it was just CPU, memory, vid card, sound card. No go, same symtom (4 LEDS).

 

Next, I pulled out the motherboard (still on the tray) and low and behold, it goes through the LED sequence. I added CD drive and it booted. I went to BIOS defaults. A step at a time I added everything in the sig and all seems well (overnight mem test, an hour of Prime95).

 

Anyone have an educated guess at what caused the power drop outs? Or why it wouldn't boot till the MB was out of the case (perhaps disconnecting and reconnecting to PS)? I did some searches and found some similar symtoms, but no definitive causes. Some resulted in RMA of MB or PS. I bought most of the components from Newegg and I think they have a 30 day RMA policy, so I only have a few days to decide to take that path.

 

Thanks for looking.

[THunDA]

So its working ok out of the case now ?

 

If so put it back in and be sure that nothing is shorting out anywhere.. Its possible something might have touched the case housing and caused this..

 

Did you try and clear the cmos ?

[mhz]

Its all back in the case and working. I never had to remove the MB from the tray. I was looking for something that may have caused a short all along and never found anything. When I tried the MB out of the case, it was mounted to the tray and the tray was resting on the case, so if it was grounded out to the tray, it should have behaved similarly to when it was in the case. At least, it seems so to me.

 

As far as clearing the CMOS, I didn't get to that step. It started booting again before I got around to that. I did try the safe boot option (MB jumper).

 

At this point, I want to tidy up the cables a bit and then i plan to wait and see if it happens again before departing from the default BIOS settings.

 

If I live happily everafter, I guess it will all remain a mystery.

 

Any suggestions?

[wevsspot]

I had a similar experience with my Ultra-D motherboard when I first got it (sans the thunderstorm part of your story).

 

Strange infrequent power downs over the course of a week. Sometimes just sitting idle, sometimes overnight, sometimes when gaming or surfing the web.

 

I replaced my aging power supply with a new OCZ 520W and problem has not resurfaced. I see you're using the same power supply, so I doubt that's the problem.

 

If you're unsure I'd RMA the psu just in case since it's still under warranty. If not, just chalk it up to the DFI gremlins and hope all stays well.

[Burky]

I had a similar experience with my Ultra-D motherboard when I first got it (sans the thunderstorm part of your story).

 

Strange infrequent power downs over the course of a week. Sometimes just sitting idle, sometimes overnight, sometimes when gaming or surfing the web.

 

I replaced my aging power supply with a new OCZ 520W and problem has not resurfaced. I see you're using the same power supply, so I doubt that's the problem.

 

If you're unsure I'd RMA the psu just in case since it's still under warranty. If not, just chalk it up to the DFI gremlins and hope all stays well.

The exact same thing happened when I got my SLI-D board too, and once I took it out of the case it worked great. When I put it back in the case after running it out in the open for two days it worked fine.

I learnt the solution off someone else too, so it sounds like this is not uncommon for these boards.

[Slammin]

The exact same thing happened when I got my SLI-D board too, and once I took it out of the case it worked great. When I put it back in the case after running it out in the open for two days it worked fine.

I learnt the solution off someone else too, so it sounds like this is not uncommon for these boards.

 

 

I don't think 'these boards' have anything to do with it. Sometimes something get's shorted when you assemble it. Be glad the board is robust enough and you had a good enough PSU to survive such misfortune.

 

I had a fan controller short out on a system that had an Ultra X-Connect PSU in it and the short burned every fan wire attached and the PSU never shut down, though the CPU finally got hot enough for a thermal shutdown. Needless to say, that mobo was shot.

[mhz]

Just a quick follow up here. I ran Prime95 torture test overnight (about 9 hours) and no errors. I guess the important thing is that it was still running the next morning!

 

CPU temp got up to 49C. It idles at 38C.

 

One thing that occurred to be was that I set the BIOS to shut down if the CPU got to 65C. I don't think that it got that high, but could an erroneous temp reading cause it to shut down unexpectedly?

 

With the current default setting for the BIOS, the shut down feature is disabled. This could explain the more polite behaviour. Just a theory.

[boatasiaus]

I would enable the thermal shutdown function.

 

Better safe than sorry, especially if you leave the comp on overnight unattended

[mhz]

Thanks for the suggestion, but I am going to leave it at BIOS defaults for a while. If I don't have any power drop outs, at some point I will enable the thermal shutdown again. If I start having the power issues again, I will suspect that BIOS feature or an erratic thermal sensor to be the culprit. It really bugs me to not know. Its possible that I had a short on my previous build and that it will not show up again. Only time will tell.