OUCH! HELP!!

Gents,

My Primary rig has been acting up a bit lately. I've been having the 6GB showing up as 4GB issues, and over the last Day and a half or so, I've been experiencing random restarts (with no BSOD). This morning I woke up to the sound of all my fans going full blast and nothing but a blank screen.
Upon inspection I found this!!!!!!



Sorry for crappy pic(just sold camera), but essentially that's the 8Pin ATX 12V with all four outter pin sockets burnt to a crisp and the corresponding plugs either melted or exploded.

Now, this is a 24/7 crunching machine with reasonable voltages 1.275Vcore, 1.3v VTT. CPU and Chipset temperatures have always been kept quite low with reasonable ambients and overhead fans blowing over the heatsinks. PSU is behind a surge protector, and everthing else on the surge protector is fine.
The reason I am posting this in the benching section is because I figured this would be the place with the most exposure to things frying, melting or exploding. That and I know alot of you are more hardware savy than the average member.
Any ideas on what the hell caused this? Bad contact at the plug? Too much voltage? Cummulative effects over time at full load? PSU?

It goes without saying not only the board, but the PSU is pretty much toast, unless it tests ok and I can install a new plug on the ATX 12v cable.

I didn't have time this morning to lift out the CPU and check for burnage in the socket, I hope my trusty 950 is ok.

This is definitely the first time I have ever had this happen on a crunching machine (or any machine for that matter) and I'm somewhat at a loss.

This board was going in for an RMA for the Dead DIMM slot, but obviously now it's a bit more than that.

System in sig.

Thanks in advance for any input based on prior experiences.

This is usually caused by a surge or a power supply pushing to much voltage into a spot.

I've seen some go up into flames when this happens. Your lucky that it was just a burnt plug.

As for checking to see if it runs, I wouldn't try it. You may end up shorting something on the board and hurting the CPU or GPU or Memory etc. Just RMA it.

Yes I am lucky. It's really making me re-consider keeping my machine at load 24/7 crunching Rosetta.
Lol, like I said, I really don't know how long it was running like this as I just found it running dead after I woke up.

Bummer Pat THis is the rig in your sig?

Bad, it's all about what your comfortable with.
I'm no expert (far from it) but I have had this happen a couple of times. 1 time, cheap psu "popped" and took the board. May have been a failing psu or a surge, I don't know. 2nd time was a bad connection in the socket. Board was toast, but psu, ram, cpu all worked in a new board.

The way it was explained to me was the bad connection created resistance, resistance caused heat, heat caused more resistance. Or something like that, it's been a number of years since I had that happen.

Ouch, that sucks.

My money is on a bad connection, due to one of the following:

1) Not plugged in all the way. It happens to the best of us, i discovered my modular 24pin connector wasn't plugged into the psu all the way a while ago. Oops.

2) Pins in one half or the other were slightly bent and not making good contact.

3) It got slightly wet at some point and formed some corrosion, most metal oxides don't conduct for beans, if at all.

From what i've read, quality PSUs can often soak up a surge these days, so while it's possible a surge killed it, i wouldn't bet on it.

Sadly, i doubt the motherboard still lives, that little section got HOT. I wouldn't trust it.
I think the rest of the stuff should be fine, though that connector on the PSU is probably dead, too.
If you're lucky your >1kw psu is like mine and has two of those 8pin connectors.

At thoes voltages I'd have a hard time beliving that the overclock caused the failure but you never know how long something will work beyond its intended limits...

You already know that you need to RMA the mobo but I suspect the PSU had a part in that failure. Wether it was the only responsible party I dont know. Pull the PSU, compeltely disconnect it from your system and test it. Only problem you have right now is not being able to test it under the same load it was under when the failure happened. I'd probably try to RMA it as well JIC.

Sorry about the burn out man, good luck!

Thanks for the input guys.
Just to make things clear:
1. Yeah, there's no way in hell this thing is getting turned back on. That's a given.
2. Pins were ALWAYS all the way in. I always wait for the positive click and visual verification on Power and data connections b4 plugging the PSU back in.
3. Lord knows what happened to this board b4 I bought it, so yeah, maybe it did get wet/corosion.

Anyhow, waiting for my new board from Ghostrecon so I can get this one RMA'd, and currently in talks with O.W. to get a new PSU expidited out west here so I can get this thing back up and running hopefully.

Keep it coming if you have other thoughts.

Cheers,

IMO, a positive click doesn't mean all the pins are as tight as they should be. The click is the 2 plastic connectors. How often have you had a molex connector fully seated, but still have to wiggle wires to get a fan to run?

I am curious though, anything else burnt on the board?

Nothing else burnt as far as I can see.
And I can definitely see what you're saying about pins and sockets not engaging even if you hear the click and verify that the plug has fully engaged the outer socket. It very well could have been something that was not visible except if you really looked close.

Regardless, it looks like Benny might just beat me to 2million points after all. I'm totally outta commish.