At my wits end with failing components...

Northwind

Northwind

Northwind

Associate
Joined
31 Mar 2006
Posts
1,793
Apologies if this is a bit fraught :) Just done putting my backup PC back together yet again after my main rig stopped POSTing. Really just at a loss, any help would be much appreciated.

(History: It's a DS3 motherboard, Pentium D 2.8, 7950GT, 3gb of RAM. It was in a nice stable but loud build with an old Akasa PSU in til right before Christmas, when I decided to replace the overloud PSU with a nice new Corsair 520W. Fitted it up, it ran fine for a few hours, left it off overnight, and the next day all was dead. Christmas day as it happened.

So, called OC, and they said "It's the board, send it back". Got a replacement eventually, fitted it, and it nearly caught fire! Wee chip near the back smoked up and left a blackened spot the size of a 10p on the board. Not so good. So phoned OC again. "Ooh, you've got no luck, send us the board and we'll replace it"

So we did, same thing happened though this time I stopped it faster. Board still dead. Returned PSU and board and got a nice shiny new Corsair and a refund on the 'board. Got a new board off Trust.)

Just fitted it up last night and... Stop me if you've heard this one. Worked flawlessly for about 6 hours. Powered it down... And today it's dead, and I'm exactly back to square one and Christmas Day, but without the presents.

I'm pretty much at a loss, absolutely out of enthusiasm but also I've got no idea where to go with it. Nobody else reports faulty Corsairs... But it's not like there's much you can do wrong with a PSU! Plug 3-pin into ATX_12V, plug 24-pin into ATX (Power connector). But it seems like it's the logical place to point, the original board was fine for 3 months of constant use till I changed the PSU, and the new one was (I think) used in another machine without problems. The other 2 were brand new.

All other components are absolutely fine and now purring away in my orrible Dell... But the motherboard's almost certainly cooked I think.

Any advice?

This pic shows the chip that cooks itself, arrowed- after about 10 seconds powered on it's hotter than hell, it's only actually burned up the once but on the last but one board solder was visibly melting around it, and this time I killed the power as soon as it failed to POST- about 5 seconds. Even then, it burnt my skin instantly when I touched it, which I'm presuming is not right at all- soldering iron hot.

IMG_0528.jpg
 
CPU is fine, at least it works fine in the other chassis, no reason to doubt it. It's not lapped or anything. Also, everything was sunny last night, the only thing that changed was that it was powercycled. Not for the first time though, it was powered up succesfully 3 times before it stopped. This is pretty much identical to the original failure way back in December, it ran fine for the evening of the build (probably just one or two cycles there, it was a minor rebuild) but then 12 hours later was dead when I went to power back up.

Every other part that's in the main system is now in the Dell, right down to cables, apart from the motherboard, cooler and PSU. I think the might be compatible with the Dell's non-ATX board but tbh I'm not risking it, since the PSU is my main suspect and I can't afford to kill another board to test it.

The system doesn't post at all- could've been clearer there, sorry about that, what happens is it spins up the fans and then, well, nothing. USB ports aren't powered, no monitor output (it stays on standby) no beeps, no hard drive activity after the first second where they make the power-up noises but don't spin up after, none of the usual booting noises. I've tried it without RAM and there's no change. This seems to be the standard failure mode of a dead Gigabyte board these days...

The PSU's under warranty so ideally I'd like it to be that which has failed :) Though I'm sure OC would disagree :D So I'm thinking I probably need to go through the RMA game again, but it's really starting to stretch "bad luck"... The drain on the PSU shouldn't be the problem, the wee Akasa was only 400W and had way weaker rails.

Cheers for the help! All much appreciated.
 
Thanks for the input folks- epic response follows:

IAmATeaf said:
The mobo isn't shorting in the case is it? Have you checked things like mobo stand offs to see if there's 1 that shouldn't be there. Also how is your HSF mounted? Does it have a backplate that could be shorting? I've seen people recommend a couple of layers of insulating tape on the backplate to make doubly sure that the backplate won't cause a prob.
Click to expand...

The HSF has a backplate but I've previously insulated it, heard that one before. And all standoffs are present and correct.

neo202020 said:
Well, if you've gone through several PSUs and are having exactly the same problem I doubt it is the PSU. Shorting seems fairly likely at this stage.
Click to expand...

Just 2 PSUs, and the last was apparently tested by OCUK and confirmed faulty. The previous PSU was the constant through all the motherboards as OC pointed the finger at the boards.

Also, thinking along these lines further, neither the motherboard or the HSF were disturbed when I fitted the Corsair- it had worked fine for about 4 months with the Akasa then failed within a few hours with the Corsair. I'd have thought a short along those lines would have revealed itself earlier, it seems to push coincidence a bit far...

ALienBC said:
- there's a problem with the mains supply you're using
Click to expand...

Possible I suppose, but it'd have to be something that doesn't affect the Dell rig as it's all the same, right down to the same mains lead.

Big.Wayne said:
Surely that's a 4-pin? (or a typo?).

Buy yourself a PSU-Tester, the cheapest one I seen in an Antec PSU-Tester costs about £10.00.

What are those wires just to the left of the smoking chip? I think its the audio connectors for the case, are they fitted as a block or is it a bunch of seperate wires?
Click to expand...

Yep, typo, good catch. Didn't know there were PSU testers available. Good idea, I'll look into that... The wires are the audio connector, and it's a block connector.

Slime said:
Hi Northwind, sorry to be a pain but any chance you could give us a close up on the chip that's burning?
Click to expand...

I can have a crack later but tbh on this one it just looks normal... On the one that burned up completely it was a bit discoloured though.
 
Hmm, yep, got front-mount USB sockets but no bent pins... Used one of them in the brief period before it died too. I think I can discount the CPU tbh, since it's rock-solid reliable in a lower-quality board.

I could build out of the case but tbh I believe the board's dead. Still, something to try.

One thing which I should maybe stress is that the symptoms are exactly the same, right down to running for a while then failing after the next power down, as the first PSU which was confirmed faulty by OC. Now, I don't know exactly what they do to test it but for me that makes it the new PSU the most likely suspect...
 
I'm not allowed to send PMs, apparently...

I went through a stack of boards... Let's see, 3. The first one, the one that burned and the one that fizzed. 4 now. But remember, 3 of those were with the same PSU which was declared faulty and replaced. Even if you don't have a huge amount of faith in the RMA testing process, it's much more logical to work from the assumption that the PSU was faulty and that the diagnosis was correct, and that the boards were killed not faulty.

So I wouldn't jumpt to the assumption here that it's your DS3 that was the problem- it could be that this is the standard failure for a DS3 that gets zapped by a faulty PSU or power surge, and that yours has just gone the way of the dodo too. It'd make sense to have a sacrificial part that'll protect the CPU, expansion cards etc.

Apparently my local Electrical Goods Chain have a PSU tester in stock, so hopefully I'll have one tomorrow to test this PSU with. Apparently this particular Electrical Goods Chain is considered a competitor, despite the fact that their PC stock's antique and was overpriced when it was new ;)
 
I was dealing with OCUK for everything, not really sure of the legal standings on this- they say it's faster to go straight to the companies involved though.
 
Now entering WTF territory :) PSU tests fine. 12Vs are both a bit over but within acceptable deviations. So bang goes my suspect, and also any hope of fixing it on warranty :D TBH I really wonder now if the last one was faulty... Course, I don't know for a fact that the tester works right but that's just needlessly overcomplicating things.

I'm going to check the shorts line of enquiry... Does seem very unlikely since nothing was changed, but you never know.

BTW Slime, a Popular High Street Electronics Store has a PSU tester for £20 and apparently it's a stock item in most of their branches. That's the sort of tool that makes you more of a man, like owning a torque wrench or a welder.
 
You must log in or register to reply here.
Back
Top Bottom