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
 
That little chip is a voltage regulator (they have a technical name - but i cant remember it) If its getting hot, that would suggest its getting too much voltage. So it sounds most likely to be psu problem to me.
 
Start the build and try turning it on with;

PSU Mobo CPU - Beeps to say "No RAM"
PSU Mobo CPU 1 ram stick - Beeps to say "No Graphics"
PSU Mobo CPU 2 ram stick - Beeps to say "No Graphics"
PSU Mobo CPU 3 ram stick - Beeps to say "No Graphics"
PSU Mobo CPU 4 ram stick - Beeps to say "No Graphics"
PSU Mobo CPU 4 ram stick & GFX card
then with sound card etc etc

Keep checking to see when the problem starts.

If it starts even with just the psu/mobo/cpu... and you've changed the mobo/psu a couple of times now, then maybe it's the CPU?

Is your CPU lapped? (copper from during lapping could be shorting pins?)

Edit: But it works fine in your dell mobo..? Have you tried powering the dell with the corsair psu? It's wierd. I had a nightmare build over xmas too mate, only got it sorted last month. My method ended up just keep troublshooting, selling, replacing bits until you've got rid of the **** that's causing all of the problems :p
 
Last edited:
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.
 
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.
 
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.
 
Yeah, did you use the little metal stand-offs which hold the motherboard off the case? (they're like screws with screw holes in):

images


My mate didn't use them and he smelt burning on his mobo, realised what he'd done and was fine (luckily) after putting the standoffs in.
 
Indeed... I'd say, with all these components failing, it has to be something else common to all of them. Three things occur to me immediately:

- there's a problem with the way you're mounting the motherboard, introducing shorts as mentioned above

- there's a problem with the mains supply you're using

- you have an evil ninja cat or something that sabotages the system whenever you're not looking

abc
 
If he did'nt have the standoffs fitted the I/O ports and pci slots would'nt line up. You can fit them where there are'nt supposed to be any however. I have had a couple of boards that have the holes but with no metal (solder?) around the edges. These are not supposed to have screws through them and can cause shorting if screws are used.
 
Hi Northwind, sorry to be a pain but any chance you could give us a close up on the chip that's burning?

My computer was working fine until sunday when it shut itself off and now refuses to post. I smelt burning but i cannot for the life of me figure out what it was. Upon closer inspection, the very same chip looks discoloured on my board and i was wondering how yours looks.

Alternatively, if anyone else could provide me with a pic of one that is working, i'd be eternally grateful.

Thanks
 
Northwind said:
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).
Click to expand...
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?
 
Just had a look in my DS3P manual and it is the front panel audio header. It's a block of wires by the looks of it. My TT Armour has similar blocks of connectors.
 
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.
 
The Antec PSU testers are great but don't keep it plugged in for too long as they can burn out.
However, I have tested power supplies with the Antec that have passed on all power rails and the PSU's have still been faulty.

Are you using front mounted USB connectors on the case?
I've yet to see one kill a motherboard but I've seen them with bent pins where the customer has been heavy handed and it's caused a short to the motherboard stopping it from powering up.
I've come across CPU's that have developed a short and killed a motherboard but that's exremely rare.
As you've had the PC powered up for a few hours before it dies, it doesn't sound like a dead short but perhaps an intermittent fault which can be a real pain to ressolve.
Another odd one I have seen before is static build up in a case that has caused problems.
If you can, how about building the PC out of case onto a desk or a bench?

In all the years I've been building PCs I've yet to see an optical drive, hard drive or floppy to be the culprit to kill a motherboard but it is possible.

Good luck and let us know when you find the culprit.
 
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...
 
Northwind, don't worry about the picture.

I've had another look at the same chip on my board (also a DS3) and it's completely burnt out. There's a brown mark surrounding the chip and it looks slightly black on the metal part of the chip.

I'm using an Enermax Liberty 500W so it looks more likely that it might be a problem with the board.

Just how many different DS3's have you been through? Have you tried any other boards? Not sure what to suggest but as i'm in exactly the same situation i'd be very grateful if you could let me know via PM or on the boards if you find a solution.

Cheers ;)
 
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 ;)
 
Right, cheers for the reply. Think i'll pop out and get hold of a PSU tester also. If the PSU proves to be faulty and considering that i've had it for roughly 5 months, i'm assuming i'd send it back to Enermax rather than OcUK?

Does the same go for the motherboard which is definitely completely broken?

:)
 
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.
 
You must log in or register to reply here.
Back
Top Bottom