So my Graphics Card exploded. Literally!

[awppsu]

Was playing some Wargame Red Dragon when I noticed my fans started speeding up! Suddenly smoke and a little burst of flame came from my computer. Before I could shut it down, it shut itself down. Opening the case I immediately spotted the fried part of the card. Looking over the rest of the internals revealed no obvious damage. There was some dust accumulation in the case, nothing major, no tumbleweed balls of dust or the like but it could have used a cleaning.

I proceeded to plug my monitor into the integrated Graphics card and everything ran fine.

My graphics card was a Nividia GTX 560 Ti with a Galaxy Cooling system. The graphics card came with a program called XtremeTuner HD which I used to slightly overclock the card following this guide: http://www.hardocp.com/article/2011/02/01/galaxy_geforce_gtx_560_ti_gc_overclocking_review/1

The computer and card have run fine for about 2 and a half years without any problem.

Before I go out and buy a new card, does anyone here think it could have been something other then the Graphics Card that could have caused this sudden damage and breakdown?

 

[blastingcap]

It's most likely due to a fault of the the graphics card.

However, what is your PSU model number, and how old was it?

And in order to get a better idea of what your power draw was: what is the rest of your system (CPU, mobo, RAM, HDD, SSD, optical--everything)? Were you overvolting your CPU or GPU and by how much? Were you overclocking your CPU and by how much? How much were you overclocking your GPU?

 

[KingFatty]

Well that burned sections is suspiciously close to the voltage pins where the card receives electricity from the power supply.

So it's possible that something silly like a metal staple or other piece fell into the computer, maybe hit a cooling fan and got flung directly on top of the video card, and directly contacted the raw power pin and shorted itself directly to somewhere else on the video card, so then your power supply blasted full juice through there and melted away everything.

Or, the card malfunctioned, as indicated by the increasing fan speed, where you had a failure that caused runaway current to flow which simply overwhelmed the power delivery part. Hmm I can't tell if there was a chip there or something where everything is melted?

 

[Insomniator]

'slightly overclock'

lol right!

 

[Techhog]

'slightly overclock'

lol right!

Great contribution. Simply overclocking on its own wouldn't cause this. It's clearly either a defect or a bad PSU.

 

[lavaheadache]

makes me all teary eyed

 

[skipsneeky2]

'slightly overclock'

lol right!

I have heard these 560 ti cards can be unpredictable. My own experience with a Asus DC2 560 ti failing with a weak overclock well in the safe zone surprised me.

That Asus card technically is the only card I have ever owned that failed on its own accord and this one failed with zero warning, funky colors to a no post all within a hour.

 

[Cloudfire777]

Over 3 year old graphic card that have trouble with many games today.

You got some serious patience there grasshopper.
That said, bad luck I guess. Maybe it was one of them timed feature by forcing people to buy new graphic cards?

*On with the tinfoil hat*

 

[awppsu]

It's most likely due to a fault of the the graphics card.

However, what is your PSU model number, and how old was it?

And in order to get a better idea of what your power draw was: what is the rest of your system (CPU, mobo, RAM, HDD, SSD, optical--everything)? Were you overvolting your CPU or GPU and by how much? Were you overclocking your CPU and by how much? How much were you overclocking your GPU?

I'll get this info asap, once I get home. What I can say is I have a TX650W, I am currently overlocking my CPU and my GPU was overlocked the specifications in the one link I posted. I have 8 GB of RAM, a Intel Core i5-2500k 3.3GHZ and an ASRock Z68 Extreme 4 MBO.

When I initially overclocked it I fully understood that it would shorten the lifespan of the card, I just didn't expect it to go out with such a flash and bang. As to the card having trouble handling modern games. Well... it seemed to me it had zero trouble. Rarely a hiccup or slow down on most games.

Also there is a row of chips in the burned section that you can barely see if you look closely. Also if you flip it over and look between the cooling fans and the card you can see how part of the card is charred right behind the power pin connectors. I'll get more pictures if anyone else wants to see.

Is the fact that I can still run the computer fine, albeit with the monitor plugged into the onboard graphics built into the Motherboard, draw anyone away from the idea it could have been the PSU?

 

[Z15CAM]

I had a loose PCB Bolt that wondered around inside the PSU case that nearly shorted out my MB at one time.

If you give your PSU a shake and heard something rattling that's a sigh of a possible short.

That PIC kinda reminds me of what happened to Old VooDoo's - In spite of that I think nVidia should reincarnate the VooDoo - LOL

Other that said I'm glad you escaped.

 

[96Firebird]

It does indeed look like that is near the PCI-E power connectors. If the PSU surged through one of the PCI-E connectors, that would probably be the first area to go pop.

Is the fact that I can still run the computer fine, albeit with the monitor plugged into the onboard graphics built into the Motherboard, draw anyone away from the idea it could have been the PSU?

You aren't using the PCI-E connectors now, so you can't fully rule out the PSU. However, it is going to be hard to figure out is it was the PSU if it was a one-time surge.

 

[dn7309]

my factory overclock galaxy GTX 660 is for sale if need a new card

 

[Silverforce11]

That region has a few small chips, looks like a small capacitor and resistor and not mosfets/VRM. That's actually quite rare for it to explode like that. Normally its the mosfets that explode due to too much power draw for it to handle (especially with OC).

 

[Soulkeeper]

can you take a picture of the other side of the card ?
I'm kinda curious here

 

[awppsu]

can you take a picture of the other side of the card ?
I'm kinda curious here

Unfortunately I would need to remove the third-party fan system to get a good picture, and I didnt have the proper screw driver to do that. I'll try and get one and get some pictures.

 

[awppsu]

Here are the pics. Sorry for the bad quality. Let me know if anyone can spot what happened or give me a general idea!

 

[KingFatty]

The white box, was that originally white? Or is it just another one of the black boxes like it's neighbor, and it turned white when it burned up? Like Gandalf.

 

[cbrunny]

Love it. Not really but it is interesting.

Similar thing happened to me: http://forums.anandtech.com/showpost.php?p=34757134&postcount=37

http://forums.anandtech.com/showpost.php?p=34757288&postcount=45

 

[Mark R]

Looks like one of the VRM MOSFETs (the things that look like small 8 pin chips near the power sockets) failed short circuit.

These are normally very tough, but they can experience something called "thermal runaway". As they get hotter, they get less efficient, and so produce more heat, causing the temperature to rise, until they either stabilise at a new higher temperature, or something pops. Normally, in a multi-phase VRM, the overheating phase will work less effectively so take less of the load, and the other phases will work harder (so you get a stabilising effect). But if there isn't proper cooling, or one phase is malfunctioning, then it can still overheat to the point that it is damaged and shorts out.

In the event of one of them malfunctioning and getting stuck "on" (short circuit), then massive, uncontrolled amounts of power will start flowing through it, and the MOSFET typically ends up catching fire, and the PCB traces running to it end up charred or melted.

 

[awppsu]

The white box, was that originally white? Or is it just another one of the black boxes like it's neighbor, and it turned white when it burned up? Like Gandalf.

Honestly... I don't know. There is some kind of paint on the box that blew off when I sprayed it with a bit of duster, and it easily scrapped off using a finger nail. Sadly no Balrogs were found.

 

[Soulkeeper]

I think Mark is right on.
Any number of those components could have been the initial cause of this, but it's no surprise that the most confined looking area (hottest, least airflow) looks to be where the phoenix spawned

I've got a video card with the same kind of mosfets, if I overclock it too much it seems to have some kinda current sense mechanism and will just shut off. Maybe yours lacked that or the resistor rating was just a tad off.

I had an asrock motherboard's mosfet melt on me recently.
could literally see some kinda gold colored metal which had melted out of it, and blue discoloration on the leads. This happened exactly at the hottest point in the day when room temps hit 80F, walked away and came back 30min later and it was turned off.

 

[Blue_Max]

This is one more reason I like power efficiency! The more power being passed through, the higher the chances of it blowing out like this... but that's just my humble opinion.

 

[awppsu]

Heres an update for anyone curious. Bought a new GTX 660 with an EVGA set up. Everything is running fine. I have been monitoring voltage, temperature, and wattage, and everything is in the norms. Looks like it was the video card. Thanks for those who lent their opinion!

 

[PhIlLy ChEeSe]

Well that burned sections is suspiciously close to the voltage pins where the card receives electricity from the power supply.

X2
Something shorted out, you don't do that kind of damage with out something Very wrong going on inside the computer. I.E. PSU failure ext..... That is what is called a "DEAD SHORT" when power is feed to ground over a period of time, this is what you get.