NVIDIA GPU Failures Caused by Material Problem, Sources Claim

[D_o_S]

NVIDIA has recently been experiencing higher-than-normal failure rates with some of its GPUs. Charlie Demerjian from The Inq. speculates that NVIDIA could actually have a much larger problem than what it admits to - it is quite possible, according to Demerjian, that not only mobile GPUs are affected, but the desktop G92 and G94 could also carry the fault.

The failures, some sources say, are caused by a solder bump that connects the I/O termination of the silicon chip to the pad on the substrate. In Nvidia's GPUs, this solder bump is created using high-lead. A thermal mismatch between the chip and the substrate has substantially grown in recent chip generations, apparently leading to fatigue cracking. What supports the theory that a high-lead solder bump in fact is at fault is the fact that Nvidia ordered an immediate switch to use eutectic solders instead of high-lead versions in the last week of July. Eutectic solders are believed to solve the problem of fatigue cracking.

View at TechPowerUp Main Site

 

[jbunch07]

Quite interesting. Its good their fixing it.

 

[Darkrealms]

Guess I should check and see when my card was manufactured, lol.

This is definately bad timing for Nvidia now that ATI has a nice set of cards with their 48xx series.
Glad to see they did something about it when they found out. Who made the original design Nvidia or the Fabs?

 

[timta2]

I am now glad that the card I bought has a lifetime warranty. I know that usually the warranty on video cards is trivial but considering this news it might become more important to people when choosing which card if you are in the market for one of these.

 

[choppy]

Charlie Demerjian from The Inq. speculates that NVIDIA could actually have a much larger problem than what it admits to

thats all i had to read, dont jump the gun just yet

 

[INSTG8R]

I recently had my 8400M G 256M replaced in my Acer 7520. I HOPE it was replaced with a newer part. I dont wanna have to send it back in AGAIN...

 

[WarEagleAU]

Nice switch, but Im hoping it doesnt affect Nvidia desktop chips. All the nvidia owners on here might want to find out though. Id hate for it to happen to any of yall.

 

[tkpenalty]

The INQ usually is biased towards AMD. This is most likely FUD as the faliure rate is extremely low, not enough to be a huge issue.

 

[Hayder_Master]

what a hill , they must be careful when they made , this is graphic card's not a mouse pad

 

[candle_86]

its inquirer, if you take it seriously your stupid, also i thought posting from fud sites in the news was a no no?

 

[OnBoard]

I am now glad that the card I bought has a lifetime warranty. I know that usually the warranty on video cards is trivial but considering this news it might become more important to people when choosing which card if you are in the market for one of these.

You do know that lifetime warranty is not your lifetime, but the cards manufacturing lifetime. When it goes EOL, lifetime warranty stops as well.

http://www.fudzilla.com/index.php?option=com_content&task=view&id=6753&Itemid=1

Might not be that far for 8800GTS 512MB. 8800GTS 320MB was released February 2007 and killed off (EOL) in October 2007. That comes just to 9 months lifetime and actual warranty was longer

 

[Darkrealms]

Nice switch, but Im hoping it doesnt affect Nvidia desktop chips. All the nvidia owners on here might want to find out though. Id hate for it to happen to any of yall.

I've got a good warranty on my card, maybe they'll let me upgrade if it happens : )
(optimisms great, huh)