I will try to keep this post as unbiased as possible, as many people on this forum have accused me of being extremely biased & I would just like to show that it's not always true. 
This thread will address then debunk the rumor about the 780Ti that supposedly fried it's self & the RiVE motherboard it was in... Please excuse the terrible format, i'm not very good @ these things but some one needs to do it. It's making OCN & it's fellow members look incredibly stupid, so many of us have decided it's time to place the fact's, speculation, & fallacies all in one place.
On november 10th this photo was published on ChipHell... Shortly after that a rash of stories started showing up on tech forums, tech blogs, & journalistic tech pages. This post exists to discredit any of those stories. Please take this knowledge and spread it among the communities you visit daily so that we can get the truth out.

So, here we have the original photo with the affect areas squared off in red & BURNT ORANGE

Here is a photo of what a reference 780Ti PCB looks like. ALL reference 780Ti pcb's will look identical to this one, that is what reference means.

here is a photo of a reference Tahiti PCB., this photo is from a 7970BUT ALL WE KNOW IS THAT THE ORIGINAL PHOTO IS A REFERENCE TAHITI PCB
Provided by Kinaesthetic

Hard Forum thread
LinusTechTips thread
The original Chiphell thread
I have taken the time to create hard & LTT accounts to post the evidence we have gathered. As you can see, the original thread has been 100% removed, this is not to protect nvidia, this is to protect chiphell from a lawsuit. They originally banned & scrubbed the thread clean, but I guess they decided that wasn't enough so they straight up deleted it.
The only argument left is this: "Galaxy must of socketed the wrong chip into the wrong PCB" I am here to tell you that the chances of that are one in 1,000,000,000... I'm not an engineer, so it would be nice if some one with the proper knowledge could chime in. I'm almost positive that even if something like this were to happen it would never leave the factory as it would not pass the in house QA. The chip would likely not even boot.
Let this be a warning to the person who started this rumor. Next time don't be lazy. Next time you should actually blow up the correct GPU, or else we will debunk it in a matter of hours every time.
I want to take this moment to apologize to anyone in Taiwan/China who has not received their GPU due to this terribly sloppy attempt @ a smear campaign
I ALSO WANT TO POINT OUT THAT THIS THREAD IS TO ADDRESS THIS SINGULAR EVENT, I AM NOT SAYING THAT THESE CHIPS MAY BE FLOPPING, I'M SIMPLY SHOWING EVIDENCE THAT THIS SINGULAR EVENT IS A FLUKE
If I have left anything out please feel free to tell me, I will update the OP.
Please feel free to comment & further speculate below...
UPDATE: Night of Nov-11th
An image has surfaced with a SUPPOSED list of defective batch numbers. If anyone knows the origin of this image please let me know!
IF ANYONE CAN ACTUALLY READ THE LANGUAGE ON THIS FORM LET ME KNOW, AS I NOW BELIEVE (with help from galaxy rep) THAT IT'S A BATCH LIST, NOT A RECALL LIST. As in it's the entire product line as of xx/xx/xx date.

another update from a fellow OCN Member, Dbaer who is in fact a circuit engineer, in the business for a very long time
He followed up the statement with
here's what appears to allegedly be another official recall notice from Galaxy. as stated above, & below. It turns out only one batch was recalled, and it was due to a failed stick placement over the mosfet section
this is an alleged first hand account incident from a chinese forum. Translation, scroll down for original post.
Last but not least, straight from the Galaxy rep
So, unless some one here can discern the language in those supposed recall lists, they are looking much more like an entire product batch list more than anything else. As the rep said, ONLY ONE BATCH WAS RECALLED, AND I WAS DUE TO THE STICKER OVER THE MOSFET
Edited by skupples - 11/15/13 at 12:46pm
This thread will address then debunk the rumor about the 780Ti that supposedly fried it's self & the RiVE motherboard it was in... Please excuse the terrible format, i'm not very good @ these things but some one needs to do it. It's making OCN & it's fellow members look incredibly stupid, so many of us have decided it's time to place the fact's, speculation, & fallacies all in one place.
On november 10th this photo was published on ChipHell... Shortly after that a rash of stories started showing up on tech forums, tech blogs, & journalistic tech pages. This post exists to discredit any of those stories. Please take this knowledge and spread it among the communities you visit daily so that we can get the truth out.

So, here we have the original photo with the affect areas squared off in red & BURNT ORANGE
Here is a photo of what a reference 780Ti PCB looks like. ALL reference 780Ti pcb's will look identical to this one, that is what reference means.
here is a photo of a reference Tahiti PCB., this photo is from a 7970BUT ALL WE KNOW IS THAT THE ORIGINAL PHOTO IS A REFERENCE TAHITI PCB
Provided by Kinaesthetic
Hard Forum thread
LinusTechTips thread
The original Chiphell thread
I have taken the time to create hard & LTT accounts to post the evidence we have gathered. As you can see, the original thread has been 100% removed, this is not to protect nvidia, this is to protect chiphell from a lawsuit. They originally banned & scrubbed the thread clean, but I guess they decided that wasn't enough so they straight up deleted it.
The only argument left is this: "Galaxy must of socketed the wrong chip into the wrong PCB" I am here to tell you that the chances of that are one in 1,000,000,000... I'm not an engineer, so it would be nice if some one with the proper knowledge could chime in. I'm almost positive that even if something like this were to happen it would never leave the factory as it would not pass the in house QA. The chip would likely not even boot.
Let this be a warning to the person who started this rumor. Next time don't be lazy. Next time you should actually blow up the correct GPU, or else we will debunk it in a matter of hours every time.
I want to take this moment to apologize to anyone in Taiwan/China who has not received their GPU due to this terribly sloppy attempt @ a smear campaign
I ALSO WANT TO POINT OUT THAT THIS THREAD IS TO ADDRESS THIS SINGULAR EVENT, I AM NOT SAYING THAT THESE CHIPS MAY BE FLOPPING, I'M SIMPLY SHOWING EVIDENCE THAT THIS SINGULAR EVENT IS A FLUKE
If I have left anything out please feel free to tell me, I will update the OP.
Please feel free to comment & further speculate below...
UPDATE: Night of Nov-11th
An image has surfaced with a SUPPOSED list of defective batch numbers. If anyone knows the origin of this image please let me know!
IF ANYONE CAN ACTUALLY READ THE LANGUAGE ON THIS FORM LET ME KNOW, AS I NOW BELIEVE (with help from galaxy rep) THAT IT'S A BATCH LIST, NOT A RECALL LIST. As in it's the entire product line as of xx/xx/xx date.
another update from a fellow OCN Member, Dbaer who is in fact a circuit engineer, in the business for a very long time
Quote:
Originally Posted by DBaer 
Skupples, I just read your comments on the "questionable" PCB flare up and I agree with you. In fact I believe that that burn area was caused by a small torch or some other open flame. You state that you are not an engineer, well I am, I have been in the printed circuit board industry for almost 50 years. I can tell you that in order to do that kind of damage where the Au plated copper traces are gone and the FR-4 laminate is fully carbonized would take much more than a burned out daughter board flare up. Remember FR-4 is fire retardant Etc.
Anyway, for what it is worth, I fully agree with your conclusion.:

Skupples, I just read your comments on the "questionable" PCB flare up and I agree with you. In fact I believe that that burn area was caused by a small torch or some other open flame. You state that you are not an engineer, well I am, I have been in the printed circuit board industry for almost 50 years. I can tell you that in order to do that kind of damage where the Au plated copper traces are gone and the FR-4 laminate is fully carbonized would take much more than a burned out daughter board flare up. Remember FR-4 is fire retardant Etc.
Anyway, for what it is worth, I fully agree with your conclusion.:
He followed up the statement with
Quote:
here's what appears to allegedly be another official recall notice from Galaxy. as stated above, & below. It turns out only one batch was recalled, and it was due to a failed stick placement over the mosfet section
Quote:
Originally Posted by SandGlass 
And here is the official galaxy announcement:
http://szgalaxy.com/2013/GTX780Ti/index.html


And here is the official galaxy announcement:
http://szgalaxy.com/2013/GTX780Ti/index.html
this is an alleged first hand account incident from a chinese forum. Translation, scroll down for original post.
Quote:
Originally Posted by SandGlass 
I read the Chinese forms, apparently:
Translation:
I finally got a 780TI, it was made by galaxy, after putting it in, I tried to turn on the computer, but after a few seconds whirring, the whole machine powered off. I called the seller of the 780Ti, and he said that of the 5 that he sold, 4 were faulty. He also recieved an official phone call from Nvidia, the information was roughly as follows: There is a huge problem with the stock pcb, it does not matter which AIB partner, there will be a recall. The user can choose to return it now, or exchange it when the new PCB is available, this statement will be public on the website next week.
I'm returning this card by mail tomorrow.... Sigh...I have bad luck...help me...
780Ti: Warning: Spoiler! (Click to show)

I read the Chinese forms, apparently:
Translation:
I finally got a 780TI, it was made by galaxy, after putting it in, I tried to turn on the computer, but after a few seconds whirring, the whole machine powered off. I called the seller of the 780Ti, and he said that of the 5 that he sold, 4 were faulty. He also recieved an official phone call from Nvidia, the information was roughly as follows: There is a huge problem with the stock pcb, it does not matter which AIB partner, there will be a recall. The user can choose to return it now, or exchange it when the new PCB is available, this statement will be public on the website next week.
I'm returning this card by mail tomorrow.... Sigh...I have bad luck...help me...
780Ti: Warning: Spoiler! (Click to show)
Last but not least, straight from the Galaxy rep
Quote:
Originally Posted by Galaxy 
Actually yes I do. With all the speculation that's been going on I think you'll be surprised to hear how simple the problem actually was. This batch of cards failed because the serial number sticker was somehow placed over the back of a MOSFET. Seriously. No faulty parts, no bad GPUs or anything like that. You'd be surprised how many problems a sticker being slightly out of place can cause. I'm sure it goes without saying the guy who calibrates the sticker sticking machine has gotten more attention than he's used to recently. Either way there's no risk of this happening to any cards in North America, Europe, Australia, or anywhere else. Just the one batch in the Chinese market was affected.

Actually yes I do. With all the speculation that's been going on I think you'll be surprised to hear how simple the problem actually was. This batch of cards failed because the serial number sticker was somehow placed over the back of a MOSFET. Seriously. No faulty parts, no bad GPUs or anything like that. You'd be surprised how many problems a sticker being slightly out of place can cause. I'm sure it goes without saying the guy who calibrates the sticker sticking machine has gotten more attention than he's used to recently. Either way there's no risk of this happening to any cards in North America, Europe, Australia, or anywhere else. Just the one batch in the Chinese market was affected.
So, unless some one here can discern the language in those supposed recall lists, they are looking much more like an entire product batch list more than anything else. As the rep said, ONLY ONE BATCH WAS RECALLED, AND I WAS DUE TO THE STICKER OVER THE MOSFET
Edited by skupples - 11/15/13 at 12:46pm










