Inquirer claims MacBook Pros may use problematic NVIDIA chips
The Inquirer, which has been sticking it to NVIDIA for some time, is again taking the chip maker to task this week, claiming that it's supplying Apple with graphics chips for the new MacBook Pro that use materials that have caused problems in the past.
Following its own investigation into the matter, the British technology tabloid concluded that the dedicated NVIDIA 9600M GT graphics chips in the unibody MacBook Pros use the same non-eutectic solder contact bumps as the GeForce 8400M and 8600M family of chips .
Those chips were proven prone to long-term heat damage, which led to notebook recalls from Apple, HP and Dell, and caused NVIDIA to take a $200 million charge back in July as a result. In these cases there were instances where the bumps, or tiny balls of solder, that hold the chip to its circuit board would crack under thermal stress and cause the hardware to become defective.
According to NVIDIA documentation, non-eutectic solder contacts (or "bad bumps") are comprised of mostly lead (95%) and a bit of tin (5%), which can lead to clumping when the material cools. By contrast, eutectic solder contacts (or "good bumps") are made of roughly 63% tin and 37% lead, a composition which cools more uniformly and produces a more consistent grain that's not prone to the same long-term heat damage.
Since the bumps sit permanently sandwiched between the chip die and their green fibreglass package, the only way to determine their composition would be to take a MacBook Pro, disassemble it, desolder the chips, saw them in half, encase them in lucite, and run them through a scanning electron microscope equipped with an X-ray microanalysis system.
The Inquirer claims to have done just this with a MacBook Pro it bought off the shelf in California shortly after the systems were announced in mid-October. It was reportedly aided by a team of unnamed scientists who have access to the multi-million dollar tools required to properly examine the chips.
A profile of the materials found in the bumps used on the MacBook Pro's NVIDIA 9600M GT show a huge spike of lead and only a tiny spike representing tin, leading the Inquirer to conclude that the chip "is unquestionably using bad bumps." The same test run on the MacBook Pro's second, integrated NVIDIA 9400M graphics chip turned up a profile consistent with eutectic solder contacts, and therefore that chip is said to be free of the issues that may plague the 9600M GT.
A profile of the bumps on the MacBook Pro's integrated 9400M (left) compared to that of its discrete 9600M GT counterpart (right).
The Inquirer suggests that the bad bumps used on the 9600M GT are causing a problem on the new MacBook Pros often referred to as the "black screen of death." The issue, which manifests itself when the systems heat up during game play, sometimes causes the notebooks' screens to go black after just a few minutes of gaming, while the systems lock up and the audio enters into an infinite loop. Apple is said to be investigating the problem.
For its part, NVIDIA has vehemently denied the Inquirer's assessment, claiming that the "GeForce 9600 GPU in the MacBook Pro does not have bad bumps. The material set (combination of underfill and bump) that is being used is similar to the material set that has been shipped in 100s of millions of chipsets by the world's largest semiconductor company."
A representative for the chip maker had not responded to AppleInsider's individual request for comment as of press time.
Following its own investigation into the matter, the British technology tabloid concluded that the dedicated NVIDIA 9600M GT graphics chips in the unibody MacBook Pros use the same non-eutectic solder contact bumps as the GeForce 8400M and 8600M family of chips .
Those chips were proven prone to long-term heat damage, which led to notebook recalls from Apple, HP and Dell, and caused NVIDIA to take a $200 million charge back in July as a result. In these cases there were instances where the bumps, or tiny balls of solder, that hold the chip to its circuit board would crack under thermal stress and cause the hardware to become defective.
According to NVIDIA documentation, non-eutectic solder contacts (or "bad bumps") are comprised of mostly lead (95%) and a bit of tin (5%), which can lead to clumping when the material cools. By contrast, eutectic solder contacts (or "good bumps") are made of roughly 63% tin and 37% lead, a composition which cools more uniformly and produces a more consistent grain that's not prone to the same long-term heat damage.
Since the bumps sit permanently sandwiched between the chip die and their green fibreglass package, the only way to determine their composition would be to take a MacBook Pro, disassemble it, desolder the chips, saw them in half, encase them in lucite, and run them through a scanning electron microscope equipped with an X-ray microanalysis system.
The Inquirer claims to have done just this with a MacBook Pro it bought off the shelf in California shortly after the systems were announced in mid-October. It was reportedly aided by a team of unnamed scientists who have access to the multi-million dollar tools required to properly examine the chips.
A profile of the materials found in the bumps used on the MacBook Pro's NVIDIA 9600M GT show a huge spike of lead and only a tiny spike representing tin, leading the Inquirer to conclude that the chip "is unquestionably using bad bumps." The same test run on the MacBook Pro's second, integrated NVIDIA 9400M graphics chip turned up a profile consistent with eutectic solder contacts, and therefore that chip is said to be free of the issues that may plague the 9600M GT.
A profile of the bumps on the MacBook Pro's integrated 9400M (left) compared to that of its discrete 9600M GT counterpart (right).
The Inquirer suggests that the bad bumps used on the 9600M GT are causing a problem on the new MacBook Pros often referred to as the "black screen of death." The issue, which manifests itself when the systems heat up during game play, sometimes causes the notebooks' screens to go black after just a few minutes of gaming, while the systems lock up and the audio enters into an infinite loop. Apple is said to be investigating the problem.
For its part, NVIDIA has vehemently denied the Inquirer's assessment, claiming that the "GeForce 9600 GPU in the MacBook Pro does not have bad bumps. The material set (combination of underfill and bump) that is being used is similar to the material set that has been shipped in 100s of millions of chipsets by the world's largest semiconductor company."
A representative for the chip maker had not responded to AppleInsider's individual request for comment as of press time.
Comments
Seriously though, I would think Apple did their homework on the 9600 before using it... "Fool me once," etc. Either way, one more reason to go with the smaller Macbook unless you absolutely need a "Pro" (which isn't really a Pro product anymore anyway).
Do we really need to cite the Enquirer for tech stories? Is there no minimum standards @ AI?
Reading the story, the journalists do seem to have done their homework this time out. I don't have a problem with the source.
...
My early '08 MB Pro (that I just got about four weeks ago on Amazon) has an nVidia 8600M GT graphics chip in it, which would seem to be one of the defective ones. Though my early '08 MBP *is* from the end of the production run, though. You'd like to think the prob got fixed by then.
Hmm... even so, guess I should consider getting the Apple Care if I still can, huh?
I think $350 is an awful lot of cheddar, tho'. 25% of the cost of the machine.
...
Reading the story, the journalists do seem to have done their homework this time out. I don't have a problem with the source.
...
Yeah. We are careful in selecting stories from certain publications, but the Inquirer has followed the NVIDIA story pretty well and their evidence in this case is clearly sound. NVIDIA has also not been very forthcoming on subjects similar to this, and I can attest to that first hand. They have every opportunity to tell their side of the story. In addition to contacting them this morning, they know us well and how to get in touch with us. If they issue comment, I will have the story updated. I have a feeling they will not, however, citing the privacy of their client (Apple). This was the response I was given while researching the problems with the MacBook Pro's 8600 chip, many weeks before Apple ultimately confirmed the problem independent of NVIDIA.
Best,
K
Apple may well have known that the 9600M GT still uses the old bump material, but may have concluded that it was an acceptable risk. The 9600M GT is produced on a 65nm process instead of 80nm like the 8600M GT and it's heat consumption should be lower reducing the risk of fracture. There may also have been some tweaks made between generations to mitigate the problems even if the same material is used. I believe the new Unibodies also generally run cooler than the old MBP too. Combined, the risk may still be there, but may be close enough to tolerances that Apple and nVidia felt no additional measures were needed.
Or maybe nVidia tried to pull a fast one on Apple (and others), because nVid didn't want to eat the (pretty tremendous) cost of fixing the issue.
That wouldn't surprise me at all, particularly if the costs of fixing it would jeopardize nVidia's continued existence, as opposed to merely its' bottom line. That, however, is still not really a good excuse. If you effed up, you should admit it as soon as you're aware of it, and fix it.
No doubt nVidia's lack of candor on the issue is going to really hurt them going forward. Who's going to trust them anymore?
...
If they issue comment, I will have the story updated.
Request: When a story is updated can a forum post be made indicating that it has been updated or at least have my your RSS present the story as new again?
So much for being Green.
Request: When a story is updated can a forum post be made indicating that it has been updated or at least have my your RSS present the story as new again?
Solipsim, we usually append the title with "[u]" which means "[updated]." That should reflect in the RSS feed.
Best,
K
Or maybe nVidia tried to pull a fast one on Apple (and others), because nVid didn't want to eat the (pretty tremendous) cost of fixing the issue.
That wouldn't surprise me at all, particularly if the costs of fixing it would jeopardize nVidia's continued existence, as opposed to merely its' bottom line. That, however, is still not really a good excuse. If you effed up, you should admit it as soon as you're aware of it, and fix it.
No doubt nVidia's lack of candor on the issue is going to really hurt them going forward. Who's going to trust them anymore?
...
While it may be in nVidia's interest to keep it quiet from the public, I can't see them withholding that information from Apple. I thought Steve Jobs publicly said they are supporting nVidia's platform and are going to use it in future models. It's almost a certainty that Jobs extracted guarantees from nVidia that this won't blow up in his face, before he gave his support, especially given the prior history of the 8600M GT. Just like Jobs supposedly told Motorola during G4 negotiations that he couldn't wait for Apple to never have to work with them again, nVidia would probably perceive Apple pulling away from them and going completely ATI as a real threat, and given Apple's media influence, it'd certainly be a PR catastrophe. If bump material is still an issue, I can't really see nVidia keeping it from Apple and risking Steve Job's wrath.
Or maybe nVidia tried to pull a fast one on Apple (and others), because nVid didn't want to eat the (pretty tremendous) cost of fixing the issue.
That wouldn't surprise me at all, particularly if the costs of fixing it would jeopardize nVidia's continued existence, as opposed to merely its' bottom line. That, however, is still not really a good excuse. If you effed up, you should admit it as soon as you're aware of it, and fix it.
No doubt nVidia's lack of candor on the issue is going to really hurt them going forward. Who's going to trust them anymore?
...
That's what bothers me, why would they manufacture all these chips while knowing that they have "bad bumps" and running the risk of losing another few million? Or did they have to rush it out so Apple could announce it on time?
Either way, I'm really disappointed by the way Apple didn't notice this and how nVidea simply keeps lying about it...
(btw; the news tip was no biggie
While it may be in nVidia's interest to keep it quiet from the public, I can't see them withholding that information from Apple. I thought Steve Jobs publicly said they are supporting nVidia's platform and are going to use it in future models. It's almost a certainty that Jobs extracted guarantees from nVidia that this won't blow up in his face, before he gave his support, especially given the prior history of the 8600M GT. Just like Jobs supposedly told Motorola during G4 negotiations that he couldn't wait for Apple to never have to work with them again, nVidia would probably perceive Apple pulling away from them and going completely ATI as a real threat, and given Apple's media influence, it'd certainly be a PR catastrophe. If bump material is still an issue, I can't really see nVidia keeping it from Apple and risking Steve Job's wrath.
It's possible that nVidia wasn't aware of just how bad the problem was at first.
Or perhaps they were worried that admitting it and disclosing fully would kill the company, and thought they could somehow skate by, hoping that the problem might end up being not so bad. Companies make strategic calculations like that all the time. Sometimes it even pays off for them. I'm thinking of Ford Explorers and SUV rollover deaths, for example. That wasn't entirely caused by defective tires, but rather, also by Ford's design of the vehicle. But even so, Ford stonewalled, and pretty much got off scot-free.
I dunno, there are a lot of "what ifs" here, but if I were Steve, I'd be taking a long hard look at ATi graphics chips right now. If nothing else, nVidia does not seem to be able to get it together right now.
...
Given that Apple have stated that the systems are RoHS compliant I highly doubt the accuracy of The Inquirer's report.
Kasper, you (and The Enquirer) have missed a much more serious consequence of this story if it's true. Lead solder is banned in Europe as part of RoHS, so if the report is accurate then Apple would be breaking the law in several European countries (any country that's a member of the EU) by selling this system there.
Given that Apple have stated that the systems are RoHS compliant I highly doubt the accuracy of The Enquirer's report.
That would be a good question for Kasper to put to the Inquirer. He did say that he and the Inquirer could get ahold of each other.
Edit- But wait a sec... RE-READ the article:
The Inquirer claims to have done just this with a MacBook Pro it bought off the shelf in California shortly after the systems were announced in mid-October.
Maybe that clears it up? I assume lead solders are still legal in the US? Though I don't know that one off the top of my head.
...
Kasper, you (and The Enquirer) have missed a much more serious consequence of this story if it's true. Lead solder is banned in Europe as part of RoHS, so if the report is accurate then Apple would be breaking the law in several European countries (any country that's a member of the EU) by selling this system there.
Given that Apple have stated that the systems are RoHS compliant I highly doubt the accuracy of The Enquirer's report.
From the official website:
7.
? lead in high melting temperature type solders (i.e. lead based alloys containing 85% by weight or more lead)
So yeah...
Kasper, you (and The Enquirer) have missed a much more serious consequence of this story if it's true. Lead solder is banned in Europe as part of RoHS, so if the report is accurate then Apple would be breaking the law in several European countries (any country that's a member of the EU) by selling this system there.
Given that Apple have stated that the systems are RoHS compliant I highly doubt the accuracy of The Enquirer's report.
There are exceptions where there is not a suitable alternative and I would say this is why they can still use lead.
...
From the official website:
So yeah...
Oops. My bad. Looks like lead solder inside a chip that bonds the die to the package is exempted. Carry on...
What kind of material is "led"?
I believe it is used in the construction of zeppelins.