We are probably all familiar with the Rossmann video on reballing, but I'll link it below just in case.
I have yet to see or hear of anyone actually showing any research, studies, or other types of investigation on the subject. It really seems just like everything else- a strongly worded statement of one's own personal opinion, with nothing to back it up other than a firm conviction. There are plenty of alternative possibilities to what the problems may be, and what heating or reballing the die may actually be doing. Despite the similarities in his logic to those he is disagreeing with, people tend to rally behind Rossmann's opinion on this video. He has not properly disproved the general idea that reballing BGA chips "fixes" them, nor has he supported his own claims about the "little bumps" inside the chip being the actual root cause. His claim that heating the chip well below the point at which solder melts somehow "proves" his point is simply not true.
I am a fan of Rossmann and the lion's share of what he has done on the topic of repair, freedom, and related subjects. His claims may very well be true, but he has not supported them.
What are your thoughts on the topic? Does anyone know of any studies or research that may be more valid-- or any other convincing arguments in general? Any ideas on ways to have a better idea on what is going on, even if it is not quite to a proper scientific level?
He is not nearly the first who said this simple truth. This reball repair BS is going on since around 2007
. Starting from laptops with failing Nvidia GPU or chipset.
He proven his claim easily, IIRC he said something like "heat up GPU way below solder melting point, and voila it works again". Which would not be the case if there was a soldering issue between PCB and chip.
Thanks for the image. What is the source? Is any pdf possible?
On a flip chip BGA it's the balls that connect the die (i.e. the bit generating the heat) to the substrate which suffer from failure due to differential thermal expansion induced stress. The thermal expansion coefficient of the substrate and PCB are more closely matched so stress is lower, and the balls are larger.
I googled with some of that image text and found the
journal source. anyone have an arxiv pdf or another source? :p
Using that I also found some more info, like the image below. I'm going to do some more reading down this rabbit hole.
I did a bit of googling into this at the time, and it appears he is right. It's an inherent flaw in flip chip design of a certain time, and what was breaking was not the solder balls under the chip, but the solder balls between the silicon and the package substrate. No amount of reballing would fix that.
http://www.sonoscanchina.cn/us/images/18FCBumpPassivationFailIRPS97.pdf
I dont think you need any more evidence than practical evidence.
AMD chips repair in 99% cases if done correctly and the chip is still working and stay working for a lenghty period of time (we are yet to have one return after years of servicing) , where NVidia chips fail within a few days or weeks after the exact same procedure. The most trouble originates from users abusing their failing hardware for days or months untill it completely fails before trying to service it.
Some times on nVidia chips due to bad cooling and constant usage at those higher temperatures a typical reballing fixes the problem as it was caused by bad bottom connections and not those within the chip.
I dont think you need any more evidence than practical evidence.
AMD chips repair in 99% cases if done correctly and the chip is still working and stay working for a lenghty period of time (we are yet to have one return after years of servicing) , where NVidia chips fail within a few days or weeks after the exact same procedure. The most trouble originates from users abusing their failing hardware for days or months untill it completely fails before trying to service it.
You mean success that it lasts long enough that service customers are left without a repair warranty? If something fails within a few months people likely will just choose another repair shop, they don't have a warranty anyway. Some AMD chips from early 2010s were not that much better than Nvidia from 2006-2008.
Some times on nVidia chips due to bad cooling and constant usage at those higher temperatures a typical reballing fixes the problem as it was caused by bad bottom connections and not those within the chip.
In rare lucky+low stress cases might last like 6-12 months instead of 2 weeks. Issue still within the chip. Actual soldering fail is quiet rare and even with unaffected by plague flip chip BGA you never know for sure if the issue wasn't in the chip.
Join just about any repairman's crash course,here in India,
the first thing they are gonna teach you is reballing,but before that,they'll tell you that many chips will go bad some time afterwards, but it'll be late enough that the customer won't suspect that you didn't repair it fully and wont blame you.
If a quick&easy reflow fixes it, and it does, who cares? How is it better to replace the chip with another new but also defective chip? I've done a zillion reflows to iMacs and PowerBooks and it just works(TM).
https://www.eevblog.com/forum/beginners/getting-started-with-bga-reflow-and-reballing/msg472940/#msg472940
Got to be nice to be someone who is ripping off his customers. If you know that you cannot buy a good replacement, you need to have guts to refuse a repair, not milking customer.
Right, I've ripped off my friends, and they're still very thankful. They just had to choose whether to pay many$ for a mlb swap or a a let's try this and cross fingers. And they chose wisely because an mlb swap would more than likely have failed again and this hasn't, as has happened to too many previous gen iBook G4 users. No doubt there's morons who may prefer to put it in the trash and get a new one... just because it might fail again.
If a quick&easy reflow fixes it, and it does, who cares? How is it better to replace the chip with another new but also defective chip? I've done a zillion reflows to iMacs and PowerBooks and it just works(TM).
https://www.eevblog.com/forum/beginners/getting-started-with-bga-reflow-and-reballing/msg472940/#msg472940
Got to be nice to be someone who is ripping off his customers. If you know that you cannot buy a good replacement, you need to have guts to refuse a repair, not milking customer.
Right, I've ripped off my friends, and they're still very thankful. They just had to choose whether to pay many$ for a mlb swap or a a let's try this and cross fingers. And they chose wisely because an mlb swap would more than likely have failed again and this hasn't, as has happened to too many previous gen iBook G4 users. No doubt there's morons who may prefer to put it in the trash and get a new one... just because it might fail again.
Something does not add up. You certainly talk BS in one way or another.
Zillion refrows, for friends? I could believe that if you do a reflow every 2 weeks for the same few laptops.
Join just about any repairman's crash course,here in India,
the first thing they are gonna teach you is reballing,but before that,they'll tell you that many chips will go bad some time afterwards, but it'll be late enough that the customer won't suspect that you didn't repair it fully and wont blame you.
You're not to blame if you explain it's a design error and you can only do what you can do but the error is going to remain there and may strike again. Twelve years ago an Apple mlb swap was very, very expensive, and didn't warrant a better outcome than a reflow, it paid to gamble and try the reflow first.
I dont think you need any more evidence than practical evidence.
AMD chips repair in 99% cases if done correctly and the chip is still working and stay working for a lenghty period of time (we are yet to have one return after years of servicing) , where NVidia chips fail within a few days or weeks after the exact same procedure. The most trouble originates from users abusing their failing hardware for days or months untill it completely fails before trying to service it.
You mean success that it lasts long enough that service customers are left without a repair warranty? If something fails within a few months people likely will just choose another repair shop, they don't have a warranty anyway. Some AMD chips from early 2010s were not that much better than Nvidia from 2006-2008.
Some times on nVidia chips due to bad cooling and constant usage at those higher temperatures a typical reballing fixes the problem as it was caused by bad bottom connections and not those within the chip.
In rare lucky+low stress cases might last like 6-12 months instead of 2 weeks. Issue still within the chip. Actual soldering fail is quiet rare and even with unaffected by plague flip chip BGA you never know for sure if the issue wasn't in the chip.
Each and every customer is fully aware of what an reballing procedure and an chip replacement procedure is. I can warrant that my work was done correctly with the tools I have in an unlimited time period but I can not warrant for your crappy chip as it is not new nor has factory warranty any more. Warranty can only be given on new chips that have their own warranty. In personal experience no one has ever opted for chip replacement, it is too expensive. Keep in mind that in the region that I am from there are only a few % of those who have high end machines.
Zillions in my book is a synonym for plenty or dozens :-)
https://www.eevblog.com/forum/beginners/getting-started-with-bga-reflow-and-reballing/msg472940/#msg472940
I'm not an expert in BGA reflow, but I have repaired macbook pros, macbook airs, white iMacs and aluminium iMacs, all with these nvidia chips that develop cracked solder joints, a few dozen Macs in total, with this simple "recipe" I was given:
Zillion is certainly is more than plenty of dozens. But I still won't ever believe you have so much "friends" that they could bring in plenty of dozens of iMacs and PowerBooks. And at the same time you don't do this professionally for living, yet pay off for a right equipment. Not to say that old PowerBooks are completely irrelevant to the problem.
I always had a sneaking suspicion that graphics card failures in iMacs that come good after heating the card might have been due to tin whiskers from lead free solder causing shorts that were then melted during the heat process as they are so fine. Just a theory, I have no proof though.
Each and every customer is fully aware of what an reballing procedure and an chip replacement procedure is.
Then I know why you don't have returns. Because they sell them to another suckers before they die again.
I always had a sneaking suspicion that graphics card failures in iMacs that come good after heating the card might have been due to tin whiskers from lead free solder causing shorts that were then melted during the heat process as they are so fine. Just a theory, I have no proof though.
Tin whiskers, particularly on GPU. When there are tons of other components with finer pin pitch.
Each and every customer is fully aware of what an reballing procedure and an chip replacement procedure is.
Then I know why you don't have returns. Because they sell them to another suckers before they die again.
Do they or do they not I do not know nor care. But I do know that those who purchased an repaired laptop from my self , as I sell repaired laptops as well , those who bought reballed laptops returned for system maintenance and purchase of additional devices later on. The shop is open and on the same address for years as it has always been, it is their own flaw not to return and get problems solved.
Zillion is certainly is more than plenty of dozens. But I still won't ever believe you have so much "friends" that they could bring in plenty of dozens of iMacs and PowerBooks. And at the same time you don't do this professionally for living, yet pay off for a right equipment. Not to say that old PowerBooks are completely irrelevant to the problem.
Dear, almost thirty years being an Apple employee, since long before the Macs even existed, I've known hundreds of Apple users with many of which became friends (mostly nerds). Retired in 2008. Just so you know iBook G4s had the same problem and aren't powerbooks. What's the purpose of you trying to break my balls at every opportunity? It's tiring...
I always had a sneaking suspicion that graphics card failures in iMacs that come good after heating the card might have been due to tin whiskers from lead free solder causing shorts that were then melted during the heat process as they are so fine. Just a theory, I have no proof though.
Yes, same here, I would have liked to be able to see under the chip!
Also, dear wraper, BTW, how do you think Apple itself fixes these mlbs? They do exactly the same thing...