Q: 2011 MacBook Pro and Discrete Graphics Card

If cracked solder is the reason we won't need reballing at all, reflow should be enough because it will fix the cracked solder but that doesn't happen, reflow is always a temporary fix and that because it doesn't eliminate the root of the problem which is the dirt under the solder of the chip that causes the bad connection most of the time and reballing is the only way to remove that since the chip and the old solder will be removed , the soldering pads will be cleaned then the chip will be reballed and soldered again to clean pads and that is why it is a long term fix.

I faced the same issue last week. Right now I am running my system be removing AMD Radeon driver from extension folder. MBP, early 2012, i7 2.2GHz, Radeon 6750m.

Hi All

Just had my June 2011 Macbook Pro repaired by Apple again.

Oct 2013 (28 months old) the graphics issue reared its head and after talking with Apple, they fixed it under warranty using our Australian Statuary warranty as I don't have applecare.

The logic board failed again in Feb this year and just had it replaced as Apple has 12 month warranty on repairs.

So if it fails between now and Sep 2015 they will replace it again.  This may go on for years with the notebook away for a few days at a time having replacement logic boards

I spoke to the case office and they will tackle what can be done if and when it becomes an issue again.

Dave

My opinion... Dust is not the cause. How can dust penetrate metal? It cannot.

If your analogy is correct then other BGA components would also suffer the same fate? CPU, VRAM, I/O Chipset....etc

This is due to heat cycles.... This causes fracturing... The reason the GPU is affected as it has the most heat cycles and differences in temp. Especially with a switching setup where it is switching from Intel to AMD all day long. The CPU does not have this issue as it is generally 50C-70C all day long.

Lead solder is less prone to fatigue during heat cycles. This is why it is still used in Aviation. Going from minus temps to very hot temps in minutes.

bga_repairs, ps3specialist,

you both, and others that know what they talking about, have different opinions/views on the cause and the cure but in the end, the cause should be Apple's problem and not the customers.

since you both can repair these failed computers, this is all that matters while Apple doesn't offer anything else than paying too much money for just swapping the entire logic board for another one with the same defect.

so please, you won't ever agree on leaded vs lead free soldering debate anyway.

what matters, again, is that you can do your job, that it's cheaper and more solid than Apple's and most of all, that mbp 2011 owners can have a working computer back.

bga_repairs wrote:

 

My opinion... Dust is not the cause. How can dust penetrate metal? It cannot.

 

If your analogy is correct then other BGA components would also suffer the same fate? CPU, VRAM, I/O Chipset....etc

 

This is due to heat cycles.... This causes fracturing... The reason the GPU is affected as it has the most heat cycles and differences in temp. Especially with a switching setup where it is switching from Intel to AMD all day long. The CPU does not have this issue as it is generally 50C-70C all day long.

 

Lead solder is less prone to fatigue during heat cycles. This is why it is still used in Aviation. Going from minus temps to very hot temps in minutes.

I agree dust isn't it - though the grey dust under the balls is most likely the fine particulates of the excess thermal paste which after a lot of cycles turns into a fine grey powder.  It's the bad pasting and the poor uneven condition of the contact surfaces which go on the GPU and CPU that's the problem which causes big heat build up with nowhere to go as the bond simply isn't good enough for these big dies of the 2011 models.  With the 2 plates properly lapped to a nice shine and decent paste applied the CPU will idle in the low 40c range for a 15 and even lower for the 17, and the GPU around 50-55c.  With OS X discrete switching you have a very hot Sandy Bridge CPU which is overpowering the heat-pipe then the AMD coming on means it can't draw away that extra heat fast enough.  That heat has to go somewhere and it's not the heat-pipe.

Lead is also used in the power generation industry and mission critical telecoms kit for the same reasons - and still not one of the 9 I have referred to you has gone faulty so far even still.  That's my preference so they don't go wrong again and it works like no doubt it does for you using Pb solder.  But even a lead free with a decent pro like ps3specialist will be a better job than Apple's depot work.

ps3specialist wrote:

 

These logic boards use leadfree solder in every onboard component so if solder is the only reason for the GPU failure then every other chip on board should fail , they all go through the same number of heating and cooling cycles, make sense ? that gives us the conclusion that it is not the solder that causes the problem and if it is making some of the factors it is not more than 5% of the factors that affecting these logic boards, heating cycles are also not an important factor and that because no matter how hot these logic boards go they can not go over 70 to 90 degree Celsius - before the computer protection cycle will activate and shuts the computer off - which is a very low temperature for the design of these logic boards and their materials , compare that to 217 degree Celsius which is the melting point of the leadfree solder that should actually lead to the conclusion that leadfree solder will have much higher stability comparing to the leaded solder which has a lower melting point  of 183 degree Celsius even if we disregard its lead material that is considered safety and environmental hazard material, so sorry to tell you that your conclusion and advise of using leaded solder is scientifically and factually wrong. You are entitled to your opinion like myself so we will just leave them to for people to decide for themselves which facts are true and they can google the numbers too.

I agree to an extent - but I put the primary causes of these GPU's failure are the poor surfaces of the die plates on the palm tree heat-pipe and Apple's pasting guidelines which aren't up to the job for the pair of chips in this chassis.  Other factors such as a lot of poor lead free solder swimming in the market after the 5 year exemption of the 2006 ROHS (The Restriction of Hazardous Substances Directive) expired in 2011 for using lead solder another possible cause.  My opinion on preferring Pb is that 217C in your case with lead free is still far more than than the 183-5C using lead when reballing, and for the sake of the logic board surface and rapid heat cycles I prefer using it as they go wrong less.

Whenever I do my re-pasting in essence I am re profiling the heat-pipe assembly plates to make them much, much more efficient. Shiny means you've got gaps on that plate of tens of microns or less in terms of being even, the dull finish the stock plate has normally the gaps, pits and troughs on the profile of the plate are probably 5-10 times more without work. As thermal paste is meant to bridge those gaps between the die and heat-pipe it's simple thermodynamics as to why after lapping they improve an awful lot.

I cannot see why Apple engineering don't find out for themselves by not only sending the depot boards away to be reballed they send a few hundred of the heat-pipe assembly away to have the plates lapped and left with tape on top to be removed shortly before fitting with a lot less paste.  I know from my experience there is not a single MacBook Pro CPU and GPU in that chassis for from 08-12 period that doesn't improve its cooling a lot having it done along with better paste.  I knew that from the very start in 2008 when you could almost cook an egg on the left speaker of the first gen 15" unibody!

Add our early 2011 mbp to the list... again.  original failed within a about a year.  replacement board #1 presented the same symptoms within a week or two.  replacement board #2 failed after 7 months of light use.  discussions with their customer relations folks have been cordial and helpful with the last repair being covered by apple although i was long out of warranty on the first repair.  we don't plan to reiterate the fix-fail cycle with apple.  maybe look at independent repair shops.  maybe pitch the thing out and begin slipping back into the Linux and Windows worlds.

well crap... My 2011 15" MBP started doing this 2 days ago...  Has anyone seen any change in Apple's stance or anything about the petition?  I haven't been to the local store yet, but I'm pretty sure I know what they're gonna tell me and the price attached to it...

Add one more frustrated once-happy Apple user to the list of disappointed affected many....

It all started months ago with a little overheating that eventually required I disconnect my external monitor...

Now it's progressed to the point where the display freezes and eventually goes black, the top center and left get very hot and I have to reboot to get it to work again... but usually only a few hours of use will result in the same situation again and again. It's been this way for the last 2-3 days. I can no longer plug in my external monitor or it will crash immediately. I can't watch video or it will soon crash. Anything graphic intensive causes the machine to melt down. I can only use for very basic Internet access without video.

I have late 2011, 2.4 GHz Intel Core i7, 8GB DDR3, 17-inch Intel HD 3000 512 MB graphics

Started having issues in early 2011 17 inch MBP in august and somehow it was working fine until now. But the problem again came today.

Can anyone suggest whether reballing the gpu will fix the issue or logic board replacement is required,

I know you will always disagree and that is fine if you disagree with an opinion but it is not when you disagree with a fact , yes every BGA chip on the logic board will have dust around it and they all fail and that is why in some logic boards they get CPU failure but it happens more often with the GPU for a reason , the next two pictures might be enough to show why ..

you can clearly see that the dust concentrates more around the GPU chip due to its location on the board at the rear edge of the computer close to the ventilation area and the air circulation inside the computer  , the CPU has a lot of dust as well but still far less than the GPU , that dust doesn't just stay around the chip , it actually penetrates under the chip around the solder, the nano particles of that dust is what goes under the solder with the help of humidity and over time starts to form a layer of isolating dirt and at the end one or more of the solder points will fail and result in all the problem that people are suffering with these computers , I repaired at least 25000 BGA packages in every electronic device and the reason I am saying it is not an opinion is that its what I find almost every time I get to reball one of these failed devices so its a fact and I don't know why you keep denying it, the material if it is leaded solder or leadfree won't make any difference and always remember that leadfree solder lasted for over three years in these computers and I don't think that leaded solder will give a better result or will even similar.

Hi gerrycramer -  Word to the wise.  In addition to backing up via TimeMachine, back up by dragging your entire user folder over to an external back up hard drive. Our IT originally backed up using Time Machine. TM was tested, found to work, then when my computer baked due to defective build, my Time Machine would not restore. My MBP was brought to Apple and they could not restore from TM. I decided to bring the MBP to my old digs and had some scientists work on the HD for 9 business days with no joy. The third time my MBP died, IT was able to restore from the user folder as the Time Machine did not restore (yet again). There is an unresolved bug that people calling Apple will likely be informed that they are the first person to call with this problem.

Apple has over $100B in cash reserves and is not offering a conditional out of warranty repair and explanation thus far.

I apply blame on quality control within manufacturing.  It is customary for computer manufacturers to model circuit boards using specialized circuit modeling software. During this phase, among other elements, heat dissipation is calculated, circuit board layout is established with heat sinks properly sized and chips placed with sufficient annular space to properly dissipate heat. Lifespan of every component is modeled. The exact migration of ions within each chip is measured in terms of months and years. This is nothing new. We've been performing this on computer and cell phone builds since 1986 that I directly know of, and likely before then as well.

ps3specialist appears to be the most accurate at estimating the source of the problem. A thin layer of dust or oxidation between the shiny chip surface and heat sink likely manifests prior to the application of the heat sink causing the attenuation of heat therms from migrating out through the heat sink. This will cause the chip temperature to rise above max temp and fail.  The temperature measurement is remote to the point of heat caused failure so unless an IR camera is used to monitor, (which we also utilize during design build of computers and cell phones) it is possible this problem shall continue unchecked.

Apple's Mac Book Pro mother boards should to be assembled in at least a class 1000 clean room. I the case of the 2011 defective mac book pros, perhaps the clean room was not maintained at low enough dust and or low enough humidity. Or, perhaps chips arrived with dirt or oxidation? it is possible the protective covers may have been removed long enough before the application of the heat pipe causing.

I do not believe that Apple engineered the product incorrectly such as accidentally sizing the heat sink improperly. I estimate the problem to be quality control. Apple might consider going back through the video images of the circuit boards that are captured at the assembly.  Apple is able take a few MBP serial numbers, look at the BOM, trace back to supplier, and supplier will be able to product details regarding production. All well controlled and accessible to Mr. Tim Cook and anyone he so directs. His challenge is allocation of resources. He would need to justify pulling resources off another project and few in Apple know how to pull this data and are busy on other tasks. So this will go unresolved until something gives. 

As this is a manufacturing quality control issue, and not poor engineering, I face Apple and expect them to offer up repair and more importantly to revisit manufacturing going forward. Keep up the good circuit design, but improve manufacturing ISO controls (perhaps via trusted third party audits by trained engineers, and not "consultants"?)

I have my own failed 2011 MBP sitting beside me here in my corporate HQ as a talking piece to other C-suite when they tour our organization for sharing of best practices.

I can't agree with a lot of your points - firstly these MacBook Pro chassis have been built on the same production (foxconn probably) lines as the other models from 2008-2012, millions of them in total.  The parts for all these models despite the different generations are roughly the same size so I don't think its an issue, and also the engineering guidelines in regard to thermal paste and cooling are almost the same too. Oxidisation occurs due to condensation,  but I too upon opening pretty much all MacBook Pro's and iMacs too see the very fine light grey almost white powder that is on these logic boards almost everywhere - the excess dried up shin-estu Apple stock thermal paste.

What makes the most stark difference in the generations of the MBP from 2008-12 is the type of CPU/GPU parts used, the amount of energy they consume and correspondingly the amount of thermal energy generated.

These 2011 models have a 45 Watt Sandy Bridge CPU with a massive die to cool, and a 40 watt AMD GPU, by far the hottest pair of chips to ever sit in this chassis.  Compare that to the 35 Watt Ivy Bridge CPU and 26 watt Nvidia Kepler GPU in the 2012.

Apple's error was thinking that they could get away long term with sticking two red hot bits of silicon in this chassis and do nothing about altering their engineering guidelines with using too much paste and poorly engineered heat-pipe plates originally designed for parts that run probably 40% cooler back in 2008. Re profiling these same poorly finished heat-pipe plates in conjunction with less and improved paste produces startling improvements in cooling.

But what do I know - I only fix these for a living and ignore bad advice and silly engineering guidelines when I see they aren't right.  Stupid I would say more for our 2011's.

Hi GavMackem - You state some strong facts and I certainly defer to you. The 2012 could likely be utilizing cooler running chips and perhaps I am splitting hairs as to how Apple is at fault whether the heat didn't flow aware from the two piping hot chips in close proximity, or if the heat doesn't get transported away through poor implementation of a heat-pipe.  I just returned from a walk and a thought occurred. Some boxes run hot, others not so hot. I noticed an inordinate amount of cycles and wonder if bad coding is also to blame? I wonder if the processor is running nested loops?  Whatever the cause, Apple is sitting on over $100B and needs to step up and turn this negative situation into a positive. If they do, they will reap substantial reward as they have in the past.

Apple needs to respond.  Their silence on this issue is unacceptable. Tim Cook is demonstrating poor leadership through his actions, or shall we say in this case, his lack of action.