Will I have a bricked logic board after reflow of the GPU v-core U8900 chip on my Mid-2012 15" Retina Macbook pro?

[korben_dallas]

Hi everyone! I'm trying to fix my black screen / failed U8900 GPU v-core on my Mid-2012 Macbook Pro 15" Retinaby doing a rework/reflow of the chip (once I determine that's the real issue; but it likely is).

I've been disabling the GPU in NVRAM for a few years... but it's just time for a better solution.

I'm new to soldering but the only feasible way to fix this issue is to do something myself.

MY QUESTION:
I've been seeing some people (on this forum mostly) run into issues with the logic board needing a serial number or it being deactivated (?) after reflowing the GPU v-core. However, I've also been seeing elsewhere that people don't have this issue at all.

Also, one repair shop I called said they dont this repair anymore and just replace the logic board. But other people I've talked to (or seen Youtube videos of - namely Louis Rossman and ddosdude) don't run into logic board issues at all.

So what am I looking at here? What's the real answer? Will I be running into a bricked device after refow of the GPU v-core?

Or perhaps I misread things and these "disabled logic board" issues are only for certain MBP models (not mine)...

Full model: Mid-2012 Macbook Pro 15" Retina A1398 (EMC 2512) MC975LL/A

[Sephir0th]

Finally, Post 1000

But back to the topic: Most likely yes.

You probably first need to realize that these are two completely different things when either an experienced chip-level technician like Louis Rossmann or a beginner with little experience works on large BGAs.

An incredible amount can go wrong in such a process. It's not for nothing that BGA rework is one of the absolute supreme disciplines at chip level.

However, there seems to be a second big misunderstanding, namely the "reflow" itself. Surprisingly, the theory that reflow would be of any use in the case of a black screen still persists these days. A dead chip is a dead chip and no reflow in the world can change this state.

Of course, experienced colleagues know that this is all nonsense and in most cases only makes the problems worse or even renders the device unusable. The only cases where a reflow would (temporarily) bring an improvement are the old NVidia chips with the Bumpgate issue, which were mainly produced between 2006 and 2008.

See also here: https://www.badcaps.net/forum/troubl...vidia-bumpgate

My understanding is that your MacBook is from 2012, so there is no indication that it would do anything. And even if it did, a reflow can never be called a "repair". It rarely lasts for more than a few weeks.

Even if there is an indication, for example in the event of a fall, the BGA would have to be completely desoldered and then the pads checked for damage.

All in all, this is a very bad idea and I strongly advise against it.

[piernov]

OP is talking about the GPU VCore buck controller IC, which suffers from bad soldering joints cracking under mechanical stress on the 820-3332 board. Still does not mean this is an easy fix for someone with no experience at all though.

The only cases where a reflow would (temporarily) bring an improvement are the old NVidia chips with the Bumpgate issue, which were mainly produced between 2006 and 2008.

Also this is not true, it's not reflowing that does something (as in melting the balls), it's the heat, and many ICs react to heat (same thing for the well known failures of AMD from around 2008-2012).

[korben_dallas]

Congrats on post 1000! I feel honored 🙂

Of course, experienced colleagues know that this is all nonsense and in most cases only makes the problems worse or even renders the device unusable. The only cases where a reflow would (temporarily) bring an improvement are the old NVidia chips with the Bumpgate issue, which were mainly produced between 2006 and 2008.

1. How would a reflow render my device unusable? In what ways can a logic board become unusable after a reflow?

2. But I thought the Bumpgate chips were just... faulty chips. How would a reflow "bring an improvement" to a faulty chip? I don't think my Nvidia Geforce 650M falls under the Bumpgate thing. The 650M was made in 2012. The 650M isn't the problem, it's the way Apple soldered on the chip (cheaply). This issue with GPU chips in Apple MBP's around this time was it's own big deal like "Bumpgate". They did a full "repair" program that expired, and afaik, all Apple did was add a "spacer" that would push the GPU vcore U8900 chip into the board so that there was a connection with the soder.

I've read comments from people (without soldering experience) have done this exact repair and been okay. It's only now seeing posts on this forum where a GPU vcore repair has resulted in a bricked logic board. I'll have to search these forums again to see what models they were repairing.

[korben_dallas]

OP is talking about the GPU VCore buck controller IC, which suffers from bad soldering joints cracking under mechanical stress on the 820-3332 board. Still does not mean this is an easy fix for someone with no experience at all though.


Also this is not true, it's not reflowing that does something (as in melting the balls), it's the heat, and many ICs react to heat (same thing for the well known failures of AMD from around 2008-2012).

piernov - Assuming I do the reflow successfully... are there any reasons why the logic board might become "bricked" or unusable after? Searching these forums, I see a lot of posts about serial number issues (?). What are those about?

[Sephir0th]

OP is talking about the GPU VCore buck controller IC, which suffers from bad soldering joints cracking under mechanical stress on the 820-3332 board. Still does not mean this is an easy fix for someone with no experience at all

Ohh, missed this detail. Then nevermind please...

Next time I have probably to look into the boardview before start writing...

[piernov]

If the repair is performed improperly, of course it can cause the problem to get worse, like any other kind of repair. But board repair with no experience at all can easily go wrong, so it's always advisable to practice on a worthless board first.
While it is not very common, the soldering joints issue on the GPU VCore buck controller can cause bad voltage regulation and can damage the GPU itself, and improper soldering when performing the repair can do the same.

[korben_dallas]

So to clarify, there shouldn't be any BIOS/SMBIOS/EFI issues if a reflow is performed on the discrete GPU, right?

[piernov]

No, but no reflow should ever be performed on a GPU, it won't fix any issue and it can easily make it worse by completely killing the GPU, damaging the balls when they were not the problem, or even damaging the board.