820-01700 A2141 No WiFi, BT or iSight

[ugamazing]

Working on an 820-01700 board that has no WiFi (WiFi equipment detected, but when I click to turn WiFi on/off, the OS doesn’t do anything—doesn’t turn on or off, never shows WiFi networks, etc), no Bluetooth (missing completely) and no iSight (“no camera detected” and “no capture devices” in system information). Figured an issue with the 3v3/5v supply as with older boards, so swapped it just in case as I’ve seen odd issues even with all enables present. Still no dice.

Everything else works fine, including an external drive that I copied data over to without issues. Just no WiFi/BT or iSight. Tested with 2 known-good units/LCDs etc. Also, I can NOT get the unit to PRAM reset, but wasn’t sure if maybe some funky T2/etc issue might be conflating that.

Unit had liquid to top-case that was cause of no-power, but replaced without issue. Liquid to keyboard/tpad might be related. Board did have VERY minor liquid, but it didn’t touch anything at all (literally no components, just minor residue). Keyboard/tpad took brunt of the damage.

Any input or recommendations on where to start? Trying to avoid wiping board and reflashing T2 if at all possible.

Thank you in advance!

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

Going to dive a bit deeper into this board today; if anyone has any tips in the meantime, I'd be very appreciative! Will post some more info on board today.

[mon2]

Re: 820-01700 A2141 No WiFi, BT or iSight

WIFI on this logic board is a PCIe device.

Confirm the attached referenced caps are present for the PCIe interconnect.

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

Thanks, mon2!

I had more time to sit down with the board today, and made some progress. I did first confirm all of the caps for PCIe are in-tact and undamaged.

Eventually found PP3V3_G3S_X shorted to ground completely; removed everything on the rail before finally splitting the connection for PP3V3_G3S_X-->PP3V3_G3S_WLAN @ R5530. Once I did this, short went away. Of course with PP3V3_G3S_X being pulled low, that means our RTC signals (for cam, BT, etc) can be all screwed up.

Starting to suspect U3730, I found P1V2_WLANBT_VLX also shorted to ground completely. Not much on that line, so pulled everything except U3730, which I'll do next.

I had a question about U3730, however: Is this chip programmed at all? I have several 820-00850/01041 boards but only ONE 01700 parts board. So, I can pull U3730 from the 01700 board of course, but wanted to check if U3720 had programming in case I damaged that single 01700 U3730, or if I had "several chances" with my other 00850/01041 boards. Does this make sense?

I'll go ahead and get U3730 pulled now to confirm it resolves the shorts. Will then put everything back before replacing U3730. If there's no programming on U3730, I'd rather pull it from an 00850 because it will be easier.

[mon2]

Re: 820-01700 A2141 No WiFi, BT or iSight

Sorry but do not know if the part is programmed or not. My guess is that it is not. However, it is one massive part to desolder -> reball and place back.

We recently de-crated our massive BGA rework station (took 4 of us to lift it) but still reading up on how to use the beast before pulling our first BGA device. Apparently this one can support 0.5*0.5mm BGA to the much larger CPU & PCH devices. Hope to have it working this week.

On the P1V2_WLANBT_VLX short, can you remove L3702 and confirm if the short remains ?

What is the resistance to ground with L3702 removed for

L3702,

pin # 1 ?
pin # 2 ?

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

Thanks, mon2! Yeah I actually replace many BGA chips by hand, lol. We have a rework station as well, but I haven't touched it in years. Might need to break it out again now that Apple is consolidating so much into larger chips. Looks like these U3730 chips have a funky pin layout as well, so should be interesting. I am going to try an 00850 chip first and see if they do have programming or not. Will not be fun, but it's necessary!

Also yes, I removed L3702 to confirm STG on P1V2_WLANBT_VLX when diagnosing. Pretty sure the issue is U3730 at this point, as short was on pin 1 of L3702, and there's nothing else on that line. I did inject a bit of voltage (2v/1a) into PP3V3_G3S_WLAN and U3730 got slightly warm. I'm sure it would get warmer if I up'ed the voltage and current a bit.

Thank you again for your help!

[mon2]

Re: 820-01700 A2141 No WiFi, BT or iSight

No worries. Very curious on your progress. Keep posting your updates.

The reason for our investment for the BGA station is primarily to allow for proper alignment of the devices. Our station is fully automatic once we use the up / down camera to align the pads with the BGA device. Hope to also use the same tool to upgrade macbook / iMAC logic boards to support higher ram. There are a number of such logic boards with soldered memory devices. We outsourced the project to a firm in Malaysia (found them on youtube) and the results were very positive. Great company to use and the client was super pleased to upgrade his iMAC to 16GB. Personally have done enough BGA rework with just a hot air tool but we really need to invest into a nice microscope soon My eyes do not cut it these days.

Which microscope do you use? Not to take this thread OT but Amscope is all over the place but no offense - they appear to just be flipping imported scopes from Asia. I think we can do better to source direct from the mfr in China. Alex (NorthRidge Fix) suggests to use the table top scope for soldering but that is very different than using the eye piece version of the same. The resolution is much better on the table top models vs. trinocular style. Doing too much homework on this but do not want to get burnt either.

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

After comparing chips, I think U3730 might actually have programming...but I'm not sure.

There seem to be three U3730 variants for 2018/2019 00850/01041/01700 boards: 339S00610 (01700); 339200428 (some 00850 & some 01041); 339S00458 (some 00850 & some 01041).

All the 01700 boards I can find seem to have only 339S00610. Might not risk it and just pull it from the 01700 board so we're replacing like-for-like.

Yeah I've hand-soldered 100% of the BGA repairs we've done the past 2-3 years. Have replaced PCHs by hand, SSDs by hand, RAM by hand, etc. It's a huge time-investment, but we don't do a TON of BGA repairs as it is. I don't actually mind removing/replacing (I do use reference photos to help align, eyes holding up thankfully), but reballing by-hand is 100% my least-favorite thing to do, lol. I would MUCH rather just pay more and have all the chips reballed and ready to go, but I get nervous about them being "good" chips, so tend to do all my own sourcing/prep.

We do use Amscope microscopes, but we bought them 6-7 years ago. I'm not sure if they've declined in quality over that time, but we've not had any issues with ours. We don't use any digital/USB/camera microscopes at all.

[mon2]

Re: 820-01700 A2141 No WiFi, BT or iSight

On the reballing - what is the secret? We brought in some razor thin stencils from Asia - will confirm the brand name tomorrow. This along with a stone block to hold down the memory BGA device with the stencil. After pasting the solder paste and applying gentle but enough heat, the stencil would spring board up due to contraction and destroy the perfect solder balls. Wasted the effort everytime. Complained to the 'factory' of these tools and they claim the setup works. From our research, the materials are NOT direct heat style of solder stencil. It was a poor design for sure.

We have since then brought in everyone's stencils, including my mother's for further testing. The 'direct heat style' stencils are rigid and thicker for sure. Have not had enough free to time to experiment but hope to do so soon. We are design engineers and this repair stuff is a local sideline and hobby to learn more about the industry. The design work is keeping us very busy these days, especially with the semiconductor shortages. OEMs need their products reworked to support parts that are still flowing in the marketplace. Covid messy blues.

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

OK, so I got U3730 replaced on the board and all seems well. WiFi, BT and iSight all returned, so looks like U3730 was in fact screwing with the clock/RTC signals for several things as well as not allowing power to WLAN of course.

Here are a few photos: https://imgur.com/a/03Tyk7E

Five photos:

1. Replacement U3730 chip pulled from our lone 01700 parts board
2. Reballed U3730; I hand reball these by tinning both the chip pads and the board pads. Not the "correct" or preferred way, but without stencils and relatively small pitches anyway, it's the easiest way on these larger chips. The underfill is a pain, but just heat it and either slowly remove with your iron or slowly chip it away with tweezers; I do both, depending.
3. Removed chip site on 01700 board; not cleaned up yet, just after pulling bad U3730. Confirmed both shorts disappeared once chip was removed.
4. Prepared site; tinned pads and cleaned as much underfill as I could without risking damaging any pads/traces. I did pull one disconnected pad in bottom right of photo during cleaning. I also damaged one more pad (top right) while actually cleaning all the flux away, and not during preparing the pads. Not really sure what it got caught on, but I looked down and it was gone, so had to run a quick wire for that one.
5. Last photo shows new chip in place. I took this photo before securing the wire with new conformal coating, but it's now sealed up.

So, not a fun job--around 2 hours start-to-finish--but a necessary one. I did go with the identical U3730 replacement, and it seems to have worked out. Going to let the unit run for a while to be sure, but it seems OK so far!

As for the secret on reballing: It really all depends on the chip and the site. Reballing an old SMC or MUX or CD3215 (stencil and solder balls) compared to larger chips or ones with more complicated (or less-uniform) pin layouts like the CD3217, U3730, or smaller BGAs like ISLs, etc, then I just do a "poor man's reball" by heavily tinning all pads. You do need to make sure they're all roughly the same size/amount of solder, but it's actually quite simple to visually identify under- or over- filled pads/pins. It certainly takes some practice, but it's doable, and is usually quicker than reballing with a stencil and balls (or paste).

I have never used solder paste in my life; I use solder balls. I can grab one single ball with my tweezers, set it in place, heat for a second and it's in-place. So sometimes I'll just quickly throw balls onto a chip by-hand. It sounds like it would take FOREVER, but it really doesn't. You can grab a solder ball without putting down the hot air and just grab-->place-->flow-->repeat for 10 minutes and you're done.

I do prefer smaller, thicker direct-heat stencils. I do not use larger or mutli-chip stencils for the reason you mentioned--warping and general hassle. I have cut stencils up before that came as a small sheet/etc, because they bent/flexed too much.

But patience is definitely a key part of replacing larger BGA chips by-hand. You have to be patient while the chip preheats, then you have to choose an end to lift (I typically go for the one that will release "first"--i.e. less ground pads and less thermal mass) and just be very slow, methodical and controlled. Don't slip! Apply your heat as you're GENTLY lifting--I let the weight of the board pull the chip off. If there's underfill, you will need to remove the edge strip on each side first, and then clean up any that got too-far under the chip from the factory. Sometimes the underfill stops a bit inside the chip footprint, other times it goes ALL the way through which is a hassle. How we clean it depends on the site we're cleaning and fragility of the pins/traces. It's no fun running multiple wires underneath a BGA chip, so it pays to go slowly.

Will post one last time on this one after it passes all diagnostics! Thanks again for your assistance, mon2!

[ugamazing]

Re: 820-01700 A2141 No WiFi, BT or iSight

This unit passed all diagnostics today and went back out this afternoon. U3730 was a bear to replace but it went well. It was shorted at PP3V3_G3S_WLAN (-->PP3V3_G3S_X), which knocked out RTC circuit for iSight. It also was shorted at P1V2_WLANBT_VLX, and WiFi and BT both weren't working. All issues resolved by replacing U3730. Not the prettiest repair, but not many are these days, lol.

[mon2]

Re: 820-01700 A2141 No WiFi, BT or iSight

Wow. Well done! Thanks for the advice on the stencils. Facing a nasty a1708 repair with no ppbus_g3h. Down to smbus sda line being stuck low at 0v8. Will replace the 2k pull up tomorrow once they arrive from Arrow. Life sucking repairs...