A1990 820-01326 stuck at 5v and 0.19A

[ahoonid]

Hi everyone.

I have a MacBook Pro 15" A1990 touchbar with (I believe) a dedicated Vaga GPU. The exact Boardview and schematics doesn't seem to exist but I've been using the one for 820-01814 so far.

The MacBook doesn't turn on, stuck at 5v and drawing only 0.19A.

All 4 USB Cs ports behave exactly the same.

There's no PPBUS G3H available.
The resistance to ground from PPBUS G3H is 0.4 ohm. I have found the same resistance to PPBUS G3H from two coils near cpu/gpu area.

Since the boardview I have is different exactly in this area, I'm only guessing the two coils to be LA650 & LA640.

Should I assume the board brain dead at this point. Problem is I have no doner board if I need to replace any mosfets for the area.

[mon2]

Based on the above, very few parts left that could potentially be the fault.

Carefully, flux and remove C7081. Do not lose the part. They are cheap to source if lost.

[ahoonid]

I think my initial assumptions are not correct at all.

I removed the F7000 and F7001 fuses and now the resistance to ground from the PPBUS G3H pad of the fuse is in K ohms and rising. From the other pad of the fuse (PPVBAT_G3H_CHGR_REG) I have 0.2 to 0.4 ohms resistance to ground. Does this mean the short is on this side? Should I inject 1V 1A to this pad?
Could the issue be U7000 (Isl9240)?

[mon2]

Has U7000 been reworked or replaced?

yes, short is on the side with the low resistance to ground.

[ahoonid]

U7000 hasn't been touched. Yet. I injected 1V 1A to the shorted side but nothing gets hot and 0.93A is drawn. I changed it to 1V and 3A and no change in the current drawn and no heat anywhere to be found. I don't have a thermal camera but used the alcohol test carefully with no luck! Everything is cold!

[mon2]

Remove all power. Meter in resistance mode.

What is the resistance to ground in ohms @ L7030 (either side is ok)?

Suspecting that there is a shorted capacitor in either the group of caps mated to the D1 of Q7040 or in the group of caps @ C7066-C7069.

If you are comfortable, flux and carefully remove the current sense resistor @ R7060. Doing so will isolate the cap groups. Then test again the resistance to ground of the PPVBAT_G3H_CHGR_REG rail.

Respectively, test the resistance to ground of C7066. If the short remains, suggest to carefully flux and remove each cap starting with C7050 (flux and lift up one leg to remove the part from the rail). Bulk tantalum caps are common points of failures. We do not use them at all in our product designs and have devices running 24/7 for years and years and still working.

[ahoonid]

Thanks mon2.
Resistance to ground @ L7030: 35 k ohm

Ok I removed the R7060 current sensor now. PPVBAT_G3H_CHGR_REG's resistance to ground hasn't changed. I have 0.2 to 0.4 ohms again.
Resistance to ground of C7066 is in Mega ohms.

Probably a very stupid question but I need to learn and confirm anyways : D
Before taking out the current sensor, on continuity mode of my multimeter, both C7066-C7069 were beeping non stop. After the sensor is removed, they no longer beep continuously. Does this also confirm the short is no longer from this side? Because on the other side, all C7050 series are still beeping non stop.

Is there another way to get a sense of which cap to start removing first?

[mon2]

Correct on your understanding. Since you have removed the current sense resistor, the C7066-C7069 bank of parts are no longer mated with the shorted part. This confirms that C7066-C7069 are not suspects.

There are not too many caps in the remaining cluster of caps. We are assuming that one of the caps is guilty so you will just have to patiently flux and lift up one leg of each cap till you (hopefully) locate the shorted capacitor. Essentially, you are plucking parts from the shorted rail in hopes of finding the bad one.

Another method which we have not yet used is to purchase a milli-ohm multimeter. They are available for $50-$100 (estimated). Using this meter, you can test each cap while it sits on your board. The part with the lowest resistance is the guilty one. Amazon and Aliexpress offer these tools and is on my wish list to buy.

[ahoonid]

Well I pulled one leg on each and every one of these bastards and I still have a short. C7054 was a tiny nightmare that I had to use the heatgun for!


PPVBAT_G3H_CHGR_REG's resiatnce to ground is still 0.4 ohm.

[mon2]

Can you confirm we are as per the attached diagram on this repair?

[ahoonid]

That's the exact current situation mon2 👌

[mon2]

Based on the above, very few parts left that could potentially be the fault.

Carefully, flux and remove C7081. Do not lose the part. They are cheap to source if lost.

[ahoonid]

Done. No longer continuous beep. Resistance to ground is OL now! Was that the issue!?

[mon2]

Yes sir. Must have been an in-rush current and was too much for this wimpy capacitor.

[mon2]

Suggest to now, slowly add back the parts you have removed earlier but do so one at a time and test the rail as you go. Fairly confident this was the only fault.

[ahoonid]

The MacBook lives to see another day man. Your help and inputs are much much appreciated mon2!!! It was indeed C7081!

I would love to hear and learn how you put your finger on this cap specifically! Why this one and not C7080?? Not a problem if you're not in a mood to further explain. You've helped more than enough so far.


Anyways thanks again and cheers mate.

[mon2]

No magic. Just searched for the netname that you shared in your first post. Any part that touches that rail and also connects to ground is a suspect. This meant every cap so it was a process of elimination. One by one. Really did not want the nasty tiny wlcsp regulator to be shorted and we got lucky. If you wish to expand your toolchest, consider the milli-ohm capacitance meter. When I was in HK last fall, met a really nice tool company with gorgeous large LCD displays on their meters and they offered such a tool. May consider one sooner than later. With that tool, only would have to test each suspected part to measure the resistance to ground. This part would then have had the lowest resistance - it is that sensitive. Can measure in milli-ohms. Amazing tech.

Glad you are up and running - will ship some repairs to your side when piernov is too busy

Can you believe that this one tiny part caused this rather costly macbook to fail?? Mention it next time you are at the Apple nerd herd cafe at their mall stores.

[ahoonid]

Got it 😁
Yeah expensive af but funny enough I got the laptop for dirt cheap man! I got it off of ebay for like 80€!! 32TB RAM 2TB SSD! Can you believe it ))) I couldn't allow myself for it to stay dead.

Shipping from France to Germany will make any repair economically not viable 😆. I'm studying my master's and only do some repairs as a hobby from time to time. But who knows maybe I'll make a side job out of it some day. For now I'm just happy I'll revive stuff and not let these become land waste.

I'll definitely think about the advanced multimeter. I actually need a new one anyways as this one gives me a headache from time to time. However I'm trying to buy a thermal camera for now. I've decided on a Infiray p2 pro but am totally open for suggestions at the price range.

[mon2]

There is a great and very in depth review of assorted thermal cameras on EEVBLOG.

< will take a while to traverse through these threads but worthwhile >

https://www.eevblog.com/forum/thermal-imaging/

We purchased the high end one from Qiani and cringe on their s/w and very tardy refresh rates. In HK I met a vendor who offers 25Hz refresh rates (rather than the 8Hz). The 8Hz was imposed by the US DoD. The one from China is sold on Amazon and is noted in the EEVBLOG forums. Next time, would consider to use one that mates with a phone to be more portable.

Good deal on your Ebay sourcing. It is a great skill to have to learn on repairs. Confident you will be busy if you take on more work in your region.

[ahoonid]

Thanks for suggesting the forum. I'll give it a read before buying the infiray for sure 👍.
8Hz!!! DoD is being selfish here 😆

[ahoonid]

One last question mon2.
I don't know if it happened by me or by the previous owner while pulling out the back plate, but two caps were knocked off of their pads. Most probably I'm not to blame cuz I usually am very careful with this and there were dents on the edges near the screws so it was definitely forced before.

The caps were C9591 and C9592. One of their legs were still connected and upon checking I noticed one of them was shorted. I fixed the healthy one back in place but I had issues finding the exact replacement for the shorted one.
The caps are both 20 µF 20% 10V.

No other doner boards I have had this exact cap. I couldn't even source the exact thing from the internet. The closest thing that exists seem to be the more standard 22 µF 20% 10V. Since this cap is near the nand ssd, I was worried if the smallest irregularities corrupted the information on the ssd. I did however find 2 x 10 µF caps with exact other specifications and connected them together before mounting them on the same pads. Which together become the exact one cap I need

This was only a temporary fix to see if the other issue would be solved which it did. Now I wonder if I should let it be the way it is with 2x 10 µF caps or buy a 22 µF cap instead?