A1708, 820-00840-A, 20V but stuck at 0.045A

[Theking3737]

Hello everyone,

I've recently acquired a 13 inch MacBook Pro 13 2017 (A1708, 820-00840-A). I bought it with the intention of trying to fix it and learn a thing or two about electronics repair along the way. So I've never repaired or even had another MacBook or Apple product before actually. However, I do have some experience with designing and producing my own PCB's so I do have some equipment like hot air station, oscilloscope, microscope, hot plate, etc at home. Today I received my USB ammeter and got to work.

The USB ammeter first showed 5V and after a few seconds it was up to 20V, so if I understand correctly that means that both of the CD3215 IC's are doing great. But of course I bought the thing as defective and that showed... When it reached 20V I noticed it got stuck at 0.045A. Weirdly enough it got stuck at exactly 0.04500A. Sometimes it would briefly go a few decimals off but then quickly go back to exactly 0.04500A.

The seller mentioned that the device broke after trying to upgrade the replaceable SSD.

When measuring with the USB ammeter I had already removed the logic board from the case so I moved on to measuring some voltages with a multimeter.

PPBUS_G3H	12.96V
PP3V3_S5	0V
PP5V_S4	0V
PP3V3_G3H	3.38V
PP5V_S5	0V
PP1V8_S5G	0V
PPBUS_G3H_CPU	12.96V
PPBUS_G3H_SSD	12.96V
P3V3S5_EN	0V
P5VS4_EN	0V
P1VOPC_EN	0V

I wonder if anyone has some tips for me on how to continue and what to measure next. I know someone who has a FLIR camera which I could borrow if necessary to better visualize voltage injection.

Thanks in advance

[mon2]

Remove all power. Meter in resistance mode. Measure the resistance to ground of each inductor (usually grey in color; 2 leads). Checking for lower than normal resistance to ground. Report any suspects for a review.

[Theking3737]

Remove all power. Meter in resistance mode. Measure the resistance to ground of each inductor (usually grey in color; 2 leads). Checking for lower than normal resistance to ground. Report any suspects for a review.

Here are some measurements. The ones I left out were all super high resistance, usually in the MΩ range.

L7430	2.1 Ω
L7410	2.1 Ω
L7420	2.1 Ω
L7210	4.1 Ω
L7370	6.5 Ω
L8060	0.4 Ω
L7900	0.4 Ω
L8030	212 Ω
L7960	35.9 Ω
L7690	0.5 Ω

L7900 seems to be dead short to ground. This kinda worries me since I saw it's connected to the PCH...

[mon2]

L7690

0.5 Ω

= PP3V3_S5

This is a show stopper. You have a shorted part on this rail. Would you have an adjustable power supply? Consider to use it at 1 volt (to be safe; not higher) and inject onto this inductor. Whomever is shorted will heat up and/or crack under the injection pressure. Often it is a shorted capacitor. The PCH may be also low but start with this review and hunt down the suspect.

[piernov]

3 power rails going to the PCH shorted to ground? Conclusion should be pretty obvious, no point in wasting time.

[Theking3737]

This is a show stopper. You have a shorted part on this rail. Would you have an adjustable power supply? Consider to use it at 1 volt (to be safe; not higher) and inject onto this inductor. Whomever is shorted will heat up and/or crack under the injection pressure. Often it is a shorted capacitor. The PCH may be also low but start with this review and hunt down the suspect.

I tried the voltage injection at L7690 but my power supply can't do more than 3A. It was near impossible to spot or feel anything heat up. I also tried pouring some IPA on the board but it didn't really help. I think I saw the alcohol evaporate a tiny bit quicker near the PCH, but it could just be my mind thinking I saw that because it seems likely to me that there's a fault there. I will try to borrow a FLIR camera next week and have a better look at it. I'm doing this purely as a hobby project and to learn a thing or two so there's basically nothing to lose at this point. I will update the post if I find something.

Could it be helpful to desolder L7900 or L8060 and see on which pad the short is on?

[Theking3737]

The PCH is dead...

[crnimrki]

HI.
I have similar problem as here described. Same board, similar resistance on inductors:
L7430

L7430	5,1
L7410	5,1
L7420	5,8
L7210	14,2
L7370	23,3
L8060	0,9
L7900	1,6
L8030	145
L7960	81
L7690	1,3
7220	14

When I inject 1V on L7690 current is 1,5 A and can not locate any heated element with the IPA method.
Does anyone have an idea how to more closely locate the short circuit?

Thank you all.

[mon2]

L7690 is a power rail (PP3V3_S5) that feeds many sections of the board including the PCH / CPU. Remove the heatsink off the CPU. Test again to see if the PCH is heating up.

[crnimrki]

L7690 is a power rail (PP3V3_S5) that feeds many sections of the board including the PCH / CPU. Remove the heatsink off the CPU. Test again to see if the PCH is heating up.

With 1V there is no warming on the CPU.

[mon2]

A thermal camera will really be helpful here. Suspecting a shorted capacitor. You can also consider a low ESR meter used to measure in the milli-ohm range for components. This process will then demand to test each and every part on the board to hunt for the shorted component.

Alternatively, but in my opinion risky, is to check if there is a path from this shorted one to the CPU rail / any low voltage rail on the board. If not, only then you can consider to use a higher voltage to 'smoke' out the shorted component. We are using a product called 'short killer' which is dialed up to 0v8 (by us) but will spike out 20A/30A surges. This combination cracks up the target part in seconds. No can hind under this pressure. Purchased it from Aliexpress. A good variable power supply should also be ample. Can you push out 5A or more on yours?

[crnimrki]

A thermal camera will really be helpful here. Suspecting a shorted capacitor. You can also consider a low ESR meter used to measure in the milli-ohm range for components. This process will then demand to test each and every part on the board to hunt for the shorted component.

Alternatively, but in my opinion risky, is to check if there is a path from this shorted one to the CPU rail / any low voltage rail on the board. If not, only then you can consider to use a higher voltage to 'smoke' out the shorted component. We are using a product called 'short killer' which is dialed up to 0v8 (by us) but will spike out 20A/30A surges. This combination cracks up the target part in seconds. No can hind under this pressure. Purchased it from Aliexpress. A good variable power supply should also be ample. Can you push out 5A or more on yours?

Unfortunately my power supply can only output 3A. I'll try to borrow a thermal camera somewhere. Thanks.
Question: Is there anyone who can borrow a thermal camera in the Nuremberg area (Germany)?

[mon2]

You can always try Amazon and return it as "do not need it anymore". There are a few good ones that mate to the phone. If you truly wish to purchase one, then review the detailed threads on eevblog:

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

These days, consider one with 25Hz refresh rate.

[crnimrki]

a product called 'short killer' which is dialed up to 0v8 (by us) but will spike out 20A/30A surges. This combination cracks up the target part in seconds. No can hind under this pressure. Purchased it from Aliexpress. A good variable power supply should also be ample. Can you push out 5A or more on yours?

It's unlikely that this would help, at 1V the current doesn't rise above 1.5 A. All that remains is for me to try with a higher voltage. What is the upper limit that would be relatively safe?
Unfortunately, my 30 power supply doesn't go below 8V.

[mon2]

Officially, the max voltage is the voltage value defined by the circuit. The concerns are, what is someone on this rail is leaking onto another lower voltage rail. For that reason, the general rule is to keep the voltage as high as the lowest voltage on the logic board which is often 0v8 to 1v0.

Suggest to inject the 1v0 onto the same rail, then hunt and measure each inductor on the logic board to test if there is a voltage from this injection present downstream.

You can also, measure the resistance from this rail to each and every inductor on the board. Do you see a short between the points of tests? No power of course during resistance checks.