A2159 (820-01598) liquid damage, PP5V_G3S power cycles to 5V

[2informaticos]

Onboard readings for resistors value are always lower than nominal value; in case of K/M range.

Do you get steady 3V3_G3S?
What about 1V8_G3S?
Also check all AWAKE and SLPS2R voltages, if stable.

 

[Fyo]

3V3_G3S, 1V8_G3S both power cycling (can barely see blip with multimeter, would have to test with scope), both kOhms to gnd.
PP3V3_AWAKE 3.25V shows for about 1s, then 0 for 2s. PP1V8_AWAKE, PP1V8_SLPS2R same, 1.8V for 1s. All mOhm to gnd.

 

[2informaticos]

So all power rails are pulsing.
Check for short on all big coils (from secondary supplies).

 

[Fyo]

The only coils near short are the CPU coils at around 3.4ohm, PVCCEDRAM around 30ohm, PVCCIO around 70ohm.

 

[2informaticos]

Can you catch any pulse on S3, or S0 power rails?

 

[Fyo]

PP1V2_S3, PP1V8_S3 nothing
PP1V_S0SW nothing
PP3V3_TBT_X_S0, PP3V3_S0SW_LCD nothing

PP5V_S0_BKLT_A, PP5V_S0_BKLT_D:

 

[Fyo]

This is what PPBUS_G3H looks like:

 

[2informaticos]

Remove audio board and check for liquid sign on both sides.
Also inspect audio flex and both connectors.

 

[Fyo]

I'm doing all this with the logic board out of the case. Nothing connected except USB ports.

 

[2informaticos]

And is dificult to check what I've asked you???
We need to know if T2 data lines might be affected for liquid spill on audio board.

 

[Fyo]

Sorry, I misunderstood. Going to check the audio board in a bit.

 

[Fyo]

 

[2informaticos]

Is that corrosion?
Check the other side, around audio amplifier.

 

[Fyo]

Yeah, looks like very slight corrosion. The liquid sensor is red. I don't see anything else that looks bad/corroded.

 

[2informaticos]

Leakage from PPBUS_G3H to T2 data lines is what probably happened.
In such case, it is garbage now.

 

[Fyo]

Any way to make sure?

 

[2informaticos]

Wait, you have 01598 board, not 01958; mea culpa, sorry.
PPBUS_G3H doesn't go to J6600, but 16V for Touch ID.
If your picture is for Touch ID connector, from audio board, very likely that corrosion damaged T2.
Just in case, check diode mode to ground on T2 data lines, related with Touch ID; all of them have "MESA" in the name.
Compare with similar working board, if available.

 

[Fyo]

All of the MESA lines are mOhm, a few kOhm, but nothing to compare to.
I tried connecting the board to another Mac and interestingly it takes 4.91V and stable 0.45A (about 2W).
The only areas that get warm are under the can (U3900, northbridge?) and L7030 (CHGR_LX1), but nothing around CHGR_LX1 is shorted.

 

[2informaticos]

U3900 is the f***g T2.

Do not use ohm scale when you check data lines to ground; diode mode is the way for comparison.

So no pulsing power on other shell?
Check all power rails above mentioned.