820-01521 MacBook Air 2018 with USB-C issues

[sekidata]

This 2018 Air came in with liquid damage in the area of the USB-C connector on the logic board. Pins 59 (20V) and 55 (Gnd) on the connector were visibly burnt, and a few nearby caps looked corroded (C6405, C6416, CC711, C3306).

After clean-up, USB is now stuck at 5V, drawing 0.08A. PPBUS is at 12.3V. I replaced both CD3215s, and I get all LDO voltages except for PP1V1_UPC_XA_LDO_BMC when plugging into port B. I replaced the Thunderbolt ROM (U2890) with no change.

The CC lines for port A (USBC_XA_CCx) are not what I see on a good board, while USBC_XB_CCx are correct. I checked resistances on the CC lines (in diode mode) and those are fine.

What should I check next to find out why PP1V1 and USBC_XA_CCx are not working?

 

[2informaticos]

1V1 BMC LDO only appears on the port where the charger is connected.
You can't get PP1V1_UPC_XA_LDO_BMC when connect the charger to U3200 port.

Did you test the board with both charger and battery connected?
Do you have S-UEM2 tester?
You may need to change J3300.

Check UPC_I2C_INT_L level.

 

[sekidata]

Sorry, I meant to write that I do get PP1V1_UPC_XB_LDO_BMC when plugging into port B, but not PP1V1_UPC_XA_LDO_BMC when plugging into port A.

When I have a battery and the charger connected, plugging into port B gives me 20V but 0A. Nothing at all on port A.

No S-UEM2 tester.

I checked for continuity of the CC signals between board and USB-C receptacle and they check out. So I suspect J3300 is fine.

UPC_I2C_INT_L is always 0V.

 

[2informaticos]

"UPC_I2C_INT_L is always 0V"
That is sign of USB-C problems.

Did you change J3300?
You've said had burnt pins...

 

[sekidata]

On a good board, UPC_I2C_INT_L only changes from 0 to 1.8V when USB-C switches to 20V, so no surprise there.

I replaced J3300, with no change.

 

[2informaticos]

If UPC_I2C_INT_L goes low, no 20V switching possible.
SMC is advised about a problem in USB-C area in this case.

You may have broken trace, or still corroded components.
Compare diode mode readings on J3300 with good board.

BTW, SMC is embedded into T2 chip.

 

[sekidata]

Diode readings of J3300 on this board match a good board exactly.

I think you're right: there's still an undiscovered corroded component or trace somewhere that's messing things up. I'll keep looking.