M3 A3114 MacBook Air 15" Very Strange Port Behavior

Hello again, this MacBook is acting strange and I'm struggling to figure it out. Here's what I've done so far:

At first it would charge fineish. Magsafe and DFU port charged just fine, drawing 70W and everything. The other USB-C port had no draw at all. Specifically, my meter went from 0.016V idle to 0.002V plugged in. It also does not recognize any USB devices on that port.I suspected the SN25A was bad after diode readings all came back normal, and since injecting voltage directly into PPVBUS_USBC2 had the expected behavior. I tried putting in a donor from a board that tbh I've kinda abused, and it wasn't booting at all. No 20V either. I put the original one back on and now it will only boot if the battery is plugged in. It won't negotiate 20V on cold boot, but once booted it will get 20V however no current draw. This is across MagSafe and USB. I have verified that both ports/cables are fine, it is an issue with the logic board.

"since injecting voltage directly into PPVBUS_USBC2 had the expected behavior"
I don't understand what this means.

You must replace SN25A chip together with its TBT ROM from donor board.

Injecting 5V directly into the USBC Power Input line (PPVBUS_USBC2) saw a 0.8A draw for a couple seconds before it stopped drawing. That indicated to me that it was a communication failure causing the drop on the ammeter. CC1 and CC2 were both normal, which brought me to the SN25A.

I thought the SN25A's were programmed? And even if not, the board I pulled from is an identical model so the ROM should be the same. All that said, I'll still try moving it, I would just like clarification.

"Injecting 5V directly into the USBC Power Input line (PPVBUS_USBC2)"
Why do you think that can help?
Do you know how USBC Power Delivery Protocol works?
The charger never injects 5V directly there, in the first moment!

CC1/2 pins are the first involved in charger detection.
Did you check diode mode to ground there?
USBC meter stays black, if detects a problem on these pins.

"identical model so the ROM should be the same"
Each chip has internal firmware, like old CD32xx series.
However, for SN25A need to change the corresponding EXTERNAL ROM too.
Seems that Apple paired it with the mainboard.

I did that to see if there was potentially a partial short that was causing the charger to drop voltage. Maybe I should read how PD works lol.

Those pins are 1.49 on diode with my meter. This is compared to 1.481 on a known working port.



Got it, didn't know the ROM was paired too. Apple bad I suppose. I'll update later when I've re-swapped it. In the meantime, in case that doesn't fix this - any ideas for why it DID charge before I attempted swapping, but now it doesn't even after reinstalling the original SN25A?

Bad solders, broken pads/traces.
The mainboard may be affected by the heat applied on the replacement process.

Alright, so I swapped the SN25A for both USB-C ports with the ones from the donor board, as well as the ROM. This board was known working, just iCloud locked. However, it has been through numerous heat cycles as I have pulled many components off it it, so the SN25A's and/or the ROM may be heat damaged.

I noticed something about this board though, only two SN25A's have a dedicated ROM. The MagSafe port and the DFU port. The other one has those pads to ground, so I guess it just piggybacks off of the DFU port's ROM. Just in case, I have a new donor board on the way. Here are my findings so far:


- With the battery plugged in, MagSafe will prompt boot, but won't negotiate 20V until reseated. Does not charge it (0.01A draw). Both USB-C ports now are not communicating with the charger at all, seen by the voltage pulldown and no current draw.

- With the battery not plugged in, and MagSafe connected, all power rails go live at their expected values, and PPBUS_AON sits at 12.5V. Interestingly, PPV1V8_UPC5_LDO is at 0V. When actually powered on, it still sits at 0V, with or without the charger connected. I'm not sure what the purpose of this net is.



I'm struggling to make sense of this behavior. Is there a power on sequence flow like there was for the T2 MacBooks?

What is the schematic used?

I have the XZZ schematic for 820-03286.

PP1V8_UPC5_LDO is generated by U5500 for internal use.
It must appear when the Magsafe is connected.
Each port controller chip should generate 1.8/3.3V LDO when gets power.

It's generating 3V3 UPC LDO and 1V5 UPC LDO. 3V3 S2 LDO is also generated. 1V8 UPC LDO is the only one that isn't generated. When magsafe is connected, 3V3 and 1V5 UPC LDO is generated at all 3 ports, but 1V8 UPC LDO is not. I don't know if they're only supposed to generate it when their port is connected or if all of them are supposed to when connected.

Either way, since PP1V8_UPC5_LDO is never generated, is that indicative of some communication failure (such as with the PMIC, since PMU is probably fine), or is it indicative of bad controllers?

Check 1V8_UPC_LDO on each chip, when the power is connected to corresponding port.

I'll do that once I have new donors for each port since both are currently nonfunctional as described above.