A2179 PPBUS_G3H low

[Gueric]

Hey mon2. I’ve got another crazy case for you that never happened to me before.

A2179 board, PPBUS_G3H short, 2 ohms.
injected voltage, 1V 5A limit, 200mA draw.
decided to increase to 2v, amperage draw instantly dropped to a few mA.

the resistance of the rail actually went from 2 ohm to 100 after that, and PPBUS went from 0mV to 700mV.

checked the resistance of the 4 current sensing resistors around the ISL, all fine. Removed the ISL, went from 100 to 160 ohm.

Replaced it, still 160 ohm.
Removed the 2 ISL mosfets. Still the same resistance. Replaced them both, still nothing.
Replaced the ISL AGAIN, went to 220 ohm.
PPBUS went from 700mv to 900mv.

im really confused now.

Also, PPDCIN_G3H is 5v as expected.

[mon2]

Hey. What is the 820-**** # you are referencing?

Remove F7000 (for now).

Remove all power. Meter in resistance mode. Measure the resistance between CHGR_CSIR_P and CHGR_CSIR_N. You can measure using the meter tips @ C7021 and C7022.

[Gueric]

mon2 I'm talking about the 820-01958 2020 Intel MacBook Air.

Remove F7000? the fuse that connects PPVBAT_G3H_FUSE and PPVBAT_G3H_CONN ?

Regarding the resistance measurement between CHGR_CSIR_P and CHGR_CSIR_N, I measured 2.6 ohm.

[mon2]

Resistance is fine.

Yes on the F7000 fuse.

Measure the voltage to ground of VDDA and also VDDP.

[dellxps15]

if u have a thermal camera would be much easier to spot where the problem is.... also sometimes a caps get shorted only when a certain of voltage is appikied, with all the risks that it give...

[Gueric]

Hey mon2 I’ve found a crazy technique to find PPBUS shorts.

I removed 2 current sensing resistors, splitting PPBUS_G3H_SPKRL and PPBUS_G3H in half.

Also another to separate PPBUS_HS_CPU from PPBUS_G3H.

I now know the shorted part of PPBUS is PPBUS_HS_CPU.

PPVIN_G3H_P1V1_S3 is also 200 ohms, since it's another name for PPBUS_HS_CPU but I don't wanna remove its current sensing resistor yet.

To remind you about the issue a little, resistance was at first 0.0 ohms, and went to 200 after I went from 1v to 2v while injecting.

Also, I probed L7210, L7220 and L7230 in diode mode, and the readings are perfect according to OpenBoardData (0.106)

How would you find the short at such high resistance ? Do you think the CPU could be cooked ?

I don't have a thermal cam btw (sadly)
thanks

[mon2]

What is the exact resistance to ground of PPBUS_HS_CPU in ohms @ C7210 (pin 1 = +)?

[Gueric]

mon2 it's precisely 203,4 ohms on pin 1 of C7210.

[mon2]

Measure the resistance from C7210 to L7210. If this is ~0R ohms, then the high side mosfet of the regulator @ U7210 has leaked and yes then the CPU is cooked from this high voltage rail.

Repeat to test the resistance onto the other same CPU power rails @ L7220 and L7230.

[Gueric]

So I just checked mon2, and resistance between C7213 ( a PPBUS_HS_CPU cap just below U7210) and L7210, L7220 and L7230 is slightly below 300 ohms.

[mon2]

Test resistance between C7213 & L7400.

[Gueric]

mon2 it’s somewhere around 600 ohms.

[mon2]

This is a tough one. I think that you will have to remove each current sense resistor between the CPU / PCH and each of the regulators, one at a time to determine which side is with the short. Test the resistance after each such part is removed - keep track of the part and their locations. As before, you are attempting to isolate the sections of the circuit.

[Gueric]

mon2 could you explain a little further ? I know where those 3 regulators are, just next to the 3 big inductors, but there are hundreds of pads under the CPU and so many different lines for the CPU/PCH that I'm a bit lost on finding those current sensing resistors.

Also, there's one chip that PPBUS_HS_CPU passes through, U7700. and there's also a mosfet fed by PPVIN_G3H_P1V1_S3 (another name of PPBUS_HS_CPU) that's obviously also 200 ohms.

[mon2]

Every switching power supply will have a current sense loop to monitor the amount of current being sinked by the load.

For example, for L7400 it is current sense resistor @ R7430.

For L7210 it is R7210, etc. Albeit a pain, suggest to flux and remove each of these resistors to isolate the producer (regulator) and the consumer (CPU) sides of the power rail. Be sure to keep track of their placement orientation (they can get funky on their SMD pads / shapes so orientation is important). Do not mix up the parts. Not a fun exercise but needs to be done with hopes to hunt for the shorted part.

[Gueric]

mon2 okay I think I get it. So removing the 3 current sensing resistors (R7210, R7220, R7230) would help making sure it's not the CPU/PCH that's shorted internally? because upon checking PPVCC_S0_CPU was not shorted and diode readings were fine.
But that wouldn't help directly with the PPBUS_HS_CPU short would it ? There are a few chips that PPBUS_HS_CPU passes through, obviously those 3 regulators, but also U7700, U7100 and Q8100.
Thanks for all your help again.

[Gueric]

I've just removed the high side mosfet to try and see (U7210) and the short is still present on PPBUS_HS_CPU. I also removed R7101 to cut off the power to U7100 and same result, short still there.
mon2

Edit : I actually went ahead and removed all 4 regulators, nothing's changed. resistance has dropped to 200 ohms instead of 230.

[Gueric]

I just also removed U7700. short still the same. 200 ohm on PPBUS_HS_CPU. I'm fully lost now

[Gueric]

from my understanding there is one single big component left on PPBUS_HS_CPU and it's Q8100.
I also haven't removed the 4 current sensing resistors. just the 4 regulators.

[Gueric]

hey mon2 i found that short by injecting voltage again. i had to step up to 3v for it to show. small cap drawing 2v at 5 amps, literally smoking. @C7401. now i gotta put every component back on and pray. By the way, do you know if SW6955 if compulsory for the board to boot? i think i damaged it while removing Q8100. its SYS_DETECT and GND.