Microsoft Surface Pro 4 1724 no power

[johnbeton]

Hello,

This tablet has no power, no visual signs of corrosion or burned components.
I notice V_ALWAY_ON_IN on PD5501 and PD5506 is 12V.
But VDD_BATA_PACK on PD5501 is 2.05V when the charger is plugged in, after a little over 1 minute VDD_BATA_PACK goes up to 6.16V
FANG_PWR_F stays 0V.
+3P3VSB_REG is only 1.98V
+5VSB_REG is 0V

I cannot find shorted coils or coils with low resistance to ground except for the ones close to the PCH/CPU.

Link to boardview and schematics:
https://www.badcaps.net/forum/troubl...009-schematics

[mon2]

+3P3VSB_REG is only 1.98V
+5VSB_REG is 0V

Remove all power -> measure their resistance to ground. Post each measurement.

[johnbeton]

+3P3VSB_REG = 13kOhms
+5VSB_REG = 12.3Ohms (not sure how I missed this yesterday)

[mon2]

Anyone who consumes from the +5VSB_REG rail is a suspect.

Start with each of the USB load switches which often get shorted.

Review:

U4504
U6401
U6403
U4703

Check for damage to any / each of the above. If they look to be visually defective, flux and carefully remove the respective load switch off the board and test again.

[johnbeton]

I checked the above components very close, but no visible damage.
I'm thinking of injecting 1v 3a on this rail to search for a short.

[mon2]

I think that is safe. My hunch is that one of the USB load switches is nuked. Continue to update this thread.

[johnbeton]

I injected 1v @ 3a, but nothing is heating or warming up (not even after 15mins). The system only pulls 0,09amps.
Some parts are difficult to reach because they are underneath the frames that hold the metal caps. But I used IPA alcohol to check for heat dispense.
I don't know if raising the voltage is safe? Otherwise I'd think to remove the load switches one by one and test if the short is gone.

[mon2]

If you are comfortable with the idea - my spidey sense is still tingling for the USB load switches. Maybe someone inserted a portable toaster oven onto the USB port?

[johnbeton]

After removing U6403 the short is gone. When plugging in the charger, the symptoms remain the same but the 5V is now 0.91V for the first minute. Then goes back to 0V.
I probably have to source replacement for U6403 to do further tests.
I measured resistance between all pins of U6403 on the bench, but find no low resistance between pin 5/7 (gnd) to other pins.

[mon2]

These load switches are not required unless you wish to use the USB port linked to this load switch. These parts act like circuit breakers for excessive current draw by the target USB widget (ie. toaster, coffee maker, etc.).

USB was only meant to power low power products by default but these days anything goes.

Have you also removed the other USB load switches? They are also suspects.

[johnbeton]

I just removed those 4 load switches, but the problem remains, I think I have to further investigate this 5v system

[mon2]

Remove all power. Meter in DIODE mode. Test each of the diodes onboard that are mating with VDD_BATA_PACK (ie. PDxx parts) using this mode of testing.

Red meter probe on the anode (backside of the diode arrowhead symbol); black to the other side of the diode. The diode should conduct.

Reverse the meter leads and the diode should block. Test each and every diode onboard. Someone may be leaking.

[johnbeton]

VDD_BATA_PACK seems good, I checked PD5501 and D7403, they don't leak.
I checked all PD parts on the board and find a leak on PD7501 (tested with PD on the board).
Red probe on Anode is 0.273 and Red probe on Katode is 1.043

[mon2]

Is it the attached diode you are testing?

If yes, flux and carefully lift one leg of the part but only if you feel comfortable. Do not believe this is a fault but can double-check.

After you lift one leg -> allow this part to cool. Then test again while one leg remains on the board. The diode must block in the one direction (not boy band).

[johnbeton]

Yes, that is the one, but it measures fine when I take it off the board

[johnbeton]

If I follow the power sequence, first in line that has 0V is VCCDSW_EN
maybe it is a defective U3601 (IT8528G).
I found this topic on this forum with a comparable problem:
https://www.badcaps.net/forum/troubl...oesn-t-turn-on

I'm just thinking out loud, the changes in voltages after 1 minute must be something controlling this... maybe this IC

[johnbeton]

I decided to order a new IT8528G and give it a try, after couple of weeks waiting the chip arrived, and now the problem is solved.
So the initial problem was the IT8528G.

[mon2]

Awesome! Excellent update. Was the IT8528G pre-flashed by the vendor or was it blank when soldered onboard? Any special updates required to use the new IT8528G?

[johnbeton]

it was pre-flashed by the seller. After replacement all voltages were fine and the device startsup and charges.