820-01700-05 Liquid Damage Partial Power

820-01700-05 light liquid damage, concentrated on U9580, U9500, U3100, and U3200.

Reflowed U3100, U3200.
Reballed U9580, U9500.
Verified 20V on all 4 USB-C ports.

Current steadily climbs until peaking at 0.31A 20V before suddenly dropping off to 0.01-0.03A 20V, and remains steady there perpetually.
Trackpad clicks when connected, keyboard is screwed up so not testing with that, otherwise no other signs of life.

Worked through the power on sequence at:

Everything is fine until PP3V3_S5 never comes online. The PMIC does not appear to have been hit, so I'm guessing this is a logical issue elsewhere? I skipped testing PMU_ONOFF_L because that's going to be a PITA with my equipment, but can do so if it's particularly relevant.

Liquid damage was quite light, tested all affected caps, no shorts, and no passives were even damaged severely enough to need replacement.

What should I do from here?
 
No, this is a data critical job and I’m concerned that dfu could negatively impact the data, or complicate the repair by necessitating another dfu revive after everything I try.

What’re your thoughts?
 
Ran out of other ideas so I gave a DFU revive a try. Error below, I've found minimal information on this code. Looks like it's now auto booting to dfu.

1645374524166.png
 
Yes, all of the voltages in the SSD 4 LANDING and SSD 2 LANDING sections of the power on sequence document are fine.

It's possible that SSD0 got zapped with PPBUS, but there's no shorts, so it feels premature to call off recovery efforts based on that.

How would I go about getting a definitive answer on SSD issues being the cause of my stalled boot?
 
Looks like skipping over portions of the power on sequence that I assumed were required for later sections was a bad decision.

The first line with an issue is actually PMU_PVDDMAIN_EN, P3V3MAIN_PGOOD and subsequent rails in that section do not come online.

The last time I saw a failure at this point in the boot sequence, it was a bad CD3217, however then I had 5V, not 20V on my ammeter. I pulled both CD3217 on the side of the board that was exposed to liquid and they were fine underneath.

Is replacing them a worthwhile idea?
 

2informaticos

Administrator
Staff member
CD3217 chips should be good, at least partially.

I suppose that DFU cannot write firmware partition on SSD.
 
I don't understand how the problem could be SSD related when the SSD is not even powered yet when PMU_PVDDMAIN_EN fails to go high?
 
Also, can you elaborate on the "partially" with regards to the CD3217, is there something I can check?

This customer is extremely interested in data recovery, so I want to explore every possible avenue.
 

2informaticos

Administrator
Staff member
You've said to have SSD voltages before.
"Yes, all of the voltages in the SSD 4 LANDING and SSD 2 LANDING sections of the power on sequence document are fine."

CD3217 chip is very complex, like CD3215.
No datasheet available, but may be some copy of another TI chip.
For example, CD3215 has the same pinout as TPS65982/87; not compatible, but datasheet helps to make an idea about it.
I doubt to have a problem with CD3217, if you get 20V on input.
 
According to the power on sequence document for this device:

PMU_PVDDMAIN_EN is supposed to go high before anything in the SSD 4 LANDING and SSD 2 LANDING sections.

However, it fails to ever go high.

My question is, how can the SSD be responsible for this fault (PMU_PVDDMAIN_EN not going high) when the SSD is not yet powered (And not supposed to be powered) at the time when the fault first occurs?

Thank you for your explanation on the CD3217! I don't get 20V after attempting a dfu revive anymore, but I did before.
 

2informaticos

Administrator
Staff member
So you've lost 20V input after touching CD3217 chips?
Did you reball, or replaced them?

U7800 may be your problem.
Check its LDOs (SYM 1 OF 4), at page 76.
 
I'm not sure, unfortunately I didn't check this between the dfu operation and CD3217 work, but my guess is that I lost 20V after the dfu revive attempt, as my soldering doesn't often cause issues and I redid the CD3217s more than once with identical results.

I did the following:
-Replaced both CD3217s on that side of the board, observed 5v and assumed it was a result of the dfu revive attempt
-Attempted a second dfu revive, same error
-Reballed and restored both of the originals as I'd like my new chips back, still observed 5v

U7800 LDOs appear fine to me:
LDO0 0.8V
LDO1 3V
LDO2 1.2V
LDO3 1.8V
LDO RTC 1.5V
 
Top