820-4924-A No ACIN on U7100 pin 3...

[Holes Flow]

Experts,

I hit a wall, and could use your insights on how to get moving.

A1502 820-4924-A
70mA when in case, 20mA when not. No fan spin.
Unit is in S3AC mode, per chart on schematic p.64 using SMC_ADAPTER_EN, SMC_PM_G2_EN[ABLE], and SMC_S4_WAKESRC_EN

Symptoms:
1. Every rareth time, system comes up. shows battery as dead (big red battery icon), not charging.
All other times (virtually all), does not come up.
2. One pin pair on the magsafe blew out-both sides black, some metal from the tip of the PS plug gone. Doesn't seem to have made a difference, except the 'other' pin on the PS plug isn't pogo any more (stuck).

pic of PS plug:



...and the LB:



Steps:

All "G3 Hot" (Always-Present) rails OK.
Starting on 5v:
1. OK = PP5V_S5_LDO
2. 0v = PP5_S5_FET mostly 0V, sometimes 5V after PS on a while!? (@U8060 pin 5/6)
3. U8060:
pin 1/3: 5v (PP5V_S5_P5VS5FET [PP5V_S4])
pin 2: 0v (P5VS4_EN in)
4. Tracing P5VS4_EN tp p.64 at C8192 = 0v
5. Tracing back to U8170:
pin 6: 3.3v irrelevant (PP3V3_S5_PWRCTL = power for OR gate 8170 (NOSTUFF))
pin 2: 0v (PM_SLP_S5_L) per chart on same page, it is only HIGH when in modes S3, S3AC, and S0. This pin Low makes P3V3S4_EN low as well.
How to make PM_SLP_S5_L go Hi?
Pulled L by R1460 (p.13), Pulled H by U0500 (Broadwell) pin AP5 (SLP_S5*) on same page.

So, I'm stuck at the Broadwell :-(
I was hoping to find out why we were stuck in S5 (even though p.64 table says in S3AC...)

TIA!

 

[2informaticos]

First of all, welcome to the forum!

You have 3V3_S5 present, but PCH requires more signals, before starting power on sequence.
Post voltage for 3V3_SUS, PCH_DSWVRMEN, PM_DSW_PWRGD and PM_BATLOW_L.

Be sure to use original 60/85W Magsafe 2 charger.

 

[Holes Flow]

Thanks @2informaticos. You guys are great.

The first 3 signals were all 3.3V, and PM_BATLOW_L was 0V, but that makes sense as the LB is not in the case. Checked the MOSFET "BATLOW# ISOLATION", nothing on any pin.
Gate signal comes from PP3V3_S4_TBT, and U3210 seems to be off, with both ENables (EN & HVEN) at 0V, and PP3V3_S4_TBTAPWRSW at 3.3V.
The magsafe LED goes from green to orange, and stays there.
The SSD is cool, like everything else on it.
Am I right that the LB is stuck at S5?
It's been a week, and it stinks that I have no distinct signal symptom to chase yet...

Thanks again!

 

[2informaticos]

PCH requires high level on PM_BATLOW_L to start power on sequence.

"Be sure to use original 60/85W Magsafe 2 charger."
???

Be aware of wrong assumptions; don't waste time with thunderbolt area.
Board MUST turn on alone on the desk, even without LIO flex connected; only needs LIO power cable.

 

[Holes Flow]

Thanks @2informaticos !
I think I made progress. But first:
"PCH requires high level on PM_BATLOW_L to start power on sequence. "
Just by the name of the signal, does this mean a battery that is not 0% is required to boot? I must not understand the signal.

re: Progress:
0. 820-4924-A p. 75/sheet 58.
1. I replaced TPS51980.
2. When checking for solder bridges after, I had a short between pins 23 (VIN/PPVIN_S5_HS_OTHER3V3_ISNS)& 24 (DRVH/P3V3S5/P3V3S5_TG). Great! Too much solder...
3. I pulled the new one off, and the short was still there!
4. Here is the relationship between the 2 pins:



...and here is the MOSFET (Q7560/FDMS3602S), just off to the right of the above image:


Should the Gate & Drain on this dual N-Channel be shorted? My instinct says no. And because the signal to pin 24 takes a direct path to/from the MOSFET Gate, there are no alternate signal paths to short indicated. With TPS51980 removed, and on the pins of the MOSFET directly, Gate & Drain are shorted.

Thoughts?

TIA!

 

[2informaticos]

Replace Q7560.

Be aware, do not power on the board without U7501, if Q7520/60 are soldered!!!

 

[Holes Flow]

Thanks @2informaticos! I screwed up and removed U7501, and cleaned/tinned the pins-all with the power on.
Ordering dual MOSFETs for both sides now. It sounds like I shorted out Q7560 (if this is how they fail).
Dohh!

 

[Holes Flow]

Well, all MOSFETs replaced, and now I have a very strange symptom-once again on the U7100. First a pic for reference:


This is from page 54 of the schematic for 820-4924-A.

At pin 1 of R7110 I have good voltage ~13v.
At pin 2 of R7110 I have nothing(!).
So U7100 has nothing on pin 3 (ACIN).

Fix:
Bad trace (open) from pin 2 of R7110 to pin 3 of U7100; repaired with jumper.

Tried:
Replaced 7110 (though it tested OK at 68kΩ.
Removed C7111-no change.
Pin 3 on U7100 is just under 100Ω to ground.
R7111 tested at 2.4kΩ in-circuit.
Re-flowed U7100, no adjacent pin shorts.


Ideas?

 

[2informaticos]

"Pin 3 on U7100 is just under 100Ω to ground."
Replace U7100...

 

[Holes Flow]

@2informaticos I just lifted it, and pin 3 is 0Ω to the pad (pin 29), but without a schematic, I don't know why it isn't also grounded to pin 22 ("GND")...

Right before I lifted it, I measured the full 16v across R7110, which I should have done yesterday. That told the story.

Really appreciate your confirmation!

 

[Holes Flow]

I was checking for pin shorts before powering up, and well, I must have had too much solder on U7100 pins 22, 23, 24, because they were all shorted together.

I sucked out enough without lifting to unshort pin 24 (CHGR_UGATE), but pins 22 (GND) and pin 23 (CHGR_PHASE) remain shorted together...



From the above, the only way I see pad 22 short to pad 23 is via Q7130.
When I checked, Q7130 pins 10 to 5/6/7 were shorted (MOSFET saturated).
As Q7130 is an Nch, I shorted the gate (pin 8) to 5/6/7 (GND), and checked to verify pin 10 to pins 5/6/7-BUT THE SHORT REMAINS (3.4Ω)!
...
Lifting Q7130 is hairy as L7130 is right next to it (already replaced it per above direction), but I guess I have to, unless you guys see another way for U7100 pin 23 to have 3.4Ω to ground?
...
Lifted Q7130, and pins 5/6/7 are 0Ω (as expected)...
But pin 10 (CHGR_PASE) is 3.1Ω to GND!
So Q7130 pin 10 isn't getting to GND via the MOSFET...
...
For reference, the shorted (3.1Ω to GND) pin/output becomes PPBUS_G3H (same schematic page, on right).
Now I'm chasing a shorted PPBUS_G3H, it seems!

 

[2informaticos]

Lift F7140 and check which side the short remains.

 

[Holes Flow]

Wrap-up. I sent this board to a shop in NYC, and there are too many layer shorts on the board to fix it. The power surge made even all the above not enough. I bought another LB, and have started a new thread about that board... Many thanks for those who helped all those hours I spent on that POS...