820-00165 doesn't power on with magsafe

[jadao]

This logicboard had water damage causing backlight issue, crap near SMC and ISL6259 but no heavy damaged components.
Cleaned everything reflowed SMC to make sure.

At this point, it boots all ok with battery, it charges when magsafe is plugged in. AHT test doesn't show any error.
It will start as expected automatically with the charger when the battery have been previously disconnected.
The problem is the second time when I power off the board with the charger still connected it will not power on after pressing the power button.

I have to disconnect the magsafe and it will power on from the battery when I press power button, if I put back the magsafe it will charge fine but again if I power off and press again the power button it won't start.

This looked to me initially a piece of cake but now I am completely stumped!

PPBUS is 8,56V without battery charged
3,42V PP3V42 present
SMC_reset_L is a nice 3.3V
SMC_ONOFF_L is also 3,4V
I can see SMC_ONOFF_L going low to 0,4V when I press power button.
PM_PWRBTN_L is also 3.4V and goes low when I press power button.
Resistance between pins 17-18 is 4 ohms and 24ohms between 27-28.
R7005 and R7006 are respectively 10 and 5 ohms I get the proper voltage on D7005 output either 8.56V either 16.5V

I replaced D7005, replaced U7100, replaced U5110, replaced Q7180 and Q7130, reballed original SMC then replaced it twice and still have the same issue...
Tried different keyboard/trackpad original 85w magsafe, I/O board & cable.
Checked again and compared expected values coming in and out of U5110 and they check fine.

The only clue I can tell is that SMC_reset_L is sensitive as soon as I probe it in voltage mode it will trigger the board and it starts, I have no more logical explanation for this as I reballed and changed twice the smc, changed the ISL and checked all traces for SMC_RESET_L. I will try an ultrasonic in case some crap under the cpu.

What am I missing or what can I check? I have no more idea and am exhausted on this one.
Thanks

 

[2informaticos]

Test the board alone on the desk, without LIO flex, nor trackpad, SSD, or airport card.
Use R5115/16 pads to trigger SMC_ONOFF_L, instead of power button.
Also keep the battery disconnected for the moment.

 

[jadao]

Hello 2informaticos, I did an ultrasonic but no change.
I tested it without flex and nothing else plugged in, only original magsafe 85w and it starts as expected without the green line.
Then I trigger SMC_ONOFF_L to shut it down and power it on: it will start sometimes but not consistently...
sometimes it will simply not start even if SMC_ONOFF_L is going low.

 

[2informaticos]

Post exact voltage for 3V42_G3H, PPBUS_G3H and SMC_ONOFF_L, when doesn't start.
Check if always starts in SMC bypass mode.

 

[jadao]

Ok so when it doesn't start: PP3V42_G3H is 3.422V, SMC_ONOFF_L is 3.419V and PPBUS_G3H is 8.59V.
However, SMC bypass mode simply doesn't work! I missed it since the board boots normally and works I didn't try it!

I tried also smc bypass with LIO flex to monitor the magsafe LED, the led stay off but no fan spin.
The magsafe led is on, green then orange in normal mode...

 

[2informaticos]

Be sure PM_RSMRST_L is steady high.
Replace R5100; try 33-47K, instead of 100K.
If you detect drop pulse at PM_PWRBTN_L, when trigger SMC_ONOFF_L, then check for pulse at all PM_SLP_Sx_L signals at same time.

Check PM_BATLOW_L, PM_DSW_PWRGD, 3V3_SUS.

Remember, all the tests when board doesn't want to start...

 

[jadao]

PM_RSMRST_L is steady high at 3.3V
Replaced R5100 with a 35k but nothing changed.
Yes I detect drop pulse at PM_PWRBTN_L, when SMC_ONOFF_L is triggered, I check for PM_SLP_S5 and found it missing 0V.

Checked PM_BATLOW_L 3.32V , PM_DSW_PWRGD 3.41V and 3V3_SUS 3.3V.

Yes all the tests done when board doesn't want to start.

Thanks

 

[2informaticos]

Looks like PCH issue.
No pulse at all at PM_SLP_S4_L, when trigger SMC_ONOFF_L?

Check PCH_DSWVRMEN and the 4 RTC signals.

You may need to put quality flux and heat PCH/CPU area; trying to burn eventually accumulated crap.

 

[jadao]

No I had no pulse at PM_SLP_S4_L when SMC_ONOFF_L was triggered.

PCH_DSWVRMEN present and the 4 RTC signal are present (checked them in the beginning but didn’t reported them).

I heated the PCH with flux but no change, probably still crap or an issue.

I returned the machine to the customer, he accepted it as-is. If the issue happens he will only have to unplug the magsafe and it will turn on when he press the power button.

thank you very much 2informaticos