820-00840: Won't boot up unless SMC reset is performed

[macRepairAmb]

Have a 820-00840 board that originally came in for power failure. Found corrosion damage around R5420. Had to repair wire traces from pin 3 and 4 to pin 5 and 4 on U5420. The board worked at the point. But got returned a few months after with a new issue. Now the board requires smc reset before it will boot up. I would have to let is remained powered off for about an hour to be able to recreate the no power issue, at which point, i have to reperform smc reset to be able to power it again.

I installed the board in another frame to confirm it wasn't the battery and i also replace the SMC with the same result.

I can't recreate the issue unless the battery is connected. Once the board is out of the frame, it comes on all the time.

Any suggestion on where to start?

 

[2informaticos]

Test if starts in SMC bypass in that moment.
Check R/C5426.

Also check for corrosion around U8510.
Be sure SMC_RESET_L and BUF_SMC_RESET_L are steady high.

Inspect J6100 for corrosion too.

 

[macRepairAmb]

There is no R5426 or C5426 on the board. But to perform the SMC bypass, doesn't it require the battery to be unplugged? The board boots up without issue all the time when the battery is unplugged.

And yes SMC_RESET_L and BUF_SMC_RESET_L are steady high at 3.41v during the time when the unit won't boot. They both go to zero when I perform the smc reset as expected.

No corrosion on J6100 and pulled U8510, no corrosion. Went ahead and replaced U8510 just in case. Same result.

One thing i did notice, not sure if this is something to be concerned about, but every line that runs off PPBUS_G3H has a diode mode reading of 0.381 compared to two know good boards reading of 0.469.

 

[2informaticos]

Check SMC_LSOC_RST level at that moment.
Also disconnect keyboard and test if board start with SW9988/89 at same moment.

You cannot expect the exact same reading on different boards; even when are the same model.

 

[macRepairAmb]

SMC_LSOC_RST is 0v both before and after smc reset.

SW9988/89 don't work, its a floating connection. These version board are missing R9988/C9988. Confirmed on two other boards.

 

[2informaticos]

"SW9988/89 don't work, its a floating connection."
I don't exactly understand what you mean.
Just need to use one of them instead of power button.

C9988 is marked as "NO STUFF", but is very strange to miss R9988.
However, just trigger SMC_ONOFF_L to ground in this case.

 

[sekidata]

I can confirm that SW9988/89 is a floating connection, with missing R9988 on the two 820-00840 boards I have looked at.