820-3787 - Chime loop wich battery

[rvanpruissen]

This board works fully (in limp mode / 800MHz) when the battery is disconnected. When the battery is connected it chimes repeatedly, so at the end of the chime, it chimes again. A different battery produces the same result. Replaced ISL6259, didn't help. Any thoughts? Could this be GPU related?

 

[2informaticos]

Does it boot in SMC bypass with both charger and battery connected?
Check exact ohm resistance between pins 17/18 and 27/28 of ISL6259.

 

[rvanpruissen]

Assuming you mean safe mode (holding shift), same behaviour. Both current sense circuits measured fine when I replaced the ISL.

 

[2informaticos]

NO, I exactly mean SMC bypass mode; explained many times on the forum how to do it.
Google can help you, in case forum search not...

 

[rvanpruissen]

It won't let me do that, it starts instantly when I press the power button. Only thing I have left to try is corrupted BIOS maybe? Any other ideas?

 

[2informaticos]

"Check exact ohm resistance between pins 17/18 and 27/28 of ISL6259."

Did you try original battery, even old?
Do not try new battery from same batch...

 

[rvanpruissen]

17/18: 3ohms
27/28: 20,9 ohms

Known good battery tested, same behaviour.

 

[2informaticos]

Does the board start on battery only (no charger)?

 

[rvanpruissen]

Yes, it starts on only the (known good) battery. Same behaviour.

 

[2informaticos]

What is the fault history?

 

[rvanpruissen]

According to the customer it just started doing this, no specific history.

 

[2informaticos]

Monitor SMC_DCIN_ISENSE and SMC_CHGR_BMON_ISENSE levels.
Any sign of corrosion around SMC?

 

[rvanpruissen]

Ok, I started testing for the suggestions above, but noticed the heatsink got hot real quick. Turns out phase 1 of the GPU is pulling ±35A from the start. GPU problem? I'm confused as to why that VR would even start at this point during POST/boot.

 

[2informaticos]

GPU should have the same behaviour for start sequence, w/ or w/o battery.
Check the mentioned sensors level and post PPBUS_G3H voltage with both battery and charger connected.
Is the battery charging?

 

[rvanpruissen]

Measurement	with batt connected	w/o batt connected
PPBUS_G3H	12.2 V rising to 12.7+ V	12.57 V steady
SMC_DCIN_ISENSE	1.62 V	0.4 V
SMC_DCIN_VSENSE	2.89 V	2.98 V
SMC_CHGR_BMON_ISENSE	0.05 V	0.05 V
Difference over R9398 (GPU Phase 2)	0 V >> 0 A draw	0 V
Difference over R9399 (GPU Phase 1)	2.6 mV >> 3.5 A @0.85V >> ±3.0 W (Miscalculated earlier)	2.6 mV
Current draw from power supply	2.25 A	1.2 A
Difference over R7150 (BMON)	-13 mV (w/o charger) >> -2.6 A 8.5 mV (with charger) >> 1.7 A	0 mV

The BMON isense does not seem to change when disconnecting the battery. I've replaced the cap and 3 resistors for battery Isensing, no change. R5423 (between BMON and SMC) measures fine and is not shorted to ground.

 

[2informaticos]

With 1.7-2.6A through R7150, CHGR_BMON should be higher than 0.05V.
ISL6259 gain for BMON is 36X.
2.6 x 0.005 x 36 = 0.468V which should get there.
Your problem could be ISL6259, or SMC...

 

[rvanpruissen]

So, I've been struggling with this one over the last few days.
* Replaced ISL > No change
* Replaced SMC > Didn't work (possibly faulty SMC)
* Replaced SMC with 3662 version and clean BIOS > Same behaviour (this one is just for testing, as I don't have any more 3787 donors)

I'm guessing some line gets overloaded, but not sure how to check this...