Ah man... you beat me to it.
I admit that I did not search.
Only read the first page... but as to parts count and power (you may have found this)
The Attiny let you run down to 1.8V and 1Mhz for super low power
Keeping the oscillator internal drops three parts at the cost of some power
I was considering dumping shunt heat across the resistor, regulator, and the chip... by using an IO leg
Interesting that you did a Resistor Zener for the supply. I was just grinding my mind looking for an LDO that would give good Q current, run down below 2V, low dropout...
I just took a drive with the ol'lady and was considering parts count and cost... I said on the way back that I may need to pivot.
My parts count is already 10 parts per channel ... x 12 = 120!! parts
Since this has already been done I am going to pivot back to a Linear chip.
I will try to make it lower parts count this time by using resistor, capacitor, and Zener arrays - layout becomes wicked hell... but population goes much faster
Sigh... LTC6802... 6803... 6804... and beyond.
Here we go again.
I lost all my original layout files and source... but luckly I shared the source with a friend in the UK so I think I can at least get my state machine back.
This time around instead of sampling at like... 20ms... I am going to go SUPER SLOW.
BMS is only going to act once every 500mS (it works... do the math...) and get the power way down.
K - thanks for the kick in the butt.
Increasing battery voltage and controller current limit will result in a non linear experience