I have been reading lots about the large capacity LiFePO4 cells and commercial off the shelf BMS units the last few days. It appears the BMS units do use an internal Coulomb counter circuit to calculate remaining capacity.
Regarding the tendency for separate LiFePO4 banks of differing capacity staying balanced, it is my understanding this can be attributed to three reasons.
First is the inherent discharge voltage characteristics of a LiFePo4 battery where the discharge voltage remains almost constant from 100% capacity to roughly 30% capacity, varying only about .2V
Second reason is the charge balancing function performed by the BMS ensures full charging of each cell, and during discharge the active balancing function of the BMS compensates the weaker cells with available voltage from stronger cells in the bank to help prevent imbalance and subsequent output shutoff with usable capacity remaining in the stronger cells.
Third is simply the available depth of discharge, or usable capacity inherent to the LifePO4 chemistry. In a less than ideal application where two different capacity battery banks are paralleled, the smaller capacity battery may have sufficient available/usable power to support large loads for a useful period of time.
Of course you probably already know this, but I figured I would share a little of what I learned while reading up on how to assemble a LiFePO4 battery bank similar to the one you put together.
Here is some information that may be useful.
https://www.electronicdesign.com/power- ... nt-systemshttps://www.solacity.com/how-to-keep-li ... ies-happy/