BMS algorithm question

[peters]

I'm designing a BMS for 15Ah LiFePO4 cells with shunt resistor balancing, and it's not clear how the BMS should manage the balancing period around the end of charging.
I have 2 options:

1.
When the voltage of any cell reaches 3.65V, then apply its shunt and switch off the charge current.
When the voltage decreases to 3.55V, then disconnect the shunt, and if all cells are below 3.55V, then switch on the charge current.

2.
When the voltage of any cell reaches 3.65V, then apply its shunt and reduce the charge current to a few 100mA that maintains the cell voltage. This current should be 3.65V/Rshunt, or a bit less.
When the voltage decreases to 3.55V, then disconnect the shunt. If all cells are below 3.55V, then increase the charge current to the original high level (e.g. 10A).

I think option 1 would balance the cells faster, but switching the cells at high current does not seem to be safe to me. Not sure, but maybe it can reduce the life of the cells.
Option 2 is is more "soft", because low current charges the cells which did not reach 3.65V yet. Despite the low current, the voltage of the cells with shunt ON would decrease to 3.55V after some time (can be few seconds or minutes), then high current switches on, but at a lower frequency than in option 1. The balancing probably takes more time in this case.
Any idea which one is better, or something else?

Thanks,
Peter

 

[dak664]

It sounds like you have BMS control over the charger current? How about turning shunts on at 3v55 and once any shunt is activated, limit charger current to 3v65/Rshunt. That eliminates small recharge cycles over the 3v55 to 3v65 range, as shunted cells will gradually rise and hold at 3v65. When all are in shunt, turn off the charger.

Many BMSs activate the shunt at 3v65 and rely on the charger taper to limit the overall current; if enough cells are low one or more high cells can reach 3v9, at which point the BMS disconnects from the charger to let them bleed down. If way out of balance, the charger blasts on after a while and the cycle repeats. Can't be good for the high cells, although damaging them a bit might help to maintain long-term balance

Seems to me the fastest way to get a top balance is to activate all shunts when charging starts, then turn off the shunt on the lowest cell until it comes up to the next lowest, then turn that off too, etc. until all shunts are off. Ideally then the charger never has to taper and can run full blast until all cells are at the maximum voltage. That would probably require central collection of all the voltages in a microcontroller.

 

[peters]

Good ideas, thanks.
I'd like to get rid off the small recharge cycles and the charging to 3.9V, and I'm planning to control the charger current from the BMS if necessary for this. And also because if the charging terminates when the cells are 3.65V at 10A, then it is only the end of CC part, only 80-90% charged state. CV is not really possible with cells in series, so need to reduce the current and limit the cell voltages with the shunts.