My system is using a JK BMS on Li NMC cells. They act very different to LiFePO4 cells, so my current settings would not be good.
But I am always looking into options for when I need to replace the cells, and LiFePO4 is a better battery for safe home energy storage. So I have been looking at how they work. Looking at the charge curve, the upper knee starts at about 3.4 volts. So that is the area where you should start the cell balancing. The idea is to only "top balance" the pack. As any cell goes above 3.4 volts, the active balancer will pull power from the high cell and push it to the lowest voltage cell in the pack. Ideally, the result will be all of the cells going above 3.5 volts at the exact same time as they are brought to the same state of charge. For that to happen, the cells would need to be closely matched an perfectly balanced. In the real world, the cell voltages will spread more as they go above 3.5 volts. 2 amps of balance current is good and my JK is keeping my 360 amp hours of Li NMC cells all balanced within 0.003 volts with no problem at all. But if the pack is being charged at 20 amps, one cell being just a few % higher state of charge can run away as it passes 3.5 volts. Pulling 2 amps won't slow it much as there will still be 18 amps charging that cell. And the way the JK balance works, it can only pull that 2 amps a bit less than 1/2 of the time. For this reason, it may take several slow charge cycles at less than 4 amps to give the balancer time to pull all of the cells to the same 100% full charge state.
With LiFePO4 cells, you want to stop balancing when the cells fall below 3.4 volts. Once the cell voltages are into the flat region, it won't do anything anyways. But, if you run the pack down to where any cells start hitting the bottom knee, the balance might then try to pull power from a stronger cell to help keep a weak cell up. But this is wrong as it will un-balance the cells very quickly. If the cells are perfectly matched, this won't happen, but any difference in cell capacity will trigger this bad behavior. Lets say these are 100 amp hour cells, but one cell is only 95 amp hours. When the pack is fully charged and top balanced, all cells will be at 100% state of charge. When you discharge 30 amp hours out of the pack, the good cells are now at 70% state of charge, but the weaker 95 amp hour cell is actually down to 68%. Rung the pack down to just 30% remaining on the good cells, this weak cell could be as low as 25% left. If the active balance was still running, it would be pulling power from the stronger cells to boost this weak one back up. While this might make it run a tiny bit longer, it is messing up the balance. If the cells were then balanced to the same voltage (state of charge) and you try to charge this back up, the weak cell get's pushed up to the same 30% SoC, but as the pack charges, the weak cell is actually charged to a higher state and will once again hit the top knee first and run away. This is due to the cell having extra charge pushed in by the balancer. If the balancer is left off, yes, the weakest cell will limit the minimum charge state by hitting the low knee first, but when charging back up, every cell will take in the same amp hours and hit full charge at the same time again. Limit the discharge to keep the weakest cell over 20% SoC and all you ever need to do is a light top balance to keep them in sync. Turning on balance when any cells goes above 3.4 volts or even a bit higher will work great.
I may change my settings to limit the balancing to the very top as well, even in my NMC cells, but it is not as critical. They are so well matched, they cells don't spread more than 0.008 volts ever any more, my balancer has not turned on in over a year since the cells got top balanced up. But I also never run them below 45% charge either. But I might in the next few days. I just got a notice that the electric company may shut off the power here due to high winds and a wild fire danger situation. So my house me be running off grid for 2 to 3 days.