One more thing. I see a 40mV drop in Cell 1 when 100mA balancing turns on. This seems too high, right? 40mV/100mA = 400mohm. Is my cell 1 damaged? Should I seek warranty replacement? Cell 1 still achieved 210Ahr, per the BMS full charg
You identified the resistance of the balancing wiring harness, not the cell. It’s pulling full capacity, meeting advertised specifications, so no you shouldn’t seek warranty replacement, at least not yet…you are covered for 10 years so if it ever does turn into a bigger issue, we will take care of you.
I will get SOK engineers feedback on your other comments.
Thanks, I also realized this last night after I sent the message. It's because the voltage measurement and the 33mohm balance share the same wires to the cell. Ideally, the voltage should be Kelvin sensed with its own wires that don't ever carry current. Why? Two obvious reasons: 1) My Cell 1 is getting overcharged to 3.74V. Because the battery is balancing, it is measuring only 3.700V when the protection MOS opens, but then the balance stops and the cell is actually at 3.74V. (See attached photo.) This is aging Cell 1 in my pack at an accelerated rate. I suppose once the cells are balanced (eventually), this problem may go away. 2) As the cells try to balance, the balance will not work as hard because it will oscillate on/off. When trying to balance, it turns on the 33ohm, then the next voltage measurement is 40mV lower, so it thinks it doesn't need balancing, so it turns the 33ohm off, then the next voltage measurement shows it back 40mV higher and still needs balancing, so it turns the 33ohm back on. IMHO, the 33ohm balance is already pretty weak for a 206Ah battery, but now it will be ON only half of the time and only at the very end of charging. Effectively, I think it is just a 100mA*50% = 50mA effective balancer, which is only 50mA/206A = 0.00024C. And it can only work in the last 1% of charging due to the 1A charging limit and the 3.375V minimum cell voltage. So, 0.00024C * 1% = 0.0000024C effective correction. Is it enough to keep up with aging mismatch? (rhetorical question) My suggestion to SOK is to use two separate wires for balancing and for voltage measurement (true Kelvin sensing). Or as a cheaper solution, keep the existing single wires, but add 40mV in software to the measured voltage whenever the balancing is turned on. (They should actually characterize the 40mV themselves and for each wire in the harness.) That will effectively eliminate the on/off toggling. Then, try to make the balancing run for a longer time. For example, use the same criteria (1A minimum charging, 3.375V minimum cell voltage, ???Vdelta minimum mismatch) to turn ON the balancing, but use different criteria to turn it OFF. For example, reduce the 1A minimum charging threshold as far as possible, even maybe 0A. Also reduce the Vdelta mismatch. Can they add a software timer to extend the balancing beyond the end of charging? That would help tremendously. Obviously, active balancing would be better, but it is more complicated and more expensive. I think with a few changes in the software, they can improve the passive balancing.