Yep Srp111 's data was unbalanced too with a 442 voltage with a 93.5% SoC
Link
Well spotted I'd forgotten about that one.
There's something in the data!
Ignore the 455V cars for now, subset of data sorted by SoC:
Srp111's data was with 7Amps of current, so I don't actually think we can use the Max/Min cell voltages or total volts, as we know the volts reduce as current is pulled.
I think there could well be something in it being based on the max cell voltage & not the total voltage:-
The charge does need to stop when the first cell gets to the preset maximum voltage. It can't use the total voltage to determine when the stop charging, as with a very unbalanced set of cells, it could end up putting a cell voltage way higher than in spec.
450v / 108 happens to = 4.1666667
So it makes sense that it stops charging as soon as any cell gets to that value.
Balancing then does it's stuff and alters cells a bit, but not too much as it's small currents and is limited to 10 hours (maybe this is why the balance is time limited to ensure a cell couldn't accidentally end up getting too high)
The SoC is consistent with the max cell voltage in the data, and not consistent with the total voltage.