whatever said:
you could set different variables for each series group, lets say one group has a bad cell, you might want to take it easy on that group and set lvc higher, all sorts of situations could be dealt with..
A "group" means cells in parallel, at 1S voltage.
A string is cells (or groups) in series, higher voltages.
Are you talking here about multiple higher-voltage modules that get combined into a bigger pack
each with their own separate BMS?
> You say its not possible to monitor parallel cells individually they have to be atomised, yes that is the challenge, how to monitor them without dismantling the pack.
No one is trying to do that within a paralleled group. Again, it is only possible to treat the group as if it were a bigger cell.
> Lets say each cell has a tiny mosfet in the top of the cell ( assume 18650 size), it would be possible to turn off all but one cell in a large parallel group of cells, and monitor that cells condition, or test that cells condition.
Sure, it is easy to isolate a cell from the rest of its group.
But you don't have any valid voltage information to spot a problem, trigger that happening.
> So its is possible to monitor cells in parallel
You have only spoken of the isolation mechanism, again, there is no capability of **monitoring** intra-group cell
> how to build such a thing though is a different story, as each mosfet would need a separate connection to turn its mosfet on/off.
It is not that far fetched to concieve of such a method being built. It might be possible to switch the mosfets on/off via some wireless method in near future.
Same again.
I thought we were discussing reality as it is, not dreaming of "what is possible." The ability to do the latter usefully requires a good grasp of the former.
wired vs wireless is an irrelevant separate topic.
> It is one of the weak points in battery packs, how to monitor cells in parallel.
Easily overcome, b going to higher capacity cells, you get many advantages, and not being "blind" to intra-group cell failures is a big one.
> Isolating a bad cell is possible using overtemperature on/off valves in the top of the cells ( fairly common), but a bad cell that doesn't get hot is going to be a bit of a problem. It will of course lower capacity of that group, but if it shorts out ( which can happen) it can cause fuses to blow
The fusible link per cell idea that Tesla has used for years is one approach, based on current.
But a temp sensor per cell?
Remember KISS complexity introduces new problems, reduces reliability.