4P16S - monitor all 64 cells individually?

Theoretically when you put 4 cells in parallel, the voltages all are the same. However, under charge/discharge each cell wouldn't each cell potentially vary (pun intended)?
I am looking at buying an Orion BMS and one model would allow me to monitor all 64 cells. That might be good so I would know if one cell was going bad.

Does anyone here monitor every cell in a parallel setup? I think it is probably fine to in my case only monitor the 16 groups cells using a less expensive 16s BMS but am not sure.

If your cells are poorly matched, there might be voltage differences internally under load. If you have the cells connected together with bus bars, you will never be able to measure a voltage difference between them.

It is a complete waste of money to install separate measurement devices on cells connected in parallel.

If you want to know how a particular cell behaves I suggest you test each one individually and map its power output through the voltage range that way you can see if a cell has more energy at any specific voltage level so that they can all be matched if necessary.

to measure a cell , you need to isolate it from the others.
it is easy when cells are serial mounted, because they are naturally isolated from each other. it is just a string. no cell is sharing is 2 poles with another single cell.
measuring an element in the string will give you a reading for this element only.
but if you mount them in parallel, this is not possible anymore unles you use some way (a diode ? a mosfet ?) between each battery.
what you can do , is to measure temperature, because a failing cell will overheat due to the fact that its lower voltage will draw the current from the other cell connected to it.
this could lead to an general over charging of all elements in parallel since the reading of the set is dropped down by the faulty element.
that is why usually there is a fuse per element, so the dying element will (expectedly) draw so much current that its fuse will burn, returning the others good cells to normal operation (with a small loss in capacity.)
Unfortunately, the faulty cell could die slowly, leading to a big strain to the other good element, long enough to make them also defective.
That is why a temperature sensor per group of parallel cell is a must to ensure the early detection of a failure.
the other solution is put the serial string first with each a bms and parallel the bms output.
in case of problem in a string, the bms will disconnect the faulty string.
the problem with this is you could get a string that disconnect, and reconnect later (when cells cooled down for example) and you would end up with several group mounted in paralle with different charge level. This is a big no no, since at reconnection, the faulty string will receive all the current of the good strings. For this you would need an additional BMS who manage the other bms, avoiding such situation.
That is why even high end battery like in tesla cars, are just using fused cell and temperature reading, because it is simple and eficient.
and putting a lot of cells in parallel allows to loose a cell, while not loosing to much capacity. (if you get 30 cells at 2A in parallel, it make 60Amps. if you loose a cell, you still get 58Amps availalble and you loos only 3.2Vx2A=6.4W).
If you choose to put 20 cells in series, you get 20x3.2V=64V, but if you loose even just one cell, it is all the string that is dead and you loose 64vx2A=128W

If each cell monitoring is important, would it be better to go 16S4P instead and use 4 x BMS devices to measure each 16S pack? Once each pack was connected in parallel, would there not still be an element of self balancing between the 4 packs ???
Just getting to grips with this myself.

yes, if you put element in parallel, you need them to be all the same.
disconnecting a serial string from a parallele group is the best way to have the disconnected element to drift from the other.
the good elements will discharge or charge while the bad one is disconnected.
then you would need to make sure it come back to a proper value before reconnecting it.
ideally, the BMS would reset, and measure the voltage from the parallel group and reconnect only when the voltage inside the bms is matching the voltage outside the bms.
it probably exists in custom batteries.
but the easiest is to put P first (with fuse and thermal sensor) and then S with BMS.
Usually battery layout is more determined by the final voltage needed. Some needs high voltage, some can cope with low voltage but high Amps.
there are also industrial rules, where if you stay under 50V , you are in the low voltage rules, with different requirements than the high voltage group.

nosys70 said:

yes, if you put element in parallel, you need them to be all the same.
disconnecting a serial string from a parallele group is the best way to have the disconnected element to drift from the other.
the good elements will discharge or charge while the bad one is disconnected.
then you would need to make sure it come back to a proper value before reconnecting it.
ideally, the BMS would reset, and measure the voltage from the parallel group and reconnect only when the voltage inside the bms is matching the voltage outside the bms.
it probably exists in custom batteries.
but the easiest is to put P first (with fuse and thermal sensor) and then S with BMS.
Usually battery layout is more determined by the final voltage needed. Some needs high voltage, some can cope with low voltage but high Amps.
there are also industrial rules, where if you stay under 50V , you are in the low voltage rules, with different requirements than the high voltage group.

Click to expand...

If it is set up to disconnect the whole bank, not simply one defective pack, then if any cell goes out of whack, you know immediately to go investigate. If you only loose one string/pack out of four, the urgency of the situation might not be quite so apparent.

but it is a pity to disable the whole bank just because one pack is defective.
most of the case, giving up a bit and continuing to provide the service is better than stopping everything.
that is what make the Tesla battery clever. You can loose cells, and hardly see the difference.
If you really need to have a total control to all elements, then my suggestion would be to reduce the number of elements.
instead using 64 elements in a 4P16S, you better will find a cell that has 4 time the capacity and get rid of the P side of the problem.
That is why you find lead acid OPvZ that are 2V elements up to 1200 Amps.
or liFepo4 packs that are up to 600 Amps