Search Results
5/18/2025, 11:16:08 PM
>>2917867
>wouldn't it be best to have the charge controller able to measure the voltage of each cell individually, no matter how many cells you have, how many are in series, how many are in parallel, etc?
For parallel groups of cells, their voltages will all be the same. By definition, they must be; they're all the same node, and a single electrical node cannot have more than one voltage at a given time. In reality, there is a slight difference due to the non-ideal nature of real conductors, but it's very small.
For series strings of cells (or groups of cells), your practical limitation is mostly voltage. Your number of choices in ADCs and/or multiplexers becomes greatly limited above about 30V, never even mind the high-voltage packs often seen in vehicles. At that point, the obvious move is to simply have individual modules that can track the state of, say, 4-12 cell groups. Those then communicate with a central controller that's galvanically isolated from them, usually via optocouplers, though some us inductive coupling (I think the Leaf's BMS does this).
I actually spent a lot of time designing a BMS (which will never see the light of day, now) that could do 18-cell strings and be linked with itself to create arbitrarily high voltage packs. This was back when the Rona was in full swing and everyone and their mother was backordered for 8 months or more on their BMS chips. I discovered one of the major problems with the "one module per cell" approach was ensuring consistent voltage measurements between cell groups. The only way to do it is to have very precise references, which balloons the cost of the system when you need one for every cell. I ended up creating my own analogue mutiplexer using a flying capacitor, instead. It should have worked, but it adds a shitton to the BoM. After the chip shortage was over, it was just straight-up worse than using one of TIs integrated chips.
>wouldn't it be best to have the charge controller able to measure the voltage of each cell individually, no matter how many cells you have, how many are in series, how many are in parallel, etc?
For parallel groups of cells, their voltages will all be the same. By definition, they must be; they're all the same node, and a single electrical node cannot have more than one voltage at a given time. In reality, there is a slight difference due to the non-ideal nature of real conductors, but it's very small.
For series strings of cells (or groups of cells), your practical limitation is mostly voltage. Your number of choices in ADCs and/or multiplexers becomes greatly limited above about 30V, never even mind the high-voltage packs often seen in vehicles. At that point, the obvious move is to simply have individual modules that can track the state of, say, 4-12 cell groups. Those then communicate with a central controller that's galvanically isolated from them, usually via optocouplers, though some us inductive coupling (I think the Leaf's BMS does this).
I actually spent a lot of time designing a BMS (which will never see the light of day, now) that could do 18-cell strings and be linked with itself to create arbitrarily high voltage packs. This was back when the Rona was in full swing and everyone and their mother was backordered for 8 months or more on their BMS chips. I discovered one of the major problems with the "one module per cell" approach was ensuring consistent voltage measurements between cell groups. The only way to do it is to have very precise references, which balloons the cost of the system when you need one for every cell. I ended up creating my own analogue mutiplexer using a flying capacitor, instead. It should have worked, but it adds a shitton to the BoM. After the chip shortage was over, it was just straight-up worse than using one of TIs integrated chips.
Page 1