I hope to be able to improve the Vectux BMS by re-arranging the cells so that the effect of low capacity cells is more obvious to the BMS.
The 102 NiMH cells are arranged in series, with two Voltage probes taking measurements between cells 27 / 28 and 75 / 76.
There is no way (that I can see) in which the stock BMS can actually shuttle current between cells.
It only uses the measured voltages to determine if there is an imbalance, in which case it might limit power delivery during riding or change to trickle charging during charging.
If low capacity cells are distributed in equal proportions through the measured sub-sets of the whole string, then the BMS will not notice an imbalance even when some cells are unable to accept further charge or are being reverse charged.
But, and that's where my idea comes in, if all the low capacity cells were in one of the measured sub-sets, then there would be a clear voltage difference once these cells are approaching full, or empty, compared to the other sub-sets.
If this works, then it would be similar to the gear box modification which reduces the running noise for next to no costs.
But there might be unexpected effects, or the performance might be reduced so much that it is impracticable.
The schematic below shows the 102 cells (in their 12 modules, in two batteries, with a 2-pole connector between them), as well as the points where the voltages V1 and V2 (and GND) are measured.
The red points mark cells with less than 19Ah capacity at 20A (1.1V cutoff) before reconditioning.
The black points mark cells which show swelling.
All cells with less than 19Ah initial capacity show swelling and the vast majority of them was located in the lowest of the three layers.
The harness for the stock BMS is shown below. (click thumbnails to enlarge)
It contains 6 temperature probes and the V1, V2 and GND cables. There is a second similar harness without voltage sensors for the front battery.
So far I have not been able to devise a way to re-assemble individual cells once the modules of 9 or 8 cells have been taken apart. Therefore I might have to be content with re-shuffling modules rather than cells.
Can you think of any potential problems that this approach / modification might cause?
The new cell distribution would look like this:
Any help in evaluating this design would be much appreciated!