openinverter forum

More questions as i aim to get the car into regular use. And this means treating the cells right.

I've got for starters 9kw bmw G30 pack, and i've got some code that can read each cell voltage using snippets of code from the web. There are some comments against bit settings about balancing but i'm not clear on how that works as yet.

In terms of balancing the cells, i think i'm lacking understanding. I understand the requirement but not how in practice this is implemented.

So some questions:-

is balance achieved by managing the incoming charge to each cell, or by discharging high cells to meet the lowest once they have been charged and have settled ? Then once balanced to the lowest, one could top up ?

Has anyone implemented this using the BMW G30 modules using their own code ?

Here is the SimpBMS code that works with these cells and contains the cell balancing commands.

https://github.com/Tom-evnut/BMWPhevBMS

There is quite a bit of checksums required to make this work.

As for balancing itself:
1. BMW implements a passive bleed resistor system.
-So current is bypassed by having resistors in parallel to cells during charge.
-When no current is flowing current is then instead drained from the cell in parallel.

Best to balance above the "flat" of the discharge curve, I usually recommend only starting above 3.85-3.9V. I would then say balancing is completed when your are below 10mV high to low.

thanks tom -that is the code i have been using to get my own code reading all 96 cells. I note that balance voltage is 3.9 with a hysteresis of 0.04 volts

i can't quite work out what the code is doing because its not clear how the balancing is structure. I imagined my code would identify the high cells and instruct the relevant CMU to discharge each cell, but it looks as if the CMU handles the balancing once you tell it the voltage to balance to ?

If balancing (line 3818) it appears to send the lowest voltage +5 (is it 5 because this just 1 more than the hysteresis value) ) to all the CMUs. Does this instruct them to balance their respective pack to this voltage, and then one just monitors it or it will report when complete ?

Yes it sends each CMU a "balance to voltage" it is set to lowest pack cell + hysterisch as to keep measurement errors from causing oscillations in balancing.

Each OEM has its own way of doing this. BMW and Mitsubishi use voltage targets.

Others use; which cell to discharge and some also require a discharge duration to be included.

Can I use 5 separate BMS/balancers to work in parallel?
For my solar system I am using a 24 battery 4 Amps active balancer- NEEY V4.
https://www.aliexpress.us/item/32568052 ... Cl6Exqq858
I would need 5 x 24 cell BMS/balancers to work in parallel. I was thinking in sharing 2 wires on each BMS with another one so for example BMS-3 knows the cell voltage of 2 cells managed by the BMS2- and 2 cells balanced by BMS-4.
Anybody doing that?

Proton wrote: ↑Tue Jun 20, 2023 8:22 pm Can I use 5 separate BMS/balancers to work in parallel?
For my solar system I am using a 24 battery 4 Amps active balancer- NEEY V4.
https://www.aliexpress.us/item/32568052 ... Cl6Exqq858
I would need 5 x 24 cell BMS/balancers to work in parallel. I was thinking in sharing 2 wires on each BMS with another one so for example BMS-3 knows the cell voltage of 2 cells managed by the BMS2- and 2 cells balanced by BMS-4.
Anybody doing that?

It will not work for cells in parallel!
I am using a single 24S JK BMS module with my LiPo cells configured into 15S 6P. This means 6 packets of 15S cells are connected in parallel by end terminals so about 55V nominal. Then to really keep cells in parallel i connected each cell with its counterpart in other 15S packs. This creates a wire nest but every cell is connected in the same way and i can simply use one BMS module.
If you really need to use 4x BMS you want to connect them separately.
Try to draw out your cell connection on a sheet of paper and make combinations...

For my solar I also have 6p16s I am trying to use them for 112s to get 358V.
was thinking first balancer connect to bateries 1 to 22, second balancer 21 to 42, 3rd balancer 41 to 63 and so on so the balancers overlap and they can measure the voltages for 1-2 cells controlled by other blancers.
Probably will not work well but wanted to see if anybody already tried that.