Taves BMS

[johntaves]

I am manually balancing the cells. I have not written the automatic algorithm yet. I figured I would do it manually to make sure I knew what to do. Note, that I am not using a charge controller at all. The BMS disconnects or connects the solar panels to the battery. The only benefit to balancing is to get the max capacity from the whole pack.

The battery is being run to the cell limit (green), instead of stopping at a % (red) and you can see it is in the run-up mode because the SoC (blue) is grayed. Note PV is ON.

Cells 3, 4, 5, 6, and 7 have each been the top cell so they are all running their resistors for a day to bring them below the next highest. At 20ohms a day of burning drops about 4ah from the cell. I will keep repeating until I have seen all 8 cells at the top. Then I will repeat with maybe an 8 hour burn and after I have seen all 8 be the top cell, they should all be within 1ah of each other. That seems close enough.

 

[johntaves]

I got tired of wires so I designed a new cell board to fit the a cell and transmit the data to the controller via bluetooth. I have not gotten the boards, so I have not been able to program them and verify they work nicely. There is just one connector on the board, besides the bolt holes for the terminals. That connector is for attaching a thermistor that you can tape to the cell. However, the board measures its own temperature, so if you don't bother with the thermistor and instead rely on ambient temp, there's no wiring at all.

I had so much area on this board that I placed another 20ohm resistor to speed up the balance burn. If the resistors produce too much heat, they will be shut off until the temp drops.

 

[Roqm]

Will you also make 14-16s monitors to accommodate other pack designs?

 

[johntaves]

Will you also make 14-16s monitors to accommodate other pack designs?

It currently handles any Ns you please.

The wired cell monitors are individual devices that happen to be placed 4x onto one board. They transmit their data serially from one cell to the next. There are jumpers that enable you to bypass the individual cell monitors on the board, so you can wire any number in series. They are placed 4x onto one board to reduce the wiring. Unfortunately the series wired solution slows down exponentially with more cells because each measurement is transmitted Nx through the cells and there is Nx as much data to transmit, so N^2 time. The new bluetooth things shouldn't have that issue.

I don't know what limitations there will be with bluetooth. I assume there is no data transmission problem because it is fundamentally 2.4Ghz. I hope there are no software library or hardware limitations with the controller to have N devices connected. I'll find out when I get my boards and do the programming.