The only problem then it’s that my banks are detachable and fold into place like drawers. I don’t know how i would be able to make a connection between them, and disconnect if i pull one for maintenance, and then it will stop working for the bank still operative?
I saw that the board has two sets of tx/rx connectors. I thought the banks could be connected to the controller independently.
I check for TCA9534A replacement. TCA9554 will be best but I2c adress will change and match with other chipset, so we need to change pinot to fix other I2c adress than 0x20 with A1-A3 adress and modify source code.
I just checked by taking out string 3 and make string 4 string 3 and time out issue remained with the one string… which newly is string 3 instead of string 4.
I replaced module as well as bus cabling but nothing changed…
And i can confirm, it is 16 modules per run which are checked.
Regards
Karl
can these buttons be added to the v4.2?
what about canbus is it working?
Yes the buttons can be added - I’ll work on instructions for that soon.
CANBUS hardware works fine, but the controller code doesn’t use it at the moment.
Trying to wait patiently for the Canbus support. I have the new Victron 10kva inverter on order and would love to have this play nicely with Venus OS.
Assume you calibrated the modules to ensure they were reporting the correct voltage?
Have you checked with a good quality multimeter what the readings are for the two low cells?
Actually these two cells show higher values as they really are… approx 150mV.
It seems that they don’t accept calibration.
I have a good uni-t multimeter which seems quite accurate.
What do you mean by not accepting calibration?
Well, i measure voltage, calibrate module and shortly after it again shows wrong voltage…
I heard from a friend that modules don’t accept calibration when they are balancing. I have to check whether this is also the Case in my situation…
I don’t think there is a reason why they shouldn’t accept calibration during balance.
Ok I’ll check that topic shortly
For ordering LEAF modules, on Github there is something missing in the gerber directory:
There is no “JLC” named BOM and CPL FIle in the Github Repository.
For the “normal” modules there are extra JLCPCB BOM and CPL files.
Finally i could calibrate, but i don’t know why it didn’t work before. Curious… 
Today testing difference of V4.21 to V4.40 module. Cool that we can mix it on the new controller, and getting an Info.
The main difference is the balancing power.
4.40 is starting with 1Amp balance power, when reaching max temperature goes to PWM mode and then jumps 200-900mA
4.21 is starting with 800mA, but holds this much longer (the large amount of resistors results in bigger surface too cool I think)
Then when reaching temperature it jumps between 300-600mA
I will do a long balance test later to look if old or new is the stronger one in long term.
4,40 took 8 minutes to balance a 300Ah Lifepo from 3,55 to 3,45
4,21 took 11 minutes.
Another difference is.
4.40 is hammering the voltage down to the set value (3,45V in my case), and it stands there fixed, maybe one or two times after-balance balancing.
4.21 is doing it also, but still needs “minimal” balance shots of a few seconds.
(after 40 minutes …) and still has 3,455 and not 3,450 constant.
The mysteries of life Karl … 
Hello, greetings from Sri Lanka. I built a esp32 controller board and some modules for my LiFePO4 batteries. I half built the controller module from JLC and ordered all the extended parts from LCSC and soldered them myself.
I am wondering if there is any way to send MQTT data to a cloud broker ( I am trying to use hivemq ) with a TLS/SSL certificate. Has any one done this before?
The thread below may help you:
Looking great.
ESP32 does support TLS, but it’s not written into the code. Can you raise a new issue on GitHub as a feature request and I’ll see what I can do.
