Introduction

BerryBMS is a small project to gather and show Schneider Conext and JK BMS data. It is meant to show what's relevant in a single page. Here's how the dashboard looks like with four JK BMS:

Installation

Installation instructions are tailored towards Debian 12 running on a Raspberry Pi. First you should create a normal user under which everything will be run.

First, install Debian-included packages:

apt-get install mosquitto python3

Then, proceed with the following instructions to setup the local Python environment as it requires dependancies not found in Debian 12 or that are way too outdated:

cd ~/berrybms
mkdir venv
python3 -m venv venv
source venv/bin/activate
pip3 install pymodbus==3.8.2
pip3 install pyyaml
pip3 install paho-mqtt
pip3 install pyserial
pip3 install python-can

If you want to have the Web GUI that uses Dash/Plotly:

pip3 install dash
pip3 install dash-bootstrap-components
pip3 install dash-bootstrap-templates
pip3 install https://github.com/plotly/dash-daq/archive/refs/heads/master.zip
pip3 install flask-mqtt

Connection to JK BMS

First, you need at least one USB-to-RS485 adapter. On the adapter, connect RJ-45 pin 6 (orange) to A+ and pin 3 (white orange) to B-.

If you want to use BerryBMS in sniffer mode, you can simply connect your USB-to-RS485 adapter directly in one of the available RS485-2 port of either BMS (master or slave). For these ports, the 015 - UART Protocol 015 protocol will be used on the master (DIP switch set to 0) and the 001 - JK BMS RS485 Modbus 1.0 protocol will be used on all slaves.

If you want to use BerryBMS in polling mode, you will have to connect your USB-to-RS485 adapter (with as many channels as BMS you have) to each BMS's RS485-1 port. The 001 - JK BMS RS485 Modbus 1.0 will be used all RS485-1 ports.

You can use sniffer mode and polling mode simultaneously.

Connection to Conext

If you have a Conext InsightHome, you can configure BerryBMS in polling mode and it will use Modbus over TCP to pull information from the Conext InsightHome.

If you want to use BerryBMS in sniffer mode, you must connect your USB-to-CAN adapter in the Xanbus network, just like any other Xanbus-enabled Conext devices. On the USD-to-CAN adapter, connect RJ-45 pin 4 to CAN_L and pin 5 to CAN_H. After connecting your CAN adapter, make you enable the can0 interface using:

/sbin/ip link set can0 up type can bitrate 250000

Optionally, you can also use socketcand (https://github.com/linux-can/socketcand) to guarentee delivery and ordering of CAN messages to BerryBMS.

You can use sniffer mode and polling mode simultaneously.

Using It

First, modify config.yaml to your needs.

Once the configuration file has been adjusted to your needs, you can use this application directly from the command-line, simply do:

cd ~/berrybms
python3 -m venv
python3 berrybms/berrybms.py

This will output tons of information and push things in MQTT. What's pushed in MQTT can be used by the Dash/Plotly frontend of this application. Alternatively, you can use the date with any other solution, like Node-RED. You can also add the -d parameter to make it run continuously, it will pause at the configured interval between each data fetch.

If you want to try the Web GUI, simply do:

python3 berrybms/berrydash.py

and open http://<IP of your Raspberry PI device>:8080/ using your favorite web browser.

Running As Service

If you want to run this program through systemd, first edit berrybms.service and adjust the path relative to the user the program will be running as. Then, do:

sudo cp berrybms.service /lib/systemd/system/
sudo chmod 644 /lib/systemd/system/berrybms.service
sudo systemctl daemon-reload
sudo systemctl enable berrybms.service
sudo systemctl start berrybms.service

If you want the Dash interface as a service, first edit berrydash.service and adjust the path relative to the user the program will be running as. Then, do:

sudo cp berrydash.service /lib/systemd/system/
sudo chmod 644 /lib/systemd/system/berrydash.service
sudo systemctl daemon-reload
sudo systemctl enable berrydash.service
sudo systemctl start berrydash.service

What's Next?

I have some ideas floating around for future improvements. Among them, there are:

1. Better error-handling everywhere
2. Adjust labelling based on XW grid input associations
3. Offer some actions through the Web GUI (start/stop the generator, enable/disable BMS charge/discharge, etc.)
4. Offer pack balancing capabilities by enabling/disabling charge/discharge on each BMS to maintain close SOC between BMS
5. Add support for a single-channel relay to start/stop the generator based on SOC