JBD BMS and Thornwave bluetooth data monitoring
JBD - Battery Management System
jbdbms.py is a backend for using bluetooth to pull data from a JBD/Overkill BMS https://overkillsolar.com and processing it for applications.
I've added separate jbdbms python scripts one for 16 cells another for 8 cell packs : jbdbms-16-socket.py and jbdbms-8-socket.py along with jbdbms-16-mqtt.py and jbdbms-8-mqtt.py. There is also a test version for 4 cell packs. (Let me know if it works as I have no way of testing)
The JBD BMS uses either serial or bluetooth to access its data and this project is for the bluetooth interface. The way JBD implemented this is not standard as in sending read requests or turning on notifications to receive data. It requires sending without data, write requests to handles (0x03 and 0x04), i.e. 'dda50400fffc77'. These messages cause the device to return a single notification response via a different handle. The returned notify is broken into 2 messages. The first is the start of the message and the second is the last half.
So far I have only programmed the 0x03 (pack info) and 0x04 (cell voltages) for system monitoring.
jbdbms.py alone is required fetching bms data. Just discover your JBD BLE address. I use hcitool lescan to find this or the Xiaoxiang app also shows this at startup. The jbdbms.py program is initiated with a (-b) device BLE address, (-i) collection interval and (-m) monitor name.
jbdbms-16-socket.py -b xx:xx:xx:xx:xx -i 10 -m jbdbms
This program only prints out ongoing bms data for testing and viewing until the socket write commands are uncommented and Telegraf is setup to listen to socket and deliver data to influxdb. The output format can be changed to your own requirements and application.
It is using bluetooth (bluepy) for fetching data from a JBD BMS or Overkill BMS and building data structures then formatting it to cvs data and writing it to a unix socket for Telegraf's Socket Listener Input Plugin. The data then preceeds via Influxdb to Grafana for graphing. Telegraf is the middleman, collecting arriving data and getting it into Influxdb. Other than configuraton files, the user only works with Grafana via a browser accessing and graphing data in dashboards. I included my Grafana bms dashboard json file that can be used as a template, as seen below.
MQTT & JSON
Now can choose either mqtt or socket output. The mqtt version has been changed from csv to json output to make using mqtt data easier with other programs. I have also included mqtt listener with json inputs in the changed section that can be added to a existing telegraf.conf file only replacing the socket listener with the mqtt listener. Of course, also need to first install Mosquitto or another mqtt server/client.
THORNWAVE
My system uses two Thornwave PowerMon bluetooth battery monitors. One PowerMon shunt for the solar charge controller and the another shunt for the inverter. These are much simpler to use, only requiring a read request at (0x15) which returns the data for processing. (these are well made and fairly inexpensive shunt monitors https://www.thornwave.com)
For Thornwave see https://github.com/mkjanke/ThornwavePy I have only modified it for cvs data, writing to Unix socket and leaving connection open for ongoing data. The meters-socket.py is for the Thronwave meters. I also included a Grafana Thornwave dashboard json file that can be used as a template as seen below.
thornwave-socket.py -b CF:E5:F3:D1:9F:87 -i 30 -m solar
SETUP
It's fairly simple with linux, download a version of jbdbms-socket.py or thornwave-socket.py or both, use and test data output, if okay, then via your app manager install Telegraph, InfluxDB, and Grafana. The default configurations are okay for both InfluxDB and Grafana.
For using InfluxDB you need to create a new database and a user with password.
Open a terminal and type :
- influx
- CREATE DATABASE battery
- influx user create -n "username" -p "password" -o "org-name"
- exit
Telegraf configuration requires the following to be added :
[[outputs.influxdb]]
- urls = ["http://127.0.0.1:8086"]
- database = "battery"
- username = "username"
- password = "password"
[[inputs.socket_listener]]
- service_address = "unixgram:///tmp/telegraf.sock"
- data_format = "csv"
- csv_header_row_count = 1
- name_override = "battpack"
- data_type = "float"
- csv_tag_columns = ["meter"]
That's it, now open Grafana at 127.0.0.1:3000 or substituting pi's ip. If data doesn't appear, launch Telegraf with --debug option, showing output with more information about any errors in processing data.
With linux it's utilizing systemd services with automatic loading on startup and restarting if connection lost.
- copy service files to /etc/systemd/system/
- systemctl start jbdbms.service
- systemctl enable jbdbms.service
- systemctl start solar.service
- systemctl enable solar.service
- systemctl start inverter.service
- systemctl enable inverter.service
These service files must be updated to show version used and location. I have only used with linux but should be okay from any python and bluetooth ready computer. I use a RaspberryPi Zero W for 24/7 and proximity to device's bluetooth signal and feeding a webserver giving access to any device on my lan.
Influx also offers a free cloud version of their database which would give the ability to access the bms from anywhere via the internet. Haven't tried this yet.
UPDATE
Using my PiZero W, being so low powered sometimes slowing down to where it was barely usable. Checking with HTOP could see that the cpu was at 100% and very heavy RAM use due to Influxdb. So I found a better time series database - Victoria-Metrics. It is very fast with a better query language and the cpu uses only 20% of what Influxdb used. The RAM usage is also 60% less.
All it required was to swap Victoria-Metrics for Influxdb. Telegraf and Grafana stayed the same. For Telegraf's configuration it uses a different port than influxdb for output.