Skip to content

Latest commit

 

History

History
 
 

rpi

README.md

Communicating with Hoymiles Micro-Inverters using Python on RaspberryPi

The tools in this folder (and subfolders) provide the ability to communicate with Hoymiles micro-inverters.

They require the hardware setup described below.

The tools are still quite rudimentary, as the communication behaviour is not yet fully understood.

This is part of an ongoing group effort, and the knowledge gained so far is the result of a crowd effort that started at [1].

Thanks go to all who contributed, and are continuing to contribute, by providing their time, equipment, and ingenuity!

Required Hardware Setup

ahoy.py has been successfully tested with the following setup

  • RaspberryPi Model 2B (any model should work)
  • NRF24L01+ Radio Module connected as described, e.g., in [2] (Instructions at [3] should work identically, but [2] has more pretty pictures.)
  • TMRh20's 'Optimized High Speed nRF24L01+ Driver' [3], installed as per the instructions given in [4]
    • Python Library Wrapper, as per [5]

Building the NRF24 Python Wrapper

You have to install the NRF24 Python Library, as a Dependency for the Raspberry Pi Version of Ahoy.

To do that correctly, I have contacted the developer of NRF24 via github Python 3 Wrapper not installing properly #845 as I could not get the Python Wrapper for NRF24 to be built.

  • Install Raspberry Pi OS lite x86 with raspberry pi imager
  • Connect nrf24 module to raspberry pi (as described in github)
  • Login with user pi
  • Execute sudo apt update && sudo apt -y upgrade
  • Execute sudo raspi-config and
    • Select "Expand filesystem" in "Advanced Options"
    • Activate "SPI" in "Interface Options"
    • "Finish" to exit raspi-config Tool, reboot YES!
  • Login as pi user again
sudo apt install cmake git python3-dev libboost-python-dev python3-pip python3-rpi.gpio

sudo ln -s $(ls /usr/lib/$(ls /usr/lib/gcc | \
     tail -1)/libboost_python3*.so | \
     tail -1) /usr/lib/$(ls /usr/lib/gcc | \
     tail -1)/libboost_python3.so

git clone https://github.com/nRF24/RF24.git
cd RF24

export RF24_DRIVER=SPIDEV
rm Makefile.inc #just to make sure there is no old stuff
mkdir build && cd build
cmake ..
make
sudo make install

cd ../pyRF24
rm -r ./build/ ./dist/ ./RF24.egg-info/ ./__pycache__/ #just to make sure there is no old stuff
python3 -m pip install --upgrade pip
python3 -m pip install .
python3 -m pip list #watch for RF24 module - if its there its installed

cd ..
cd examples_linux/
python3 getting_started.py # to test and see whether RF24 class can be loaded as module in python correctly

If there are no error messages on the last step, then the NRF24 Wrapper has been installed successfully.

Building RF24 Wrapper for Debian 11 (bullseye) 64 bit operating system

The description above does not work on Debian 11 (bullseye) 64 bit operating system. Please check first, if you have Debian 11 (bullseye) 64 bit operating system installed:

  • uname -a search for aarch64
  • lsb_release -d
  • cat /etc/debian_version

There are 2 possible solutions to install the RF24 wrapper:

1. Solution:

sudo apt install cmake git python3-dev libboost-python-dev python3-pip python3-rpi.gpio

sudo ln -s $(ls /usr/lib/$(ls /usr/lib/gcc | \
     head -1)/libboost_python3*.so | \
     tail -1) /usr/lib/$(ls /usr/lib/gcc | \
     head -1)/libboost_python3.so

git clone https://github.com/nRF24/RF24.git
cd RF24

rm -rf build Makefile.inc 
./configure --driver=SPIDEV

edit Makefile.inc with your prefered editor e.g. nano or vi

old:

CPUFLAGS=-marm -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp -mfloat-abi=hard
CFLAGS=-marm -march=armv6zk -mtune=arm1176jzf-s -mfpu=vfp -mfloat-abi=hard -Ofast -Wall -pthread

new:

CPUFLAGS=
CFLAGS=-Ofast -Wall -pthread

continue now

make
sudo make install

cd pyRF24
rm -r ./build/ ./dist/ ./RF24.egg-info/ ./__pycache__/ #just to make sure there is no old stuff
python3 -m pip install --upgrade pip
python3 -m pip install .
python3 -m pip list #watch for RF24 module - if its there its installed

2. Solution:

sudo apt install git python3-dev libboost-python-dev python3-pip python3-rpi.gpio

git clone --recurse-submodules https://github.com/nRF24/pyRF24.git
cd pyRF24
python3 -m pip install . -v     # this step takes about 5 minutes on my RPI-4 !

If you have problems with your radio module from ahoi, e.g.: cannot interpret received data, please try to reduce the speed of your radio module! Add the following parameter to your ahoy.yml configuration file in "nrf" section: spispeed: 600000 (0.6 MHz)

Required python modules

Some modules are not installed by default on a RaspberryPi, therefore add them manually:

pip install crcmod pyyaml paho-mqtt SunTimes

Configuration

Local settings are read from ahoy.yml
An example is provided as ahoy.yml.example

Example Run

The following command will run the communication tool, which will try to contact the inverter every second on channel 40, and listen for replies.

Whenever it sees a reply, it will decoded and logged to the given log file.

$ sudo python3 -um hoymiles --log-transactions --verbose --config /home/dtu/ahoy.yml | tee -a log2.log

Python parameters

  • -u enables python's unbuffered mode
  • -m hoymiles tells python to load module 'hoymiles' as main app

The application describes itself

python3 -m hoymiles --help
usage: hoymiles [-h] -c [CONFIG_FILE] [--log-transactions] [--verbose]

Ahoy - Hoymiles solar inverter gateway

optional arguments:
  -h, --help            show this help message and exit
  -c [CONFIG_FILE], --config-file [CONFIG_FILE]
                        configuration file
  --log-transactions    Enable transaction logging output
  --verbose             Enable debug output

Inject payloads via MQTT

To enable mqtt payload injection, this must be configured per inverter

...
  inverters:
...
    - serial: 1147112345
      mqtt:
        send_raw_enabled: true
...

This can be used to inject debug payloads The message must be in hexlified format

Use of variables:

  • tttttttt expands to current time like we know from our 80 0b command

Example injects exactly the same as we normally use to poll data

$ mosquitto_pub -h broker -t inverter_topic/command -m 800b00tttttttt0000000500000000

This allows for even faster hacking during runtime

Analysing the Logs

Use basic command line tools to get an idea what you recorded. For example:

$ cat log2.log
[...]
2022-05-02 16:41:16.044179 Transmit | 15 72 22 01 43 78 56 34 12 80 0b 00 62 3c 8e cf 00 00 00 05 00 00 00 00 35 a3 08
2022-05-02 17:01:41.844361 Received 27 bytes on channel 3: 95 72 22 01 43 72 22 01 43 01 00 01 01 44 00 4e 00 fe 01 46 00 4f 01 02 00 00 6b
2022-05-02 17:01:41.886796 Received 27 bytes on channel 75: 95 72 22 01 43 72 22 01 43 02 8f 82 00 00 86 7a 05 fe 06 0b 08 fc 13 8a 01 e9 15
2022-05-02 17:01:41.934667 Received 23 bytes on channel 75: 95 72 22 01 43 72 22 01 43 83 00 00 00 15 03 e8 00 df 03 83 d5 f3 91
2022-05-02 17:01:41.934667 Decoded: 44 string1= 32.4VDC 0.78A 25.4W 36738Wh 1534Wh/day string2= 32.6VDC 0.79A 25.8W 34426Wh 1547Wh/day phase1= 230.0VAC 2.1A 48.9W inverter=114171230143 50.02Hz 22.3°C
[...]

A brief example log is supplied in the example-logs folder.

Todo

  • Ability to talk to multiple inverters
  • MQTT gateway
  • understand channel hopping
  • configurable polling interval done: interval ist configurable in ahoy.yml
  • commands
  • picture of setup!
  • python module
  • ...

References