Custom Firmware for Elecrow Watering Kit 2.1

I purchased the Elecrow Watering Kit 2.1 from Amazon. The controller board for this kit has an integrated Arduino Leonardo.

This kit was a great starting point but I've since made numerous changes from the original code.

Waterproofing

I intended to use this kit outside. To that end, I knew I'd have to waterproof the electronics (I placed them in a plastic box that had a nice seal on the lid and made small holes for the cables). I used a similar box for the pump and vavle block.

I quickly discovered the Moisture Probes would need to be insulated / sealed. I used a piece of 3/4" heat shrink tubing along with a piece of 1/2" heat shrink tubing to enclose the electronics and connector on the probe. I used a heat gun to shrink the tubing.

I then used Liquid Electrical Tape to seal the ends of the heat shrink tubing to provide additional water protection.

Water Connections

I found the water connections to the Pump and Vavle Block had a tendency of separating. I used some E6000 to glue the connections into places.

Changes to my code

• Formatting and variable names made more consistent
• This supports the VL53L0X Time-of-Flight (ToF) Laser Ranging Sensor I2C IIC module. I connected this to the IIC port on the Elecrow board.
• Sensor data is sent over TX to an ESP8266 (if available) running the esp8266-app/esp8266-app.ino. I use a D1 Mini (clone) for this
• The ESP8266 obtains VCC and GND from port that contains MOSI and MISO.
• The ESP8266 RX pin is connected to the Elecrow's RX pin (which is actually TX, but seems to be mis-labeled).
  • This will preclude you from using the Serial monitor on the ESP8266. I do not think there is a way around this.
  • Since I am using the ESP8266 RX pin, I do not believe you need to use a level shifter (since it has a resistor in-place). If you are squeemish, you might with to use a 3.3V/5V level shifter here such as the 74AHCT125.
• Code on the ESP8266 connects to a local Mosquitto broker. From there you can get the data to wherever makes sense. I'm using Node-RED to migrate the data to InfluxDB which I can then visualize using Grafana.
• Added a moisture-calibration app to help calibrate the values coming from the moisture sensors.
• MINOR Auto-calibration of the moisture sensors, but you should still run mosisture-calibration app to get the absolute wet and dry values for your sensors.

Monitoring with MQTT, Node-RED, InfluxDB, and Grafana

Data to MQTT

• Assuming the esp82660-app is be receiving stats from the Elecrow watering kit and can connect to the MQTT broker, will publish stats to a topic. The topic I publish to is home/watering-1/data.
• I use the IOTStack project to run MQTT, Node-RED, InfluxDB, and Grafana. I run on the stack on a Raspberry Pi that uses an M.2 SSD over USB for all storage for increased speed and reliability vs using an SD card.

Node-RED move the data from MQTT to Influx DB

I use a very simple Flow in Node-Red to move my data from MQTT to InfluxB.

• Node 1 mqtt in. This connects to your Mosquitto broker and subscribes to the Topic home/watering-1/data.
• Node 2 function node. The input comes from Node 1. The purpose of this function is to take the CSV data from MQTT and convert it into a JSON object. This function contains the code

const data = msg.payload.split(",");
msg.payload = {
    'moisture-0' : parseInt(data[0]),
    'moisture-1' : parseInt(data[1]),
    'moisture-2' : parseInt(data[2]),
    'moisture-3' : parseInt(data[3]),
    'pump-0' : parseInt(data[4]),
    'water-level-0' : parseInt(data[5]),
    'water-level-per-0' : parseInt(data[6]),
    'valve-0' : parseInt(data[7]),
    'valve-1' : parseInt(data[8]),
    'valve-2' : parseInt(data[9]),
    'valve-3' : parseInt(data[10])
};
return msg;

• Node 3 influxdb out. The input comes from Node 2. This connects to your InfluxDB database. I output to a measurement home/watering-1.
• Node 4 debug. The input also comes from Node 2. I find this helpful to view the data as it comes from MQTT and gets converted to JSON.

Displaying Stats

Find a simple Grafana tutuorial and make charts from the data that is now in your InfluxDB within home/watering-1. I may expand this in the future.

My dashboard looks like

Calibration

• If you flash the moisture-calibration firmware, you can determine the WET_VALUE and DRY_VALUE for your sensors to make sure you are covering the complete range.
• Run the app on the Elecrow hardware with the Moisture sensors (and optionally the ToF sensor installed).
• For each moisture sensor observe minimum values for wet and maximum values for dry. Adjust WET_VALUE and DRY_VALUE in watering-kit-config.h accordingly.
• If using the ToF sensor to check the depth of your water reservoir, this will show you distance in millimeters. Affix your sensor over your reservoir and then get the reading from the Elecrow display. Adjust MAX_WATER_DEPTH in watering-kit-config.h to a value LESS than this (you don't want the water to actually reach the sensor at "Full").

Programming notes on Windows

• I found programming this Leonardo to bit a bit of a pain on Windows.
• On Windows, the Leonardo consumes two COM ports, let's say they are COM5 and COM6. Below I'll call these COMn and COMn+1
• COMn is for the Serial Monitor, when running a program.
• COMn+1 is for programming.
• Within the Arudino app's Preferences, make sure Show Verbose Output on Upload is selected.
• To program
  • Plug in the Elecrow board via USB
  • Make sure COMn+1 is selected (if not listed, press and hold the Reset button on the Elecrow)
  • Click Verify to compile the application
  • Press and hold the Reset button
    • Click Send
    • Wait for the build to ALMOST finish and let go of the Reset button
    • When you start getting messages listing all of the COM ports, press Reset again (don't hold)
    • Programming should start (and complete)
• To view the Serial console, switch to COMn (re-open the Serial Monitor, if necessary)

Acknowledgements

I've included a copy of the Elecrow code in the orig/ folder for comparison purposes. The firmware that was provided by Elecrow has several issues, not the least of which are graphical gitches on the display.

I found a version of the firmware modified by liutyi that fixed these gitches.

My code started it's life based on liutyi's modifications but have since undergone significant changes.

Additional ideas and code were borrowed from rfrancis97 to use an ESP8266 (or ESP32) to publish stats to an MQTT broker. rfrancis97 also implemented a way to measure water level using an an Ultrasonic sensor (HC-SR04). I liked this idea but decided to use a VL53L0X Time-of-Flight (ToF) Laser Ranging Sensor I2C IIC module, instead.