plantbuddy

nodeMCU based moisture monitoring for plants 🌱 with a serverless backend

Main Features

Collect sensor readings every 5 minutes
Submit values to serverless infrastructure using local WiFi network
Automatically fetch weather report for each reading using backend function
later have a UI as well 🙄

Hardware

The hardware setup consists of a ESP8266 Microcontroller with a powersupply and two connected sensors.

DHT22 connected to D2 digital input pin
Soil moisture sensor connected to A0
Bridge between D0 and RST to enable deep sleep for optimizing battery life
switch off sensors during deepsleep using a transistor driven via pin D1

To program the ESP8266 you need to clear the connection to RST

Power consumption

ESP8266 draws ~80mA when active and few µA when in deepsleep
Maximum output current per GPIO pin is 12mA
Photo resistor has 10kOhm resistor -> 3V3 / 10kOhm = 33mA
Moisture sensor consumes 35mA
DHT222 max current 2.5mA

Microcontroller

The ESP8266 chip is programmed with the esp8266.ino file using the Arduino IDE. Secrets for connecting to local WiFi and endpoints to post data need to be defined in a esp8266.secrets.c file which is specified at esp8266.secrets.template.c.

Setup

Serverless backend

As a backend solution the plantbuddy runs on Google's firebase 🔥 platform. The main functionalities used from firebase are:

Firestore database to store all sensor readings and heartbeats
Firebase functions for providing an HTTP endpoint as well as querying for weather data for each sensor reading

Note: As outbound traffic is required to collect weather data, the project should run in BLAZE plan. Function and database usage should be within the free tier, only outbound traffic will be billed.

Backend code

Before deploying the backend code several variables need to be set so the service can work correctly. An API token for OpenWeatherMap can be obtained from their website. A free account should suffice unless a LOT of plantbuddies are running on one account.

firebase functions:config:set owm.key=<OpenWeatherMap API Token>
firebase functions:config:set owm.city_id=2873776 // according to [OWM City ID List](https://openweathermap.org/current#cityid)
firebase functions:config:set header.secret=<some secret>

Local development

For devleoping the function locally:

# get a copy of the environment config
firebase functions:config:get > functions/.runtimeconfig.json
# start function host locally
firebase serve --only functions

Deployment

All the Firebase stuff gets automatically deployed via Travis CI. To set it up create a encrypted token and update the Travis CI Config.

# create a firebase CI token
firebase login:ci

# encrypt the token to your project
docker run --rm caktux/travis-cli encrypt "<token from firebase login:ci>" -r anoff/plantbuddy

ymmit85/plantbuddy