Raspberry Pi Weekend Projects: Build an I2C Smart Plant Monitor

Why This Build Beats Basic Raspberry Pi Weekend Projects

Most raspberry pi weekend projects stall out at blinking LEDs, basic web servers, or reading a single analog sensor via an ADC. If you want a project that actually exercises the Pi’s hardware interfaces and teaches you robust embedded Python, you need to work with I2C multiplexing, environmental sensing, and inductive load switching.

This guide walks through building a smart greenhouse monitor using a Raspberry Pi 5 (8GB variant) running Raspberry Pi OS (Bookworm, 64-bit). We will interface a BME280 environmental sensor and a capacitive soil moisture probe over the I2C bus, and trigger a 5V relay to control a water pump. Unlike toy projects, this build includes production-style error handling for I2C bus lockups and hardware timeouts.

Hardware Spec Sheet & Sensor Selection

Choosing the right sensors is where most weekend builds fail. Resistive soil sensors corrode within weeks, and DHT22 sensors rely on bit-banged 1-Wire protocols that drop packets under heavy CPU load. We are using I2C-based capacitive sensing and a Bosch BME280 for rock-solid data.

Note: Always buy the BME280, not the BMP280. The BMP280 lacks the humidity sensor. For a deep dive on the BME280 breakout, refer to the Adafruit BME280 Guide.

Wiring & Pin Mapping

The Raspberry Pi 5 uses a 3.3V logic level for its I2C bus. While the Pi 5 has internal 1.8kΩ pull-up resistors on SDA and SCL, keeping your I2C wire runs under 30cm (12 inches) is critical to avoid bus capacitance issues. Use 24 AWG stranded silicone wire for breadboarding.

Python Control Code with I2C Error Handling

To interact with these sensors, we use the Adafruit Blinka ecosystem, which ports CircuitPython libraries to the Raspberry Pi's Linux environment. We also use gpiozero for the relay.

First, install the dependencies via your terminal:

sudo apt update
sudo apt install python3-gpiozero i2c-tools
pip3 install adafruit-circuitpython-bme280 adafruit-circuitpython-seesaw

Below is the complete, compilable Python script. It reads the soil moisture and ambient temperature, triggers the pump if the soil is dry, and includes robust try/except blocks to handle I2C bus lockups—a common failure mode in weekend builds.

import time
import board
import busio
import adafruit_bme280
from adafruit_seesaw.seesaw import Seesaw
from gpiozero import OutputDevice
import sys

# --- PIN & CONFIGURATION DEFINITIONS ---
RELAY_PIN = 18                  # BCM GPIO 18 (Physical Pin 12)
PUMP_DURATION_SEC = 5           # How long to run the pump
SOIL_DRY_THRESHOLD = 2000       # Capacitive value (Lower = wetter, Higher = drier)
READ_INTERVAL_SEC = 30          # Time between sensor reads

# Initialize GPIO Relay (Active Low for most optocoupler relay boards)
pump_relay = OutputDevice(RELAY_PIN, active_high=False, initial_value=False)

# Initialize I2C Bus
i2c = busio.I2C(board.SCL, board.SDA)

def init_sensors():
    """Initialize I2C sensors with error handling for address conflicts."""
    try:
        bme = adafruit_bme280.Adafruit_BME280_I2C(i2c, address=0x77)
        # Seesaw chip handles the capacitive touch and temp on the STEMMA soil sensor
        soil = Seesaw(i2c, addr=0x36)
        print("[INFO] Sensors initialized successfully.")
        return bme, soil
    except ValueError as e:
        print(f"[FATAL] I2C Address Error: {e}")
        print("Check wiring and run 'i2cdetect -y 1' to verify addresses.")
        sys.exit(1)

def read_and_act(bme, soil):
    """Read sensor data and trigger pump if necessary."""
    try:
        # Read BME280
        temp_c = bme.temperature
        humidity = bme.relative_humidity
        
        # Read Soil Moisture (Capacitive touch returns 0-4095)
        soil_moisture = soil.moisture_read()
        soil_temp = soil.get_temp(0x30) # Seesaw internal temp
        
        print(f"[{time.strftime('%H:%M:%S')}] Air: {temp_c:.1f}C | Hum: {humidity:.1f}% | "
              f"Soil Moisture: {soil_moisture} | Soil Temp: {soil_temp:.1f}C")
        
        # Automation Logic
        if soil_moisture > SOIL_DRY_THRESHOLD:
            print("[ACTION] Soil is dry. Triggering water pump...")
            pump_relay.on()
            time.sleep(PUMP_DURATION_SEC)
            pump_relay.off()
            print("[ACTION] Pump cycle complete.")
            
    except OSError as e:
        # Catches I2C Remote I/O errors (Errno 121) or bus lockups
        print(f"[ERROR] I2C Bus Read Failure: {e}")
        print("[WARN] Resetting I2C bus on next cycle...")
        # In a production daemon, you would re-initialize the i2c bus object here
    except Exception as e:
        print(f"[ERROR] Unexpected failure: {e}")

if __name__ == "__main__":
    bme_sensor, soil_sensor = init_sensors()
    
    try:
        while True:
            read_and_act(bme_sensor, soil_sensor)
            time.sleep(READ_INTERVAL_SEC)
    except KeyboardInterrupt:
        print("\n[INFO] Script terminated by user. Ensuring pump is OFF.")
        pump_relay.off()
        sys.exit(0)

Debugging I2C Failures: Exact Errors and Fixes

When working with I2C on the Pi 5, you will inevitably hit a bus error. Do not guess; use the exact error strings to diagnose the fault.

The "First Three Things to Check" Rule

Before rewriting code or swapping sensors, execute these three diagnostic steps:

1. Verify Bus Enumeration: Run sudo i2cdetect -y 1 in the terminal. You must see 36 and 77 in the grid. If the grid is empty, your I2C interface is disabled in raspi-config or your SDA/SCL wires are swapped.
2. Check Voltage Levels: Use a multimeter to measure the voltage between Pin 1 (3V3) and Pin 6 (GND). It must read exactly 3.28V to 3.32V. If it reads 5V, you are probing the wrong rail and risk frying the sensors.
3. Inspect Pull-Up Resistors: The Pi 5 has internal pull-ups, but if you are using wires longer than 30cm, signal edges degrade. Add external 4.7kΩ pull-up resistors between SDA and 3.3V, and SCL and 3.3V.

Common Exact Error Strings

Scaling the Build: Simplify or Extend

Once your raspberry pi weekend projects baseline is stable, you need to decide whether to strip it down for reliability or scale it up for home automation.

How to Simplify (The "Set and Forget" Route)

If you just want data logging without the risk of water leaks from a stuck relay:

• Drop the Relay: Remove the pump circuit entirely. Use the Pi strictly as a datalogger.
• Log to CSV: Add Python's native csv module to append readings to a local file every 5 minutes.
• Use Cron: Instead of a continuous while True loop, strip the loop out and use Linux cron to run the script once every 10 minutes. This frees up CPU cycles and prevents I2C bus lockups from long-running Python memory leaks.

How to Extend (The Smart Home Route)

To turn this into a permanent fixture in your greenhouse:

• Add MQTT: Install paho-mqtt and publish the soil moisture and BME280 data to a local Mosquitto broker. This allows Home Assistant to ingest the data natively via MQTT Discovery.
• Add Light Sensing: Daisy-chain a TSL2591 High Dynamic Range Digital Light Sensor (I2C address 0x29) to the same bus. Plants need PAR (Photosynthetically Active Radiation) data, and the TSL2591 provides excellent lux and IR readings.
• Hardware Watchdog: The Pi 5 includes a hardware watchdog timer. Enable it via systemd to automatically reboot the Pi if the Python script hangs due to an unrecoverable I2C kernel panic.

For more details on configuring the Pi 5's underlying hardware interfaces, always refer to the official Raspberry Pi Configuration Documentation and cross-reference your physical GPIO layout with Pinout.xyz before applying power.