A production-ready full-stack IoT environmental monitoring system with 13 sensors, battery + solar power, deep-sleep optimization achieving ~92% power savings, and a real-time web dashboard. The system wakes every 30 minutes, reads all sensors, transmits data over Wi-Fi, and goes back to sleep -- designed to run indefinitely on solar power.
This is the firmware repository. See also: Backend API | Web Dashboard
Shows total sensors, online/offline count, 1,786+ measurements recorded, and live sensor cards with real-time readings.
13 environmental sensors displayed with real-time values, status indicators (green = online), and "View details" drill-down.
Individual sensor view with configurable time range (24h / 7d / 30d / All), interactive time-series chart, and min/max/avg statistics.
Filter sensors by category: Environment, Light, Gas, Soil. Showing 8 gas sensors monitoring CO, CH4, NH3, H2S, HCHO, O2, O3, and Alcohol.
- 13 Environmental Sensors with I2C multiplexing (2x TCA9548A) for address conflict resolution
- ~92% power savings via deep sleep + MOSFET-switched sensor power rail
- Solar-powered indefinite runtime -- 10W panel generates 4.3x daily consumption
- Battery backup -- 3x 18650 cells (8,700 mAh), 5+ days without sunlight
- Dual measurement cycle -- status check then data upload for reliability
- Production-ready dashboard with real-time updates, historical charts, and statistics
- Full-stack architecture -- ESP32-S3 firmware + FastAPI backend + React/TypeScript frontend
10W Solar Panel --> Charge Controller --> 3x 18650 Battery (8700 mAh)
|
Boost Converter (3.7V --> 5V)
|
+-------------------------+-------------------------+
| |
ESP32-S3 MCU N-Channel MOSFET (GPIO 14)
(always powered, ~20mA sleep) |
| Sensor 5V Rail
| |
+-----+-----+ All 13 Sensors (~300mA active)
| |
Timer Wake Button Wake -----> Double Measurement Cycle
(30 min) (GPIO 7) |
WiFi POST --> FastAPI Backend --> SQLite DB
|
React Dashboard (live updates)
| State | Current Draw | Duration | Description |
|---|---|---|---|
| Active | ~300 mA | ~8 sec | Sensors ON, reading + WiFi transmit |
| Deep Sleep | ~26 mA | ~1792 sec | ESP32 sleep + sensors OFF via MOSFET |
| Average per cycle | -- | 30 min | 13.62 mAh per cycle |
| Daily consumption | -- | 24h | ~653 mAh @ 5V (~3.27 Wh) |
| Solar generation | -- | 5h sun | ~2,800 mAh/day (4.3x surplus) |
The system uses two TCA9548A I2C multiplexers to avoid address conflicts across 13 sensors:
| Sensor | Parameter(s) | Protocol | I2C Mux |
|---|---|---|---|
| MS8607 | Temperature, Humidity, Pressure | I2C | MUX_A (0x70) |
| BH1750 | Light Intensity (lux) | I2C | MUX_A |
| MLX90614 | IR Object Temperature | I2C | MUX_B (0x71) |
| Soil Moisture | Capacitive Moisture | I2C | MUX_B |
| Alcohol (MQ-135) | Alcohol concentration (ppm) | I2C/Analog | MUX_A |
| CH4 (MHZ9041A) | Methane (LEL) | I2C | MUX_A |
| Soil EC + pH | Conductivity, pH | UART2 (RS485) | -- |
| HCHO | Formaldehyde (ppm) | UART1 | -- |
| H2S | Hydrogen Sulfide (ppm) | I2C | MUX_A |
| O2 | Oxygen (%vol) | I2C | MUX_B |
| NH3 | Ammonia (ppm) | I2C | MUX_B |
| CO | Carbon Monoxide (ppm) | I2C | MUX_B |
| O3 | Ozone (ppm) | I2C | MUX_B |
- I2C Bus: SDA = GPIO 8, SCL = GPIO 9
- Sensor Power: 5V rail controlled by N-Channel MOSFET (GPIO 14)
- UART2 (Soil EC+pH): RX = GPIO 19, TX = GPIO 20 (RS485)
- UART1 (HCHO): RX = GPIO 5
- Wake Button: GPIO 7 (internal pull-up)
- Status LED: Blue LED on GPIO 21
Each wake-up performs two measurement cycles for reliability:
- Cycle 1 -- Status Check: Power ON sensors --> Read all values --> Serial output --> Power OFF (no upload)
- Cycle 2 -- Data Upload: Power ON sensors --> Read all values --> POST to backend --> Power OFF
- Deep Sleep: Enter deep sleep for 30 minutes (wake on timer or button)
{
"ts_ms": 1234567,
"MS8607": { "status": "online", "T": 24.3, "RH": 60.5, "P": 1013.2 },
"BH1750": { "status": "online", "lux": 185.0 },
"MULTIGAS": {
"H2S": { "status": "online", "value": 0.0, "unit": "ppm" },
"O2": { "status": "online", "value": 20.9, "unit": "%vol" },
"CO": { "status": "online", "value": 0.0, "unit": "ppm" }
},
"summary": { "sensor_total": 13, "offline_count": 0 }
}SENSOR_SAMPLE_INTERVAL_MS = 1800000 // 30 min between wake cycles
SENSOR_POWER_WARMUP_MS = 3000 // 3 sec sensor stabilization
SENSOR_MOSFET_PIN = 14 // GPIO controlling sensor power
WAKE_BUTTON_PIN = 7 // Manual wake-up button
STATUS_LED_PIN = 21 // Blue status LED- Solid ON: Device awake, WiFi connected
- Rapid blink (5x): WiFi connection failed
- Slow blink (1Hz): Awake but WiFi offline (button wake mode)
- OFF: Deep sleep
Plant-Sensor-Network/
|-- Firmware/
| +-- Plant Sensor Network/
| |-- src/
| | |-- sensors.cpp # Main sensor reading + WiFi logic
| | |-- config.h # WiFi credentials + backend URL
| | +-- DFRobotHCHOSensor.h # HCHO sensor library
| +-- platformio.ini # PlatformIO configuration
|-- Schemetic/ # Circuit schematics and PCB files
|-- Images/ # Dashboard and system screenshots
|-- Documents/ # BOM, technical documentation
+-- Requirements/ # System specifications
Edit src/config.h:
#define WIFI_SSID "YourWiFiNetwork"
#define WIFI_PASSWORD "YourPassword"
#define BACKEND_URL "http://YOUR_PC_IP:8000/api/ingest"Upload to ESP32-S3 via PlatformIO.
cd Plant-Sensor-Network-Backend
pip install -r requirements.txt
python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000cd Plant-Sensor-Network-Web-Dashboard
npm install
# Configure .env: VITE_BACKEND_URL=http://YOUR_PC_IP:8000
npm run devDevice will: Connect to WiFi --> Measure --> Upload --> Deep sleep (30 min) --> Repeat
| Component | Technology |
|---|---|
| MCU | ESP32-S3 |
| Firmware | C/C++ (PlatformIO) |
| Backend | FastAPI + SQLite |
| Frontend | React + TypeScript + Tailwind CSS |
| Protocols | I2C, UART, RS485, Wi-Fi, HTTP REST |
| Power | 10W Solar + 3x 18650 batteries + MOSFET switching |
| PCB Tools | Schematic design files included |
Muhammad Zaeem Sarfraz -- Electronics & IoT Hardware Engineer
- 📧 Zaeem.7744@gmail.com
- 🌍 Vaasa, Finland
Document Version: 1.0 | Status: Production Ready | Last Updated: February 2026