A high-efficiency, battery-powered wind speed sensor (Anemometer) using the ESP32. This project leverages the ULP (Ultra-Low Power) co-processor to monitor wind pulses while the main CPU is in Deep Sleep, allowing for years of battery life.

• Ultra-Low Power: Can be as low as ~15µA in deep sleep on the ESP32 alone. Actual current depends on the board and any attached hardware. The main cores only wake up to calculate and transmit data.
• BTHome V2 Protocol: Works natively with Home Assistant via Bluetooth—no custom integration or ESPHome YAML required. See bthome.io.
• Intelligent Reporting:
  • Wind Detected: Typically reports on the next 5-second wake cycle while wind is present.
  • Change Detection: Tries to report a speed change as soon as it is noticed on the next wake cycle.
  • Heartbeat: Sends a periodic "still alive" update, usually about every 60 seconds during calm periods.
• Hardware Debouncing: ULP-based software sampling filters out mechanical reed switch "bounce."

1. ESP32 or ESP32-S3 Board.
   • For ESP32-C6 check the ESP32-C6 anemometer project.
2. Anemometer (3-cup type with a Reed Switch).
3. Battery: Li-ion 18650 or LiPo.

Connect the reed switch between SENSOR_GPIO pin and GND.

For battery monitoring, a board with an onboard battery divider is recommended.

Install the following libraries via the Arduino Library Manager:

• NimBLE-Arduino (Lightweight Bluetooth LE)
• BTHomeV2-Arduino (BTHome data formatting)

Adjust the constants at the top of the sketch to match your hardware:

const float RADIUS_METERS      = 0.071; // Center to cup middle
const float CALIBRATION_FACTOR = 2.5;   // Aerodynamic factor
const int   PULSES_PER_REV     = 2;     // Number of magnets in your sensor

Tuning Guide:

• RADIUS_METERS: Measure from the center of the shaft to the center of a cup.
• PULSES_PER_REV: Set this to the number of magnet passes per full revolution.
• CALIBRATION_FACTOR: Adjust this if you want accurate wind-speed measurements. Compare the reading against a known wind source or a reference sensor, then increase or decrease this factor until they match closely.

If you are only using the anemometer for threshold-based automation, such as closing an awning with a Shelly device, a precise calibration is usually not required. In that case, a stable and repeatable reading is enough to trigger the chosen wind-speed threshold.

Upload the code to your ESP32. Open the Serial Monitor (115200 baud) to verify the "Fresh Boot" and "Deep Sleep" cycles.

1. Ensure the Bluetooth integration is active in Home Assistant.
2. The sensor will be automatically discovered as a BTHome device named "Wind Sensor".
3. Add it to your dashboard. The wind speed will appear as a sensor entity in m/s.

If you are using Shelly devices, the anemometer can also be configured as a virtual component in your Shelly setup. This does not require Home Assistant.

Depending on the Shelly virtual component setup, you may need the ESP32 BLE address. To print it directly from the board, upload this small sketch and read the value from the Serial Monitor:

#include <Arduino.h>
#include <NimBLEDevice.h>

void setup() {
  Serial.begin(115200);
  NimBLEDevice::init("ESP32");
}

void loop() {
  BLEAddress localAddress = NimBLEDevice::getAddress();
  Serial.println(localAddress.toString().c_str());
  delay(1000);
}

Use the anemometer to automatically close an awning with a Shelly 2PM Gen3 when wind speed gets too high, without any extra wiring between the wind sensor and the Shelly device. The anemometer broadcasts wind data over BLE, and Shelly handles the automation logic on its side. In this setup, the ESP32 board and battery are hidden inside the square bar, which also supports the anemometer using an interlocking fit (no screw required).

• Deep Sleep (ULP Active): ~15µA minimum, depending on the board and attached hardware
• BLE Transmission (1.5s): ~100mA
• Total Battery Life: ~1-2 years on a 2500mAh 18650 cell (depending on wind frequency).

This project is licensed under the MIT License - see the LICENSE file for details.

• NimBLE-Arduino by h2zero
• BTHome V2 protocol by bthome.io.