Sensor Component

ESPHome has support for many different sensors. Each of them is a platform of the sensor domain and each sensor has several base configuration options.

Base Sensor Configuration

All sensors in ESPHome have a name and some other optional configuration options. By default, the sensor platform will chose appropriate values for all of these by default, but you can always override them if you want to.

# Example sensor configuration
name: Livingroom Temperature

# Optional variables:
unit_of_measurement: "°C"
icon: "mdi:water-percent"
accuracy_decimals: 1
expire_after: 30s
filters:
  - sliding_window_moving_average:
      window_size: 15
      send_every: 15

Configuration variables:

  • name (Required, string): The name for the sensor.
  • unit_of_measurement (Optional, string): Manually set the unit of measurement the sensor should advertise its values with. This does not actually do any maths (conversion between units).
  • icon (Optional, icon): Manually set the icon to use for the sensor in the frontend.
  • accuracy_decimals (Optional, int): Manually set the accuracy of decimals to use when reporting values.
  • filters (Optional): Specify filters to use for some basic transforming of values. See Sensor Filters for more information.
  • internal (Optional, boolean): Mark this component as internal. Internal components will not be exposed to the frontend (like Home Assistant). Only specifying an id without a name will implicitly set this to true.

Automations:

  • on_value (Optional, Automation): An automation to perform when a new value is published. See on_value.
  • on_value_range (Optional, Automation): An automation to perform when a published value transition from outside to a range to inside. See on_value_range.
  • on_raw_value (Optional, Automation): An automation to perform when a raw value is received that hasn’t passed through any filters. See on_raw_value.

MQTT Options:

  • expire_after (Optional, Time): Manually set the time in which the sensor values should be marked as “expired”/“unknown”. Not providing any value means no expiry.
  • All other options from MQTT Component.

Note

If you’re trying to setup filters for a sensor that has multiple outputs - for example a DHT22 which reports temperature and humidity - put the filters option into each sensor output like this:

sensor:
  - platform: dht
    # ...
    temperature:
      filters:
        # ...
    humidity:
      filters:
        # ...

Sensor Filters

ESPHome allows you to do some basic pre-processing of sensor values before they’re sent to Home Assistant. This is for example useful if you want to apply some average over the last few values.

There are a lot of filters that sensors support. You define them by adding a filters block in the sensor configuration (at the same level as platform; or inside each sensor block for platforms with multiple sensors)

# Example filters:
filters:
  - offset: 2.0
  - multiply: 1.2
  - calibrate_linear:
      - 0.0 -> 0.0
      - 40.0 -> 45.0
      - 100.0 -> 102.5
  - filter_out: 42.0
  - sliding_window_moving_average:
      window_size: 15
      send_every: 15
  - exponential_moving_average:
      alpha: 0.1
      send_every: 15
  - throttle: 1s
  - heartbeat: 5s
  - debounce: 0.1s
  - delta: 5.0
  - or:
    - throttle: 1s
    - delta: 5.0
  - lambda: return x * (9.0/5.0) + 32.0;

offset / multiply

# Example configuration entry
- platform: adc
  # ...
  filters:
    - offset: 2.0
    - multiply: 1.2

Offset adds a constant value to each sensor value. Multiply multiplies each value by a constant value.

calibrate_linear

Calibrate your sensor values by using values you measured with an accurate “truth” source.

First, collect a bunch of values of what the sensor shows and what the real value should be. For temperature, this can for example be achieved by using an accurate thermometer. For other sensors like power sensor this can be done by connecting a known load and then writing down the value the sensor shows.

# Example configuration entry
- platform: dht
  # ...
  temperature:
    filters:
      - calibrate_linear:
          # Map 0.0 (from sensor) to 0.0 (true value)
          - 0.0 -> 0.0
          - 10.0 -> 12.1

The arguments are a list of data points, each in the form MEASURED -> TRUTH. ESPHome will then fit a linear equation to the values (using least squares). So you need to supply at least two values.

filter_out

Filter out specific values to be displayed. For example to filter out the value 85.0

# Example configuration entry
- platform: wifi_signal
  # ...
  filters:
    - filter_out: 85.0

sliding_window_moving_average / exponential_moving_average

Two simple moving averages over the data. These can be used to have a short update interval on the sensor but only push out an average on a specific interval (thus increasing resolution).

# Example configuration entry
- platform: wifi_signal
  # ...
  filters:
    - sliding_window_moving_average:
        window_size: 15
        send_every: 15
  • sliding_window_moving_average: A simple moving average over the last few values.
    • window_size: The number of values over which to perform an average when pushing out a value.
    • send_every: How often a sensor value should be pushed out. For example, in above configuration the weighted average is only pushed out on every 15th received sensor value.
    • send_first_at: By default, the very first raw value on boot is immediately published. With this parameter you can specify when the very first value is to be sent. Defaults to 1.
  • exponential_moving_average: A simple exponential moving average over the last few values.
    • alpha: The forget factor/alpha value of the filter.
    • send_every: How often a sensor value should be pushed out.

throttle / heartbeat / debounce / delta

# Example filters:
filters:
  - throttle: 1s
  - heartbeat: 5s
  - debounce: 0.1s
  - delta: 5.0
  - lambda: return x * (9.0/5.0) + 32.0;
  • throttle: Throttle the incoming values. When this filter gets an incoming value, it checks if the last incoming value is at least specified time period old. If it is not older than the configured value, the value is not passed forward.
  • heartbeat: Send the last value that this sensor in the specified time interval. So a value of 10s will cause the filter to output values every 10s regardless of the input values.
  • debounce: Only send values if the last incoming value is at least specified time period old. For example if two values come in at almost the same time, this filter will only output the last value and only after the specified time period has passed without any new incoming values.
  • delta: This filter stores the last value passed through this filter and only passes incoming values through if the absolute difference is greater than the configured value. For example if a value of 1.0 first comes in, it’s passed on. If the delta filter is configured with a value of 5, it will now not pass on an incoming value of 2.0, only values that are at least 6.0 big or -4.0.

or Filter

# Example filters:
filters:
  - or:
    - throttle: 1s
    - delta: 5.0
  • or: Pass forward a value with the first child filter that returns. Above example will only pass forward values that are either at least 1s old or are if the absolute difference is at least 5.0.

lambda Filter

filters:
  - lambda: return x * (9.0/5.0) + 32.0;

lambda: Perform a simple mathematical operation over the sensor values. The input value is x and the result of the lambda is used as the output (use return).

Make sure to add .0 to all values in the lambda, otherwise divisions of integers will result in integers (not floating point values).

Example: Converting Celsius to Fahrenheit

While I personally don’t like the Fahrenheit temperature scale, I do understand that having temperature values appear in the fahrenheit unit is quite useful to some users. ESPHome uses the celsius temperature unit internally, and I’m not planning on making converting between the two simple (😉), but you can use this filter to convert celsius values to fahrenheit.

filters:
  - lambda: return x * (9.0/5.0) + 32.0;
unit_of_measurement: "°F"

Sensor Automation

You can access the most recent state of the sensor in lambdas using id(sensor_id).state and the most recent raw state using id(sensor_id).raw_state.

on_value

This automation will be triggered when a new value that has passed through all filters is published. In Lambdas you can get the value from the trigger with x.

sensor:
  - platform: dallas
    # ...
    on_value:
      then:
        - light.turn_on:
            id: light_1
            red: !lambda "return x/255;"

Configuration variables: See Automation.

on_value_range

With this automation you can observe if a sensor value passes from outside a defined range of values to inside a range. For example you can have an automation that triggers when a humidity crosses a threshold, and then turns on a dehumidifier. This trigger will only trigger when the new value is inside the range and the previous value was outside the range. On startup, the last state before reboot is restored and if the value crossed the boundary during the boot process, the trigger is also executed.

Define the range with above and below. If only one of them is defined, the interval is half-open. So for example above: 5 with no below would mean the range from 5 to positive infinity.

sensor:
  - platform: dallas
    # ...
    on_value_range:
      above: 5
      below: 10
      then:
        - switch.turn_on: relay_1

Configuration variables:

  • above (Optional, float): The minimum for the trigger.
  • below (Optional, float): The maximum for the trigger.
  • See Automation.

on_raw_value

This automation will be triggered when a new value that has passed through all filters is published. In Lambdas you can get the value from the trigger with x.

sensor:
  - platform: dallas
    # ...
    on_value:
      then:
        - light.turn_on:
            id: light_1
            red: !lambda "return x/255;"

Configuration variables: See Automation.

sensor.in_range Condition

This condition passes if the state of the given sensor is inside a range.

Define the range with above and below. If only one of them is defined, the interval is half-open. So for example above: 5 with no below would mean the range from 5 to positive infinity.

# in a trigger:
on_...:
  if:
    condition:
      sensor.in_range:
        id: my_sensor
        above: 50.0
    then:
    - script.execute: my_script

Configuration variables:

  • above (Optional, float): The minimum for the condition.
  • below (Optional, float): The maximum for the condition.

lambda calls

From lambdas, you can call several methods on all sensors to do some advanced stuff (see the full API Reference for more info).

  • publish_state(): Manually cause the sensor to push out a value. It will then be processed by the sensor filters, and once done be published to MQTT.

    // Within lambda, push a value of 42.0
    id(my_sensor).publish_state(42.0);
    
  • .state: Retrieve the current value of the sensor that has passed through all sensor filters. Is NAN if no value has gotten through all filters yet.

    // For example, create a custom log message when a value is received:
    ESP_LOGI("main", "Value of my sensor: %f", id(my_sensor).state);
    
  • raw_state: Retrieve the current value of the sensor that has not passed through any filters Is NAN if no value if no value has been pushed by the sensor itself yet.

    // For example, create a custom log message when a value is received:
    ESP_LOGI("main", "Raw Value of my sensor: %f", id(my_sensor).raw_state);