Calculated Channels

Process live telemetry with calculated channels.

Calculated channels compute values from other channels in real time:

Scale, convert, or filter raw data
Implement sensor voting algorithms
Trigger alarms or warnings from a condition

Calculated channels are virtual: their data is computed on demand and never stored to disk. Because nothing is stored, an edit applies to all historical data. A channel changed from chan * 3 to chan * 10 shows chan * 10 for its whole history.

Calculated Channel Parameters

Table

Parameter	Type	Default	Description
`name`	string	Required	A name for the channel.
`expression`	string	Required	The Arc expression that calculates the value to be written to the calculated channel. This expression must end with a `return` statement. Channel dependencies are automatically detected from the expression.
`operations`	list of `Operation`	`[]`	Post-processing operation applied to the expression result. Options include `min`, `max`, `avg` for running aggregations, and `derivative` for rate of change.

Operation Parameters

Table

Parameter	Type	Default	Description
`type`	string	Required	Options include `min`, `max`, `avg` for running aggregations, and `derivative` for rate of change.
`duration`	`TimeSpan`	`0`	How long the window runs before the operation state resets. Set to 0 for no timed reset.
`reset_channel`	number	`0`	A `boolean` channel that triggers operation reset when its value is `true`.

Create and Edit a Calculated Channel

Console

Python

TypeScript

# Create a calculated channel
avg_pressure = client.channels.create(
    name="avg_pressure",
    data_type=sy.DataType.FLOAT32,
    expression="return (pressure_1 + pressure_2) / 2",
    operations=[sy.channel.Operation(type="avg", duration=10 * sy.TimeSpan.SECOND)],
)

# Upsert to edit a calculated channel
avg_pressure.expression = "return pressure_1 * 2"
avg_pressure.operations = [
    sy.channel.Operation(type="max", duration=10 * sy.TimeSpan.SECOND)
]
client.channels.create(avg_pressure)

// Create a calculated channel
const avgPressure = await client.channels.create({
  name: "avg_pressure",
  dataType: DataType.FLOAT32,
  expression: "return (pressure_1 + pressure_2) / 2",
  operations: [{ type: "avg", duration: TimeSpan.seconds(10) }],
});

// Upsert to edit a calculated channel
await client.channels.create({
  ...avgPressure.payload,
  expression: "return pressure_1 * 2",
  operations: [{ type: "max", duration: TimeSpan.seconds(10) }],
});

Operations

Operations are optional steps that apply a running aggregation, such as a minimum, maximum, or average.

The operation keeps state across executions and outputs a single aggregated value.

Supported Operations

Table

Operation	Description
`min`	Tracks the smallest value seen since the last reset.
`max`	Tracks the largest value seen since the last reset.
`avg`	Computes the mean of all values since the last reset.
`derivative`	Computes the derivative (dx/dt) per second of the expression result. Always outputs `float64`.

The derivative operation does not support window or reset channel configuration. It maintains only the previous value and timestamp, so there is no accumulated state to reset.

Example Configurations

Table

Use case	`type`	`duration`	`reset_channel`
Rolling 10-second average	`avg`	`10s`	none
Maximum until manual reset	`max`	`0s`	`manual_reset_btn`
Minimum, time or signal	`min`	`60s`	`cycle_complete`
Rate of change	`derivative`	n/a	n/a

Reset Channels

A reset channel provides signal-based control over operation state and must have the data type boolean. When it receives true, the operation clears its state and restarts.

Manual: write to a boolean channel from a schematic button or control panel.
Periodic: use a timer or sequence to generate reset pulses.
Conditional: use a calculated channel that outputs true when a condition is met.

Reset Example

Table

Time	Pressure	`max` output
0s	50	50
5s	100	100
10s	75	100
10.1s (`reset = true`)	75	75
15s	90	90
20s	110	110

At 10.1s, the reset channel triggers, clearing the max value. The operation restarts from the current input (75) and continues tracking the new maximum.

Writing Expressions

Expressions are written in Arc. Reference channels by name, and end the expression with a return statement. The output data type is inferred from the return value. Channel references are series, not scalars, so return temperature + pressure is an elementwise operation on arrays.

scaled := pressure * 2.5
return scaled + 10

Table

Use case	Expression
Scale a sensor	`return pressure * 1.5`
Convert Celsius to Kelvin	`return temperature + 273.15`
Power	`return voltage * current`
Sum sensors	`return sensor_1 + sensor_2 + sensor_3`
Average	`return (temp_1 + temp_2 + temp_3) / 3`
Differential pressure	`return inlet_pressure - outlet_pressure`

See the Arc reference for operators, variables, and if statements.

Channels

Ranges