FTC Dashboard & Live PID Tuning

FTC Dashboard is a web page served by the robot. It lets you edit public static fields while an OpMode runs, graph telemetry, and draw on a field view. Official docs are at acmerobotics.github.io/ftc-dashboard.

To open it, connect to the Control Hub’s Wi-Fi and go to http://192.168.43.1:8080/dash.

Installation

The Road Runner quickstart already includes it. For a plain FtcRobotController project, open build.dependencies.gradle in the project root and add the repository and the dependency. Get the current version number from the getting started page.

repositories {
    maven { url = 'https://maven.brott.dev/' }
}

dependencies {
    implementation 'com.acmerobotics.dashboard:dashboard:VERSION'
}

Then sync Gradle.

Tunable variables

A field shows up in the dashboard when:

1. The class is annotated with @Config.
2. The field is public static.

import com.acmerobotics.dashboard.config.Config;
import com.qualcomm.robotcore.hardware.PIDFCoefficients;

@Config
public class LiftConstants {
    public static PIDFCoefficients PIDF = new PIDFCoefficients(0.05, 0, 0.002, 0);
    public static int TARGET_TICKS = 800;
}

Static fields belong to the class, not an instance, so when the dashboard writes a new value the running OpMode reads it on its next loop.

PIDFCoefficients is an SDK class with p, i, d, and f fields. The dashboard shows each one as its own editable value.

Put the constants class in its own file. Keep all tunable values in one or two @Config classes so there is one place to look.

Telemetry to the dashboard

FtcDashboard.getInstance().getTelemetry() returns a Telemetry that sends to the dashboard. MultipleTelemetry sends to both the dashboard and the Driver Station at once.

MultipleTelemetry tel = new MultipleTelemetry(telemetry, FtcDashboard.getInstance().getTelemetry());

Numeric values sent this way are graphed in the dashboard’s graph view.

Example: lift tuner

package org.firstinspires.ftc.teamcode;

import com.acmerobotics.dashboard.FtcDashboard;
import com.acmerobotics.dashboard.telemetry.MultipleTelemetry;
import com.qualcomm.robotcore.eventloop.opmode.LinearOpMode;
import com.qualcomm.robotcore.eventloop.opmode.TeleOp;
import com.qualcomm.robotcore.hardware.DcMotor;
import com.qualcomm.robotcore.hardware.PIDFCoefficients;
import com.qualcomm.robotcore.util.ElapsedTime;

@TeleOp(name = "Lift PID Tuner")
public class LiftPIDTuner extends LinearOpMode {

    @Override
    public void runOpMode() {
        DcMotor lift = hardwareMap.get(DcMotor.class, "lift_motor");
        lift.setMode(DcMotor.RunMode.STOP_AND_RESET_ENCODER);
        lift.setMode(DcMotor.RunMode.RUN_WITHOUT_ENCODER);

        MultipleTelemetry tel = new MultipleTelemetry(telemetry, FtcDashboard.getInstance().getTelemetry());

        ElapsedTime timer = new ElapsedTime();
        double lastError = 0;
        double integralSum = 0;

        waitForStart();
        timer.reset();

        while (opModeIsActive()) {
            double dt = timer.seconds();
            timer.reset();

            int currentPos = lift.getCurrentPosition();
            double error = LiftConstants.TARGET_TICKS - currentPos;

            integralSum = Math.max(-1, Math.min(1, integralSum + error * dt));
            double derivative = (error - lastError) / dt;
            lastError = error;

            PIDFCoefficients c = LiftConstants.PIDF;
            double power = c.p * error + c.i * integralSum + c.d * derivative + c.f;

            lift.setPower(Math.max(-1, Math.min(1, power)));

            tel.addData("Target", LiftConstants.TARGET_TICKS);
            tel.addData("Position", currentPos);
            tel.addData("Error", error);
            tel.addData("Power", power);
            tel.update();
        }
    }
}

c.f here is a constant added to hold the lift against gravity. LiftConstants.PIDF is read every loop, so changes in the dashboard apply immediately.

Tuning

1. Set i, d, and f to zero. Raise p until the lift reaches the target with a small oscillation.
2. Raise d until the oscillation stops.
3. If the lift settles below the target, raise f until it holds.
4. Leave i at zero unless there is still a steady error after f.

Watch the Error line in the graph. The goal is a fast drop to zero with no bounce.

Graph	Cause	Change
Slow approach, never quite reaches	p too low	Raise p
Oscillates around target	p too high	Lower p
Overshoots then settles	d too low	Raise d
Jittery power	d too high	Lower d
Settles short	Gravity	Raise f

Values are not saved

Dashboard edits live in memory only. When the robot restarts their gone. Copy the final numbers back into the source file.