For most rookie and intermediate FIRST Tech Challenge (FTC) teams, the short answer is no: you do not strictly need dead-wheel odometry pods to score autonomous points. Built-in motor encoders combined with the internal Inertial Measurement Unit (IMU) on your REV Control Hub can reliably deliver a two-action autonomous routine—such as parking and scoring a preloaded game piece. However, if your team intends to run high-speed multi-sample cycles, execute continuous spline paths, or reliably recover from mid-field collisions, built-in encoders will fail you. The reason is not software quality; it is the fundamental physics of propulsion wheels under load.

The Physics: Why Motor Encoders Lie

Built-in encoders (such as the hall-effect encoders integrated into goBILDA YellowJacket or REV HD Hex motors) measure the rotation of the motor's internal armature before or after the gearbox. They answer one question accurately: how many times has this motor output shaft turned?

They cannot, however, tell you where your robot is on the foam tile field. The translation from shaft rotation to field position breaks down due to three distinct physical phenomena:

• Wheel Slip: Whenever your robot accelerates rapidly, decelerates, or changes direction, the wheels slip against the soft foam tiles. When a wheel spins in place for even 50 milliseconds during a launch, the motor encoder logs forward translation that never happened.
• Mecanum Vector Loss: Mecanum drivetrains move omnidirectionally by vectoring diagonal forces through passive rollers. Under sideways strafing loads, those rollers scrub and slip significantly more than standard traction wheels. Standard encoder math cannot consistently predict the true lateral displacement when roller friction varies across worn tiles.
• Gearbox and Drivetrain Backlash: Planetary gearboxes, chain slack, and belt stretch introduce mechanical slop between the encoder tick and the tyre’s contact patch. Over a 30-second autonomous routine with ten directional changes, this mechanical hysteresis compounds into errors of 10 to 20 centimetres.

When Built-In Encoders and IMU Are Genuinely Sufficient

Before spending budget and chassis volume on dead wheels, consider whether your autonomous strategy actually demands them. You do not need dead wheels if your autonomous routine matches these criteria:

• Low-speed, straight-line motions: Driving forward at 40% power generates negligible wheel slip.
• Heading correction via IMU: Using the Control Hub’s built-in gyroscope to control all turning and heading hold eliminates drivetrain rotational drift.
• Sensor resets: If your robot drives until a distance sensor detects the perimeter wall, or aligns against the field perimeter using physical squaring (bumping), accumulated encoder drift is reset to zero.
• Single-task autonomous: Delivering a preloaded element, driving over a line, and parking does not require millimetre-level trajectory tracking.

For teams in their first two seasons, spending three weeks debugging Road Runner PID tuning on unsprung custom odometry pods is often a net loss of competition points compared to building a reliable intake mechanism.

The Breaking Point: When Dead Wheels Become Essential

Dead-wheel odometry pods (also called unpowered tracking wheels) consist of small, freely spinning omni-wheels held against the field tiles with spring tension, each wired to an independent rotary encoder. Because these wheels carry no motor torque and no robot weight, they roll freely without scrubbing or spinning out.

You have reached the limit of built-in encoders and require dead wheels when you observe the following:

• You are running spline paths: Trajectory libraries like Road Runner or Pedro Pathing require continuous, microsecond-accurate updates of both $(x, y)$ coordinates and heading ($ heta$) while the robot accelerates at maximum motor power.
• You are strafing at speed: If your auto relies on diagonal or sideways travel across the center field, drive encoder estimates will drift off target by 15 centimetres or more in a single run.
• Cycle autonomous routines: If you are attempting to pick up three or four game pieces from the ground during the 30-second autonomous window, an error of 2 centimetres means missing an intake alignment completely.
• Defensive bumping or robot contact: If another robot collides with your chassis during autonomous, drive wheels will slip against the floor. Motor encoders will assume the robot moved forward, while dead wheels will measure the actual displacement resulting from the push.

Comparing Localisation Approaches

Implementation Realities

If your team decides to make the leap to dead-wheel tracking, keep two mechanical rules in mind. First, pods must be sprung downward with constant force; if a dead wheel bounces over a seam in the foam tiles, it loses tracking instantly. Second, ensure you have sufficient encoder ports available on your REV Control Hub or Expansion Hub, as each dead wheel occupies one encoder input typically reserved for a motor.

For South African teams sourcing hardware locally, you can find competition-grade robotics components and electronics through the Sheen Robotics store, or explore our competition coaching modules at Sheen Robotics FTC support to help your programmers configure motion-profiling libraries effectively.

The Verdict

Do not install dead-wheel odometry because you saw top-tier teams doing it on YouTube. Install it when your autonomous scoring ceiling is genuinely blocked by mechanical wheel slip. If your robot currently struggles to score a single preloaded element reliably, tune your mechanisms and use your built-in encoders with IMU heading control first. When you are ready to shave seconds off your cycle paths and execute multi-element autonomous routines at full throttle, dead wheels are the single most effective upgrade you can make.