Last season · 2025–26 · DECODE presented by RTX

Last Season's Robot

Every FTC season is a new game and a new robot. This is the one we built for DECODE: a rotating drum that holds three artifacts, a flywheel that launches them, and an intake that feeds the drum from the floor. Everything below was designed, built, and programmed by students.

The 2026–27 game has just been released. This year's robot gets its own section here as soon as there is something to photograph.

The robot on a workbench with its yellow drum and launcher wired up
Drum and launcher wired up for a test, late November
01

The Drum

The drum is a rotating magazine with three pockets. It started as a cardboard cylinder held together with tape, became a plywood disc inside a yellow ring, and ended up as a 3D-printed ring on a plexi face with printed brackets.

A motor with an encoder turns it. The code steps the drum one-sixth of a turn at a time, so a pocket is always either facing the intake or facing the launcher, never in between.

  • Capacity3 artifacts
  • DriveOne motor, encoder-indexed
  • VersionsCardboard → plywood → printed ring
The original cardboard drum prototype
Version one, in cardboard
A plywood disc next to the yellow drum ring
Plywood disc, cut to fit the ring
The 3D-printed drum ring held up on its plexi face
The printed ring on its plexi face
Java code for the DrumRotor class showing intakeMode and outtakeMode methods
DrumRotor: one-sixth of a turn per step
Spinning up the drum on the bench, November 1
02

The Launcher

A flywheel launcher sits behind the drum. When a pocket lines up, the artifact is pushed into the spinning wheel and leaves at a speed we can set from the driver station.

Before we built it we did the projectile math on the whiteboard: how fast the artifact has to leave to clear the goal from the distances we expect to shoot from, and how much the angle matters. The first prototype was two motors in a channel; the version on the robot uses the same idea with better bearings and a lot more zip ties.

  • TypeFlywheel
  • Motors2, on one channel
  • Tuned byWhiteboard physics, then testing
Pencil drawing of the launcher prototype from the front
The first launcher drawing, September 20
Whiteboard covered in projectile motion equations
Projectile equations, worked out by hand
Side view of the launcher mounted on the robot
Launcher from the side, November
An early outtake test
Launcher mechanism test
03

The Intake

The intake pulls artifacts off the floor and into the drum. We tried a lot of things first: twin spinning bars, a cardboard ramp, a kicker that pushes artifacts in from the side.

What worked was a row of 3D-printed compliant "cheese wheels" on a shaft, with plexi guides that we shaped by heating the plastic with a heat gun and bending it over a form. The plexi steers each artifact into the open pocket.

  • Rollers3D-printed compliant wheels
  • GuidesHeat-formed plexi
  • PrototypesSpinning bars, ramp, kicker
An early intake prototype with two spinning bars
Twin spinning bars, October
Two aluminum uprights with a bar between them
The kicker
A 3D printer printing black compliant wheels
Printing new cheese wheels
Students heating a sheet of plexi with a heat gun
Heat-gunning the plexi guides
The prototype intake, pulling in a purple artifact
04

Chassis & Odometry

The chassis is aluminum channel on four mecanum wheels, so the robot can strafe sideways as easily as it drives forward. On day one we sketched it on graph paper; by week two we had three mini chassis so more people could build and drive at once.

Two dead wheels with encoders ride along the floor and track where the robot actually is, independent of the drive wheels. That's what lets autonomous routines end up where they meant to. In November we moved the Control Hub and Expansion Hub to spots we could actually reach.

  • Drive4 mecanum wheels
  • Tracking2 dead wheels + encoders
  • ElectronicsREV Control Hub + Expansion Hub
Graph-paper sketch of the chassis with a tape measure and pencil
The first sketch, September 13
Close-up of a dead wheel odometry pod
A dead wheel, up close
The chassis with the hubs mounted in new positions
New hub layout, November 1
Neatly routed wires around the Control Hub
Wire management
Driving the chassis in the hallway, September
05

Software

The code runs on the REV Control Hub in Java. Autonomous paths use Road Runner, tuned by driving the robot along a track taped onto the floor of the middle room until the odometry agreed with the tape.

A camera reads the AprilTags on the goals so the robot knows where it is on the field, and a color check tells green artifacts from purple ones. We watch it all live with FTC Dashboard while tuning.

  • LanguageJava, FTC SDK
  • PathsRoad Runner
  • VisionAprilTags, Limelight, color detection
A laptop showing AprilTag detection code next to the robot
AprilTag detection, September 26
A student with a laptop watching the robot drive on tiles in the hallway
Road Runner testing in the hallway
A student at a laptop running FTC Dashboard beside the robot
Tuning with FTC Dashboard
A Limelight camera mounted on the front of the robot
Limelight camera, January

See it move

Practice field

Before the scrimmage we taped out a practice field across the shop floor and ran the full cycle: intake, index the drum, launch, repeat.

Read the season journal
Practice field in the shop, November 29