CAN Color Sensor (am-5683)
Overview
The AndyMark Color & Proximity Sensor is a compact, 3.3V CAN device that reports classified color, raw RGB + clear, proximity, and ambient light. It’s designed for easy integration on robots and test rigs where you need reliable color detection (game pieces, markings, LEDs) and short-range presence sensing.Use this page to understand what the sensor is good for and how to get started. Detailed electrical limits, pinout, and mechanical drawings live on Specifications. Programming walk-throughs and Java/C++ samples live on Examples.
When to use this sensor
Detecting the color of nearby objects (game elements, indicators, tape).
Presence / approach detection at short range (e.g., “object is in the mechanism”).
Measuring ambient light to auto-adjust thresholds or avoid false positives.
Classifying states (e.g., “red vs blue” or “object present vs absent”) without external optics.
What this sensor is not
It’s not a long-range distance sensor; for that, use the AndyMark Distance Sensor.
It’s not a line-scan or camera; it gives summarized color/light values rather than images.
Highlights (at a glance)
Single sensor combining color, proximity, and ambient light functions.
Standard FRC CAN interface.
Works with common 3.3 V robot controllers and microcontrollers.
Keyed 4-pin header for simple wiring (see Specifications → Wiring for pin order).
Example code provided for fast bring-up (see Examples).
Typical integrations
Place near an intake to confirm piece acquisition and read team color.
Mount along a chute or feeder to detect jam/flow with proximity.
Face the field to classify marker colors or read LED indicators on mechanisms.
Use ambient light readings to auto-calibrate thresholds between practice and event lighting.
Best-practice tips
Distance matters: closer targets give stronger, more consistent color readings.
Turn on the LED: Check examples for how to control the LED to turn it on for better readings
Control the view: avoid direct glare from bright LEDs; a short shroud can improve classification.
Re-calibrate at events: different field lighting can shift raw values—run a quick check in the pit.
Debounce proximity: treat proximity as “object likely present” and confirm with a short moving average.
Last updated
Was this helpful?