This project implements a controller and web interface for treadmills that use the MC2100LTS-30 motor controller board (Icon part number 263165). This likely includes many Icon Fitness treadmills under various brand names such as Epic, NordicTrack, Proform, Reebok, and probably others. Use the part search form at Icon Service Canada or Icon Service to search for part number 263165 to see if your model might work.

The long term goal is to allow for control of treadmill speed and incline, easy construction of training programs, and recording fitness data, including heart rate (via a USB ANT+ dongle).

Here is a very basic schematic for hooking the Pi up to the controller board. There is no need to cut anything to do this: the cable linking the console and the controller should have a connector at the top of the leg holding the console. The Pi and breadboard fit nicely in the tray between the legs.

I can't remember why the speed and incline sensors have different filter capacitors and resistors, they likely should be the same. Get you 'scope out!

The schematic can be edited in Fritzing.

This project is a suite of three pieces of software.

The controller program is written in C++ and is responsible for interacting with the GPIO pins. It has a minimum of logic, being responsible only for calculations directly related to controlling the speed and incline of the treadmill, as well as features related to safe operations of the hardware, such as ensuring the key is in the console before allowing operation and stopping the motors if no feedback on their motion is available.

The controller program uses ZeroMQ to communicate with the web server.

The webserver is currently written in Python using Flask. This part is responsible for serving static resources such as the UI, storing information in a database to facilitate the construction of training programs and recording runs.

The front end is written in Elm. Unfortunately the Elm compiler is not available on the Raspberry Pi, so it is necessary to build the frontend on a Windows, Mac, or Linux machine.

A big chunk of this project is about learning Elm and seeing what happens when an Elm app gets big, so the frontend is doing way more stuff than it should. A lot should be moved to the other components.

Install the required dependencies:

sudo apt-get install wiringpi libzmq3-dev python-zmq
sudo pip install wiringpi2

The C++ controller can be built like this:

cd controller
make

This produces controller and fakecontroller programs. The controller program will use WiringPI to drive the GPIO pins to actually drive the treadmill's motor controller. The fakecontroller accepts commands over zmq but returns fake responses.

The UI can be built on a machine that the Elm 0.18 tool-chain can be installed on. The easiest way to install Elm 0.18 seems to be to use Node v8.17.0 and NPM 6.13.4. nvm is a nice way to manage multiple versions of node. Building a prod version of the JS file using closure apparently requires yarn, but you can install that with npm as well. This installs them both globally, which may not be desirable if you want the most recent version of Elm installed globally for other projects:

npm install -g elm@0.18.0-exp5 yarn

You can instead install yarn and elm locally in the frontend directory:

cd frontend
npm install elm@0.18.0-exp5 yarn
export PATH=$(pwd)/node_modules/.bin:$PATH

Then you can build the front end using make:

cd frontend
make

This will build a JS file and copy it into a directory where it can be served by flask. You can also build a minified version for production:

make prod

This project currently uses Flask for it's web server. Flask can be set up with the following commands:

cd server
pip install virtualenv
virtualenv venv
. venv/bin/activate
pip install flask pyzmq gevent Flask-Sockets
deactivate

Once that is set up, you can run the web server using the provided run-server.sh script.