Backyard Ice Resurfacer
Background
For my senior capstone design project, I worked with a team of four other mechanical engineering students to design and build a remote-controlled ice resurfacing device for backyard ice rinks. Over the course of six months, the team brainstormed initial concepts, performed competitor research, developed a design validated through mathematical models and simulations, and built a fully-functional prototype.
The Need
Professional ice rinks are resurfaced using Zamboni machines, however backyard ice rinks have no elegant solution for maintaining surface quality. The most common method for resurfacing backyard rinks is to simply flood the rink with water and wait for it to freeze overnight. This is time consuming, laborious, and highly inefficient. There are also a number of DIY solutions found online that mimic the function of traditional Zamboni machines, however they are not user-friendly and still require manual labor.
Ideation
I created concept sketches to brainstorm the device's overall architecture and form for internal development and presentations. The team explored different water containers, water dispersion systems, resurfacing features, and navigation methods. We ultimately landed on designing a small vehicle that would be manually driven via a remote control, equipped with a 20 gallon water tank and a mist-based water dispersion system.
Steering Model
Our device follows an Ackermann steering model, which is a type of linkage system that ensures all four tires rotate about a single pivot point when turning to minimize wheel slip - something crucial for this project in order to maximize traction while driving on ice. This is achieved by having the front right wheel turn at a greater angle relative to the front left wheel to compensate for their varied distances from a common center of rotation. I iterated through multiple CAD models for our steering system in order to optimize our turning radius to be as small as possible while establishing necessary dimensions and clearances for steering linkages and the overall shape of the chassis.
Steering Control System
I led the design of the device's steering control system, which included motor selection, electrical system design, programming a wireless controller, and integrating steering hardware into the chassis. Our primary steering mechanism was an existing rack and pinion mechanism borrowed from Northeastern's Baja SAE Club. In order to drive the pinion, I selected a 24V Nema 23 stepper motor with an integrated 15:1 gearbox. This combination provided precise control, high holding torque, and sufficient pullout torque at low RPM to turn both front wheels under maximum load (full water tank).
To power the stepper motor, I used a DC voltage step up converter to increase the output of the device's onboard 12-volt battery to 24 volts. An arduino and stepper motor driver were wired at the interface between the voltage converter and the stepper motor. I developed code that mapped zones of a joystick's potentiometer to angles of rotation of the stepper motor, allowing the wheels to turn at varying degrees according to the joystick's position.
I connected the stepper motor shaft to the pinion of the steering mechanism with a clamping shaft coupling that was slightly modified using a manual lathe. The motor and gearbox were mounted to the chassis via a water jet-cut plate and aluminum brackets I machined using a manual mill.
After validating that the motor assembly could successfully turn the wheels with a wired joystick, I adapted my steering code to work wirelessly via two radio frequency transceivers, and merged it with my teammate's code that wirelessly toggled two relays on and off to control the water dispersion system and drive motor. Though this proved to be a challenging process because of the loop-blocking nature of controlling stepper motors, the final result was a fully remote controlled prototype that could dispense water, drive, and turn simultaneously.
Testing
The team had the opportunity to test the device on campus at Matthews Arena after open-skate hours, allowing us to evaluate the device's performance in a real-use case. We also set up boundaries to simulate a maximum backyard-size rink of 2,000 square feet and explore optimal driving routes.
Results
The team was able to fully resurface a target backyard-size rink of 2,000 square feet using a 20 gallon water tank in approximately 20 minutes. The mist system filled in major surface detents and significantly reduced friction on the ice surface. The device successfully increased surface quality and user experience while reducing the amount of water, time, and effort needed to resurface ice compared to traditional backyard methods.
