Hi, I’m Austin Li

Objective

I have always been intrigued by motors (electric motors in particular), the ability to turn electricity into motion seems magical, and yet the way motors are so ubiquitous today turn them into this seemingly simple thing which we take for granted. Motors are literally everywhere, everything from linear motion to water-pumps and turbines often start as rotation driven by motors, and without them the world would be a much more stationary place (you could almost say motors make the world turn…).

I wanted to know how motors work, so I built one entirely from scratch. Everything from the stator windings to the magnetic rotor, I specced, designed, built, and tested. Moreover, to make it more challenging and to squeeze a little bit more efficiency out of it, I designed the motor specifically for Field Oriented Control meaning the need for a magnetic encoder, distributed windings, and large arc magnets.

Being a personal project, much of this build was constrained by budgets, but through various attempts to compromise and de-scope, I was able to assemble a first prototype which is currently being tested. The ultimate goal of this project is to use the motor in a larger DIY OneWheel build which has been a dream of mine for a while.

Design

I began by defining my constraints then modeled everything in SolidWorks, since this first prototype would be largely 3D printed, everything was designed with a heavy focus on DFM for FDM printing.

Field Oriented Control

I wanted to use field oriented control (FOC) to drive this motor. This meant two things that needed to be different from a typical motor: encoder to detect position, and large arc magnets in the rotor.

Finding a place to put a magnetic encoder proved difficult because it meant needing to split the shaft to make the encoder concentric with the motor, however, supporting the stator from only one end would severely compromise its rigidity which is a concern as it experiences extreme magnetic forces.

I designed a way to place the encoder inside the shaft by splitting the shaft and decoupling it from the rotor using bearings.

Wide angle arc magnets are vital for creating a sinusoidal back-EMF profile. However, large arc magnets are hard to find in the proper dimensions, which makes custom ordering the only option, but this obviously greatly exceeds my budget for this project.

To compromise, I approximated arc magnets using smaller block magnets placed adjacently. Block magnets are much cheaper.

I validated that this works by using magnetic sensitive sheets to visualize the magnetic fields, and indeed the magnetic field appeared uniform in the regions they should be.

Back-EMF Validation

Using an oscilloscope I could verify that the back-EMF from spinning the motor was roughly sinusoidal. The 6 spikes seen can be attributed to the 6 individual magnets in each arc cogging with the stator.

Objective

Before starting this project, I had no experience in PCB design, at all. I had never touched a PCB software nor taken any advanced circuits courses. However, knowing how universal the applications of printed circuit boards are, I wanted to learn the process of how they are designed.

So, I designed FORC, my attempt at a high-power ESC specifically designed for Field Oriented Control. The software I used was Altium, which I chose over other beginner friendly choices like KiCad because it is more of an industry standard. My design is based on the Vedder ESC architecture, which is the gold standard for speed controllers in the application I am looking for.

The ultimate goal of FORC is to use it to drive my DIY OneWheel build, which is why it required specific components like an IMU and FSR sensor.

Objective

The Autonomous Pool-Playing Robot was built as a group project for my final Mechatronics Project of first year. The robot was built around standard VEX components, and our goal was to push these components to the limits of their capabilities.

My group and I designed, coded and tested an omni-directional robot which could successfully locate and strike a pool ball into a pocket.

My contribution to this project was in the mechanical design of the striking mechanism and the ball detection algorithm I wrote in C++.

Ball Detection Algorithm

The ball detection algorithm was constrained by a single laser distance sensor which forced out-of-the-box thinking to be able to locate the ball. To achieve both efficiency and accuracy, I split the algorithm into two parts:

Striking Mechanism

After many revisions, the final design I settled on used a scissor mechanism to ensure the front face stayed perpendicular with the forward direction.

Full Video Demo

Replay 1×

Objective

AutoCleat is a shoe attachment that automatically deploys ice spikes when winter surfaces become unsafe. It was designed to solve the awkwardness of traditional cleats, which work outdoors but become inconvenient or damaging when walking indoors.

AutoCleat was built in 24 hours as a submission to the Waterloo Enghacks Hackathon.

The prototype uses temperature sensing to detect icy conditions and an ultrasonic foot-lift sensor to deploy TPU spikes only when the foot is raised. The spikes retract on safe surfaces, giving the user hands-free traction control without stopping, bending down, or manually switching modes.

Replay

Objective

This cycloidal gearbox was designed to be a compact reducer for use in robotic joints. The final dimensions measure 22mm in thickness.

Objective

I created an accurate model of a transport trailer in CAD (SolidWorks) which proved to be mission critical for designing fixtures to transport the Team’s rocket to the launch site.

Explore the model

Loading model…

Drag to orbit · scroll to zoom

Reset view