In the halloween season of October 2024, I came across this video of someone seemingly flying around on a witch's broom. An idea came to me that I was positive no one else had ever had. What if you used it with Heelys? It would make it the ideal transportation device to get around campus, being compact and easy to use.
I did extensive research into propulsion sticks and found they were all very expensive, and failed kickstarters no longer selling products.
So I decided to make my own. Inspired by the idea of using it as a Witch's broom, I called it the Witch Stick.
I started the project with sketches and schematics of how I thought it might work. I decided I am going to build it around a hoverboard motor since I have 2 hoverboards at home that don't work anymore and the motors in them are fairly powerful.
I originally wanted to make this as my mechatronics project, but the team gravitated towards the robot goalie we ended up making. For months this stayed as just an idea that I didn't have much time to work on
Then came Fall semester 2026, where my friend Lucas started the Doohickey and Gadgetry Club. The purpose of the club is allow students like me a place to work on their personal project ideas with support from engineering staff and other students. The first few meetings I made a few funny Doohickeys to get everyone into the spirit of the club, then started to actually build my Witch Stick.
First meeting drawing doohickey
Second meeting doomscrolling doohickey
I started by modeling what the stick will look like in its final version. I quickly realized this was a bit of a mistake though, because as I changed things it required completely redesigning almost everything. At a certain point I decided I just needed to build a very rough prototype to understand what all the final components will be and the space needed for them.
First concept model
Second concept model with battery space
Building a rough prototype instead of jumping right into modelling was needed because I was modelling without being sure of what my final components will be. Gradually building a prototype allows me to make decisions on the fly and develop faster while identifying what works and what doesn't.
I came across a useful video on a good motor controller for hoverboard motors which I ended up using in the project. In my testing with it, I realized I could do everything I wanted it to do with only the motor controller, so I ditched the idea of using a microcontroller to simplify the system.
When looking at existing propulsion sticks, I noticed I didn't like how they were controlled and thought I could make it more ergonomic and safe. The REM stick uses a ball you twist and the KickStick uses a pressure sensitive trigger you squeeze. I saw issues with both these control methods, as twisting a ball at the top of the stick can feel awkward and tiring, and squeezing a trigger can also be tiring and awkward since you need to squeeze the stick just to hold it. If you get scared you tend to squeeze harder, which could be disaterous with a stick that can send you even faster towards danger.
My idea was to implement a thumb throttle like is used on electric scooters to control the speed, and use a trigger that acts as a dead-man's switch that needs to be held down in order to turn off the brake. That way if something goes wrong and the stick is dropped, the stick won't drive itself away. This allows the user to have a brake as well, something the other sticks did not feature.
The motor controller has an analog input option so I was going to control speed through a potentiometer throttle. It also has a brake that is active high. The circuit below shows how the analog input is given and how a switch controls if the brake is on or not.
As I am not an electrical engineer, figuring out how to power the system was a big hurdle for me. When testing the prototype I would simply use a benchtop power supply with alligator clips connected to the voltage inputs. I wanted to just buy a battery used in hoverboards since that is where the motor came from, but I realized I wouldn't have a way to charge it. Instead, I figured I could use two four cell (14.8v) Lipo batteries connected in series which would provide around the 30 volts the motor needs. I got two 2200 mAh four cell batteries and some wires that let me connect them in series and wire the output. I crimped some connectors on the wires that allowed me to add a switch and connect the wires to the board.
I also joined an electrical and computer engineering senior design team so I now had friends knowlegable in electronics that could help me throughout this project.
Made from some wood I found in the EMEC, a few 3D printed parts and all the wiring I was able to make a working prototype that proves this product will work. Below is a video of me riding it through the engineering building halls with my Heelys!
The first prototype revealed a few problems to fix.
The axle of the hoverboard motor can sometimes spin which twists the wires, eventually causing them to unplug
A breadboard will not suffice. The wire for the throttle keeps popping out and other wires have had reliability problems. I will need to make a soldered breadboard, preferably with JST connections for the throttle and brake switch.
It is a bit wobbly with only one wheel, I will probably add a second "dead" wheel for stability.
On June 25th, I tried making a new wood frame intended to stop the axle spinning by implementing the fastening mechanism the original hoverboard used. Unfortunately my attempts were not desirable as the wood I used is bad quality and not very machinable.
On July 3rd I had a new idea I quickly worked to implement. The wood frame I made was still usable and the right size so I designed a 3d printed part that would slide onto the wood frame, screw in, and it would have the axle holder. It ended up working great so I moved all the components to this new frame and also re-crimped some of the wires.
July 4th I took the Witch Stick to its first public appearance! My neighborhood has a 4th of July breakfast and "parade" where kids can decorate their bikes and scooters to ride around a loop. I decorated the Witch Stick and joined in. It was great to talk to a few people about it and show it off!
Axle holding part - succesfully stopped axle spinning and wire twisting
New smaller frame with components installed
4th of July decorations
A new problem I discovered with this prototype is that I do not have a battery protection system. During my senior design classes battery safety was drilled into us constantly due to a battery fire from a previous senior design team. I have been trying to take it seriously but did not go far enough with over-discharge protection since I figured "this is an early prototype and I shouldn't use the battery enough to over discharge it."
Well, after constant use on the fourth of July, one of the batteries over-discharged, becoming unsafe to charge again as the battery cells dropped as low as 2.77 volts when the minimum reccomended per cell voltage is 3 volts. My next version will avoid this by either using a system that reads the cell voltages and turns the power off if a cell reaches 3.2 volts, or I will put on an alarm that triggers when 3.2 volts is reached so I know to immediately stop using the stick.