RM
Hub Motor
V1
RM
Hub Motor
V1
Abstract
This custom brushless hub motor capable of withstanding continuous impacts above 1 kj , spinning above 40k rpm, and slamming to a complete stop in under 5ms. Each component is machined from 7075 aluminum an mirrored to reduce production costs.
This post details the background, design, manufacturing, and validation of the hub motor along with the future of the project.
Background
In the Fall of 2023, I competed in a 1 lb combat robotics tournament hosted by the University of Illinois Urbana Champaign. My entry was an innovative design that maximized durability and control with a low-profile chassis and elastomeric treads on crowned pulleys. At the crux of the design is a high-velocity horizontal spinner with a Ti-6Al-4V cutting tooth spinning at 32k rpm and delivering 175J of energy. While the design performed well; withstanding high-energy impacts and overpowering opponents, the weapon suffered multiple failures.
Failure Analysis
The weapon was constructed from three components, a high-power drone motor, a 3d printed hub, and a titanium tooth. A captive nut and bolt within the hub created a clamping force that held the hub to the drum. Upon collisions with vertically spinning objects, the static friction between the hub and drum would be overcome, where it would axially slide along the motor until it made contact with the rail constraining the top of the motor.
Design Process
Criterion
To prevent any foreseeable failures, the replacement of the system needed to hit several criteria:
Be able to withstand forces and impacts from any loading direction
Spin above 30k rpm for 3 minutes.
Survive coming to a complete stop in under 5ms
Have a vertical height of 20mm
Manufactured with common tools for under $150
Component Selection
The two most important components that drive the design are the bearings and the donor motor. While the performance of each should be maximized, they are constrained by the overall height of the motor. The bearings should have a speed rating above 30k rpm, an internal diameter large enough to pass wires and hardware, and a minimal height. Looking through McMaster, thin-section bearings were the optimal choice for the use case. I created a large spreadsheet with many common bearing sizes. I categorized each bearing by performance and cost and graphed the curve. I decided on using 6704 bearings due to their high availability, low profile, and optimal performance.
I then created another spreadsheet containing many motors from popular manufacturers. A common outrunner wire diameter is 1mm. Leaving 0.5mm of clearance on each side and accounting for the height of the bearings allows for a maximum stator height of 7mm. The robot is powered by a 3s lipo cell which has an operating voltage of 11.1V. To reach the target rpm of 30K, a kv of 2700 is required. The combination of stator height constraints and kv narrowed the options to a handful of motors. Finally, I chose the Cobra C2204 as it had the highest maximum power of the motors.
Material Selection
The material selection of the hub is of critical importance. The clearance between the stator and rotor is under 200 µm and must remain concentric. The material must be impact-resistant and lightweight. Plastics are lightweight and impact-resistant but not stiff enough to maintain alignment. Ceramics are stiff but are not impact-resistant. This leaves metals, specifically aluminum alloys as the primary choice. There are three primary engineering aluminum alloys, 6061, 2024, and 7075. The specific strength and stiffness of 7075 was the highest and was chosen to form the main hub.
The material for the impacting tooth needed to be as hard and tough as possible to survive repeated impacts. Laser-cut AR500 steel was chosen for its material properties and low cost.
Design
Core
The core of the motor is made from two identical machined aluminum bases. They interlock through offset features that transfer shear and torsional forces. A 4mm shoulder bolt provides a bearing preload and maintains alignment
Rotor
Similar to the core, the rotor is made of two identical machined components that interlock. What started as a large base that inset into a top cap evolved. into the current design. The 6 mounting holes alternate between tapped and counterbored holes. This means the two of the same part can be offset by 60 degrees bolted to itself
Disk
The disk was initially a 1/8 in thick piece of titanium inset into the bottom component of the rotor. By changing the material to ar500 and increasing the thickness to 0.25 in, the moment of inertia was increased by a factor of 4, and allowed it to transfer shear forces across the rotor halves.
Simulation
This static simulation of the weapon disk revealed a high stress concentration at the impacting point on the disk. Potential dulling and chipping of the point is exacerbated by the heat affected zone generated by laser cutting. This weak point was addressed by increasing the fillet at the base of the tip to disperse the force.
The image simulates a 10kn vertical load on the edge of the disk. This simulates the event of a vertical spinner impacting the hub motor. The simulation shows that there would be a small amount of permanent deformation but the structural integrity of the disk will remain intact
Fabrication
In order extract the parts I needed from the motor, I designed and printed a jig. The hub was pressed out of the stator with a screwdriver and mallet. The bell was put in a drill and spun while a hacksaw was held to the seam to extract the magnet ring. Before machining the hub, I created a fully 3d printed prototype of the design. This prototype was mounted to the bot and thoroughly tested. The motor lasted the full 3 minutes and survived repeated impacts with wooden blocks. The assembly of the prototype also informed changes in tolerance that needed to be made to the final version.
After updating the design, the components were machined and assembled. The magnet ring was potted in epoxy for durability and the bearings were set in by inducing a thermal differential. The prototyping ensured that there were no clearance or tolerance issues. The motor was mounted onto the bot and repeatedly tested.
Validation
The bot was tested at the Southern California Attack Robotics January event in a tournament against some of the world's best. During the event, the motor was highly destructive and survived dozens of high-energy impacts. Several large impacts left scratches and gouges in the outer housing but such damage was expected.
The second instance the bot was tested was in the same competition that drove me to develop this motor in the first place, Robobrawl. This tournament features a modified ruleset that prevents any part of the robot except for fasteners and motors from being made of metal. In order to make the motor legal, I printed the disk out of high impact PLA and mounted it in a PLA frame. During the competition, the motor suffered no reliability issues. It continued to spin after taking massive hits that sent my 1 lb robot into the ceiling at over 20mph.
Conclusion
Overall I am very happy with the design and performance of the hub. It has proven itself to be reliable under a variety of situations and powerful enough to be competitive. There is still much to improve with the design. I want to optimize the shape of outer housing to reduce the overall weight of the design. I may also add some offset locating features to the outer housing to transfer more shear forces. Lastly, The tolerances on the bearing housing needs to be increased so maintenance is more viable in a competition.
While the motor is incredibly power dense, I believe I can further increase the power by designing my own stator to increase the KV. The eventual goal of this project is the creation of a large hub motor with supersonic tangential velocity. I believe this can be accomplished by using the rotation of the motor to spin an impeller that actively cools the core and creates an air bearing system. My next step in this direction is increasing the size of the motor to complete in the 3lb weight class.