Embedded Files
Holonomic Differential Swerve Drive
Why Build a Diffy Swerve?
Why Build a Diffy Swerve?
The goal of this project was to take the knowledge I had learned from competition robotics and use it to create something on my own. This project required me to learn and gain experience in many fields of robotics.
The goal of this project was to take the knowledge I had learned from competition robotics and use it to create something on my own. This project required me to learn and gain experience in many fields of robotics.
(Picture of a old CAD model)
Capabilities
Capabilities
The Pods that I am developing are capable of three independent degrees of freedom. What this means is that it can rotate, move horizontally, and vertically in any combination. This is in contrast to a traditional drivetrain like a car, which cannot crab or rotate freely.
Potential applications
Potential applications
A drivetrain like this - capable of moving in any direction at any orientation - has lots of real-world applications, like use as a base for warehouse robots. The drivetrain I am designing uses thrust bearings to support hundreds of pounds of weight, making it a good platform for a robotic arm. Its great mobility also means it can navigate tight spaces quickly and easily.
Img Source: MMH Staff, ed. Modern Materials Handling. Last modified August 19,2021.https://www.mmh.com/article/analyst_study_revenues_from_warehouse_robotics_to_top_51_billion_by_2030.
Mechanical Prototyping
Mechanical Prototyping
Manufacturing limitations
Manufacturing limitations
The main mechanical difficulties in this project came from the fact that I am working out of my garage, and do not have any nice machine tools. This project relied on a 3D printer, and the ability to order laser cut plates through send-cut-send. This means that I cannot go with the tried and tested bevel gear design, as I do not have the manufacturing for it.
Prototype #1 (fail)
Prototype #1 (fail)
How it was supposed to work
How it was supposed to work
3D printed Prototype
3D printed Prototype
My initial idea for the swerve drive was to have only three gears - two ring gears and one pinion connecting them. It would work by using a mismatched gear ratio to bind and prevent back driving, a huge issue with designs like this. This did not work because the device could not be driven with the ring gears, as the forces pushed the wrong way.
Prototype #2 (works)
Prototype #2 (works)
Redesign
Redesign
Physical Prototype
Physical Prototype
My second try ditched the idea of the gear ratios. It instead used what made the first one not work to its benefit, binding to prevent backdrivability. I drew sketches to analyze which way the gears would turn and settled on a design. It worked, using two motors as both turn and drive and not being back drivable.
Building a Functional Pod
Building a Functional Pod
After I finished a working prototype, I started to figure out the packaging for a real pod. I figured out where the motors would go, where I would put bearings, and other stuff.
I then moved onto the CAD, where I changed some things, like using many bearings instead of one large ring bearing.
Assembly of the first swerve module is almost complete. I ordered plates on send cut send, and printed the parts.
Last summer I spent lots of time perfecting the design of the other two pods. Even though they look similar to the first pod, these two pods underwent a major design overhaul. The construction of the pods was changed from plate and spacer to plate and standoff, and the differential tolerances were perfected. The way that the inner differential was held in place was also completely redesigned to use one large bearing instead of many small ones.
Page updated
Google Sites
Report abuse