On the CWRUBotix MATE ROV team, we build and compete with our ROV each year. There are missions core to the competition that often require complex manipulation of objects such as valves and simulated power connectors.
Actuators are core to our ability to accomplish tasks, and we have been quite limited by the cost and availability of actuators that are fully waterproof at depth.
Many commercial waterproof actuators were either not suitable for continuous submersion, or were designed for hundreds of meters of depth adding cost and size. I decided to built a custom servo for us to use on our ROV.
I began by working with the team to identify key functional and non-functional requirements:
0.5 Nm of torque at 100 rpm minimum
highly compact footprint and mounting
reliable thermal, mechanical, and electrical integrity over several dozen hours of operation
final cost $60 or less per unit, plus several hours of machine and assembly time
I then went through several rounds of research, sketches, and concept CAD before getting into design and manufacturing.
I arrived at a design that encloses a DC motor tightly in an aluminum tube, onto which machined flanges are bonded to interface with the output and sealing components.
The motor's output shaft is also replaced with a custom stainless steel driveshaft to avoid carbon steel corrosion.
At the start of this project I had little CNC manufacturing experience, so I taught myself Mastercam and learned to use my university's Haas MiniMill.
I machined a series of modular Mitee-Bite Pitbull fixtures that were used to manufacture nearly every part from waterjet blanks, with high material yield and nearly zero-post processing operations.
In addition to learning CNC machining myself, I documented my process so that others on the team could learn, as no one on the team had CNC milling experience.
Throughout the process, I incrementally tested alignment, sealing, and motor temperature.
The servo successfully operates at low depth continuously for several days.
However, at increased depth the rotary seal slowly leaks, which required improved testing methodology. To achieve this, I built a hydrostatic testing chamber and rotary seal testing assembly. Additionally, the custom PCB and closed-loop control is still in progress. Given the 2026 competition timeline, we intend to implement the custom servo into our 2027 robot and are continuing development.
Despite these delays, the project has still improved the capabilities and knowledge base of our team tremendously. We now have hundreds of hours of CNC milling experience, custom fixtures, and efficient manufacturing workflows.
This project proved the viability of ultra-compact gaskets and aluminum components, integrated ball bearings, rotary gaskets, and compact epoxy penetrations. Enhanced testing techniques, such as color-changing water detection strips, were proven out.