MATE 2019
Failed Functionality
In the final year of Subsea Rovers, we had eight members and I was once again the CEO. This year we had our website and Facebook page going at full force. By full force I mean we updated it once or twice. I headed up the physical design and used SOLIDWORKS to create CAD models and renderings which then materialized into the new, more compact frame.
Aside from the modular chassis change, we also switched from CAT-6 cables to fiber optic. There are a couple advantages to that. One, if fiber optic connections get pool water in them they still work which meant that we did not have to waterproof the terminations. Secondly, the cables were extremely flexible and the tether became extremely flexible.
We also started using network cameras (pictured left) via fiber optic instead of having separate wires for analog video feeds. Unfortunately, the latency was pretty bad and we ended up mostly using our first camera that had never failed (pictured below). It is also noteworthy that we created an image recognition algorithm using the net cams for one of the challenges.
The white microcontroller (pictured above) is an Espruino which uses Javascript instead of the C++ that we previously used. This was not a good design decision because it added control latency and bulk to the electronics suite.
The schematic to the upper left shows the final, and current, state of the ROV and details the data and power systems.
We also developed a working claw (video to the left) using a lead screw and stepper motor. I 3D printed the claw from a modified CAD model I found online and modified it. One issue I ran into during testing was the friction between the PETG pieces of the claw. This was resolved by increasing the tolerances in the CAD model and with a file. The effects of friction were exacerbated by the waterproofing of the stepper. I fully potted the rotor in a marine epoxy and got a significant decrease in stepper performance, but it still worked so what the hell it’s good enough.
The implementation of fiber optic data transmission meant that an additional waterproof housing needed to be added (pictured right) since it was already a tight squeeze in the housing (pictured below). There was definitely a better solution, but that would have required a complete re-design of the camera and control system.
