ESA's 3D-printable ExoMy mini rover has a cute smiling face but it also has a lot of great mechanical details. One that must have taken a ton of time and effort is the thought put into designing channels so wires can be tucked away neatly. Rovers projects that use commodity tubes already have a convenient channel for wires, and I see wires were tucked inside the structural tubes of rover by Krantz. Sawppy isn't as tidy, which is partly a reflection of the my messy nature but there's also a functionality concern to be balanced out against aesthetics: ease of access for repair.

I didn't have ease of repair on my mind when I designed Sawppy, and this oversight has made my life difficult across multiple episodes of Sawppy breakdown. The most dramatic one was at Maker Faire San Mateo 2019, which was a multi-day affair far from home. Little problems accumulated because I didn't have access to my home workbench at the end of each day to address them. Sawppy limped on until mid-Sunday when the little rover popped a fuse and stopped running entirely [title image for this post] requiring an emergency work session in the field.

The first order of business of any repair was to isolate the problem, and this is where I was thankful Sawppy's wiring was easily accessible. Things would have taken a lot longer if I had to extract wire from inside structural tubes and what not. But even then I became keenly aware of problems waiting within Sawppy's electrical system. Using serial bus servos made the connection logic easy, hook up all servos electrically parallel and I was done. This made for a logically simple wiring harness, one for each side of the rover. But as Sawppy racked up mileage, I realized it was only a matter of time before a broken wire or some other electrical fault develops in one of these wiring harnesses, and it would be extremely tedious to isolate the fault location within that network of wires. This is something I hope to address in a future revision.

For the Maker Faire blown fuse, I eventually isolated it to a single steering servo that failed short. If Sawppy didn't have a fuse that blew when it did, things might have become more exciting in terms of smoking electronics. Nevertheless, this was disappointing and another strike against LX-16A serial bus servos. I carried spare servos in Sawppy's field medical kit, but that just highlighted another problem: when I replace a serial bus servo, I can't just swap in a physical component and call it done. I also had to reprogram the replacement servo with the same serial bus ID as the dead servo, and I had to hassle with steering trim of the replacement servo.

I got Sawppy back up and running after about an hour of work, sitting on a corner of one of Maker Faire's speaker areas. This isn't terrible but I want to make Sawppy problems easier to diagnose and repair. Not just for the sake of my own sanity, but also for others. I've been contacted by several enthusiastic educators who are planning to incorporate Sawppy into their curriculum, typically as a big class project or something similar. When something goes wrong on the rover, I don't want the students to get bogged down in repairs, I want them to get back to having fun as quickly as they can.

For future evolution, I wouldn't mind it if Sawppy looks a little tidier. But that will be balanced against ease of repair. If there's a potential conflict, my prioritization is firmly in favor of easier repairs. And in my own defense, exposed wiring bundles on a Mars rover is actually fairly realistic.