Arduinos, Valves, and Raining Patterns: Thoughts on Principles of Engineering

Principles of Engineering (PoE) is very much a do-learn course. It's really a class about integration, starting with three labs using Arduino Unos and then spending the rest of the semester on a team project that must include significant mechanical, electrical, and software systems. I was nervous about PoE going in because I was a mechE who hadn't done much mechanical design, so I wasn't sure what role I would take on the project team. I ended up having a lot more fun than I expected, and PoE contributed more than any other class this semester to my growth as a mechanical engineering. I wouldn't have expected to say this going in, but I think PoE is an important class to have in the curriculum, and I'm really glad it's a graduation requirement.

What I did in PoE:
My team project was a digital water curtain. We had sixty solenoid valves that we controlled to recreate images and patterns that people could input via a pattern grid on the Olin network. I worked on the mechanical system, speccing the valves and the pump, working on the structure design, CADing some of the components, and doing a lot of fabrication and waterproofing. None of these are final photos, but here's a CAD render and some in progress photos:

What's not shown in any of these pictures are the tubes from the top pipe down to the valves in the box with the Olin logo on it. We drilled and tapped thirty holes on each side of the pipe, putting in fittings, then connected those to the valves with plastic tubing.

It wasn't a completely successful project, but we got more than two thirds of the valves functioning and displaying the patterns that people gave it. It took us a while to fix the code so that the valves had the right response time to not stretch out the picture too much, but we eventually got it there.

Most people take PoE during the first semester of sophomore year, so the spring class is smaller and more diverse in terms of experience, with people in every graduating class and a lot of exchange students in the course. My team in particular had a sophomore (software leaning), an exchange student (industrial engineering, but mostly worked on electrical), two juniors (the other is a RoboE who worked on electrical and software), and one super senior (mechanical). I worked pretty closely with Oliver, the super senior, on the mechanical system, and I learned a lot through that. It was definitely good for me to work mainly on the mechanical part of the project while not being the mechanical lead. I was able to contribute and grow as a mechE, getting more comfort with design and CAD and working in the shop, without feeling like my lack of experience was hurting my team.

What I learned:

Beyond generally getting more experience with mechanical design, integration, and fabrication, I learned a few key things from PoE, especially the project:

1. Integration. Like I said, PoE is really a class about integration. On the labs, a lot of this integration was really just calibration, but that was nontrivial. In the project, one of our more minor problems was leaving only a 1.5 inch by 0.5 inch rectangular hole in the side of the valve box for the wires. That's really small for 120 22-gauge solid core wires, oops. Most of the problems in our final water curtain are integration issues of some kind.

2. Going from 10 to 60 isn't as easy as multiplying by 6. It certainly includes multiplying by six, but it's harder than that. Going from our prototype with ten valves to the final version with sixty meant not only building a structure for sixty valves and soldering six times as many wires but also, for example, trying to get I/O working for sixty actuators. There were similar issues with the electrical subsystem. It turns out that 40 valves would have been much easier than 60. Even if scaling up were just as easy as multiplying the time and work, that's nontrivial. I wish we'd had a larger mechanical system assembled earlier so that we could have done more electrical and software testing.

3. Waterproofing is hard. So much water. So many things to make watertight. Some of this was just a matter of fittings and teflon tape, but even that's only simple if everything is correct the first time.

4. Arduinos are cool. We used Arduinos for the labs, and then we used one for our first prototype. They're really good for prototyping because they do so many things.

Thoughts on the class

I really like the structure of PoE. The three labs are good for figuring out integration, even if they're not relevant to most projects. They were also good indications of the kinds of jobs that would need to be done on the projects, so I got to think about whether I wanted to get better at the mechE part or do something completely different.

During the projects, we had design reviews every couple of weeks, but those were more like progress reviews in front of everyone, and people from outside the class were invited. We got a lot of valuable feedback through those reviews, especially from other professors. Our software system, for example, was almost completely determined by conversations with computing profs at design reviews. On other class days, the NINJAs and Siddhartan, the professor, checked in once or twice per class period. There was a lot of guidance and a lot of opportunity to ask for help and get feedback. That was one of the best parts of PoE.