Design Process: 2012 Shooters

[IanW]

Unfortunately, my team (FRC 2374) did not find as much success as some of the more experienced teams in designing a shooter. However, the process we used in developing our shooter resembled a proper design process much more closely than we have ever done before.

We essentially started out by brainstorming different methods of getting the ball in the hoop (First mistake: not deciding on a clear strategy and generating design requirements from it). We quickly narrowed down our ideas to a wheeled shooter, and two different types of catapults. Based on our review of previous games/the "Behind the Design" books, we already had an idea that the wheeled shooter would probably have the best chance of success. Still, we began prototyping all three shooters. We soon abandoned the catapult based prototypes and began focusing on the pitching wheel idea.

We only worked on single axle shooters mainly because they are simpler mechanically, there is one less variable to control, and because we had read that backspin would be helpful. Our first prototype barely worked, as we tried to make it too much like how we envisioned our end product instead of using it to determine what worked best. We then created a much more basic prototype that worked much better, but from which we gathered no real data (Second mistake: making prototypes and not really learning anything from them. In this case, it was failing to gather data about how changing each variable involved with building a shooter affected its ability to shoot).

At this point, we had a better idea of what capabilities we wanted our shooter to have - turret, adjustable launch angle, and capable of shooting from at least the key (We didn't really have good reasons for wanting these features, they were just what we wanted; see "First mistake"). From this, we made one more prototype that was essentially just a proof of concept to show that our packaging would work. Again, we didn't really learn anything from this prototype. Based on this prototype, I designed our final shooter with CAD, and then proceeded to build it.

From when our shooter was mounted to our first regional, we pretty much only iterated on it by changing how we mounted sensors. Before we bagged, we removed the shooter so that we could use it with our practice chassis for further development. After our first regional, we realized that our launch angle control was superfluous and therefore removed angle adjustment for our second one. Though I believe our shooting percentage increased, this probably had as much to do with our completely-rookie driveteam becoming more comfortable as it did with our shooter being more accurate. One of our main limitations was actually ball collection - our intake was not very effective, resulting in us being unable to even try very many shots.

I hope this explanation is of interest to someone. Unfortunately, I am not knowledgeable about our software development process, so I cannot share about those iterations. Hopefully I use these lessons learned if I mentor a team in college.