[nathan_hui]

The way I have approached prototyping is to create a analog that best demonstrates a concept with the least amount of effort and materials. When I designed our team's ball pickup (as a student), I used some KB channels, 1/4-20 bolts, big sheets of plywood, old polycord, c-clamps, and other junk parts around the shop. Our team's prototype of the shooter went a little better, but not much. It was actual metal, but it was just scrap metal that happened to be about the right length (or was too long and got handed to the band saw). It did have an actual motor, but because we wanted to see which motors would have the power to do what we wanted and we didn't have the math to figure it out on paper. We have generally stayed away from the 80/20 stuff because of the price, and because it generally can't do anything we otherwise can't with scrap U-channel, plywood, or whiteboard drawings.

Once you learn CAD, statics/dynamics, and/or how to mathematically model systems, that is NOT a substitute for an actual prototype. CAD (and all of the things I listed before that) merely helps you formalize what the prototype should actually test, and this gets more into test articles rather than prototypes. Ideally, for each prototype/test article you build, you should learn something valuable from each. Otherwise, you should just go and stick that prototype/test article on your robot, since you can't learn anything new (and therefore can't improve on it).