We ended up finalizing on a design that "shot" the ball with two flywheels - Lego wheels attached to motors (similar to a pitching machine). In early prototypes, we tested both a catapult design (to dunk) and a motor design (to shoot) and were surprised by the accuracy when shooting.
To find the basket, we wrote code that would strafe along one wall of the court - and stop to shoot when triggered by an IR sensor. This worked well - integration came quickly after the drivetrain was finalized (we finished same day), and we made 8/15 shots when checking off. Pretty accurate!
For the competition, we made no modifications. Despite our batteries being fully charged, power seemed to be an issue and the robot ran slowly and affected our timers/calibration. We made no shots and were out in the first round.
LESSONS LEARNED
Brad Allen: The entire class was a great experience for me. I had taken a few intro ME courses, but no electronics, and hadn't had meaningful software experience in ~10 years. I feel like I have a greater intuition for how electrical systems work (especially signal processing!) and the importance of thinking holistically (or in an integrated manner) when troubleshooting. Power, timing, torques, loose connectors - we had just about everything go wrong, and we were better off for it.
Armelle Goreux:
Sid Kumar: I feel like we made much much more progress in those last few days before we checked off than in the previous two weeks (nothing like an early deadline to get you going). Integration is a lot harder than it sounds. Electronics are really finicky and the extra noise can throw off your motors and sensors. Everything that can go wrong, will go wrong so budget that into your project plan and workload.
Mason Shaw: The class was great but for the final week during the final project it was fairly hellish. Don't try to plan everything out at the beginning–you're just setting yourself up for disaster later when none of those plans work. Instead I recommend working on developing your design ideas initially and getting something built that actually runs around the playing area. The sooner you can do this, the easier it is to integrate components step by step.
To find the basket, we wrote code that would strafe along one wall of the court - and stop to shoot when triggered by an IR sensor. This worked well - integration came quickly after the drivetrain was finalized (we finished same day), and we made 8/15 shots when checking off. Pretty accurate!
For the competition, we made no modifications. Despite our batteries being fully charged, power seemed to be an issue and the robot ran slowly and affected our timers/calibration. We made no shots and were out in the first round.
LESSONS LEARNED
- Give yourself slack in early HW designs. We redesigned our drivetrain 4 or 5 times, and each new design cost us a lot of time since laser cutters were in short supply. A few specific highlights: make sure your holes are large enough to fit a battery molex through. Random holes are helpful, but an "accordion" design (long slats) hurts the structural integrity of the board. To manage disparity between two motors, use high-torque motors for faster acceleration.
- Cleanliness really matters. When you first start, organize your bin (or workstation) between your different components/types of components/systems. As you redesign, you will end up with a lot of scrap parts, random nuts, random screws, etc. This can make a pretty gnarly nest of used goods (and you will lose things) if it is not managed well.
- Develop good support systems. Try to give yourself the opportunity to fail as quickly as possible. Many things will go wrong, but the amount of time it will take to fix is a function of your support systems. Schematics are helpful. Designing an exploded assembly (and all components - including casters and motors) prevents small mistakes. A state diagram and pseudo-code tighten up the logic. In the same vein, have an early conversation with your coach about when and how to ask for support.
Brad Allen: The entire class was a great experience for me. I had taken a few intro ME courses, but no electronics, and hadn't had meaningful software experience in ~10 years. I feel like I have a greater intuition for how electrical systems work (especially signal processing!) and the importance of thinking holistically (or in an integrated manner) when troubleshooting. Power, timing, torques, loose connectors - we had just about everything go wrong, and we were better off for it.
Armelle Goreux:
Sid Kumar: I feel like we made much much more progress in those last few days before we checked off than in the previous two weeks (nothing like an early deadline to get you going). Integration is a lot harder than it sounds. Electronics are really finicky and the extra noise can throw off your motors and sensors. Everything that can go wrong, will go wrong so budget that into your project plan and workload.
Mason Shaw: The class was great but for the final week during the final project it was fairly hellish. Don't try to plan everything out at the beginning–you're just setting yourself up for disaster later when none of those plans work. Instead I recommend working on developing your design ideas initially and getting something built that actually runs around the playing area. The sooner you can do this, the easier it is to integrate components step by step.