Andrew Huang's Final Reflection

For me, ME250 was an interesting experience. I learned a lot about designing via CAD software and then trying to go about manufacturing those designed parts. I gained plenty of hands-on experience from all the time spent in the shop, and got some insights about the principles that mechanical engineers must include in all their designs. Furthermore, I learned a lot by just asking the more experienced staff questions about machining and design.

As far as the design aspect of engineering goes, I learned the most about appropriate tolerancing and avoiding overconstraint. For example, our design used running fits and press fits- the hardstop on our telescoping arm was press fit for simplicity’s sake, the axles were all made to running fits. Often we would have to check the dimensions of the stock given to us and machine it down to the appropriate dimensions for a running fit. Also, as a result of the design of our drivetrain, the rear axles were terribly overconstrained. Each axle had two pillow blocks supporting them (as they should) but these axles were mounted via a set screw to the output axles of the double gearbox. This meant the drivetrain was constrained in 6 places- 2 for each axle, and then twice for the gearbox mountings. This made manufacturing a pain, but the parts given necessitated such overconstraint. In other parts of the robot, we avoided it like the plague.

As far as the manufacturing aspect goes, I learned a lot about the tools and capabilities of the mill and the lathe. (I also learned how much I missed my dremel, but that’s a different story.) I had no previous experience with the lathe, but I learned how it is absolutely vital in making axles to the right dimensions, cutting grooves for e-rings, and most importantly, drilling centered holes through shafts. Some more subtle things I learned about the lathe was its use in chamfering the ends of axles (with a file) and how to get a good surface finish on parts (very slowly). I learned similar lessons on the mill- how critical it was to face off parts and then acquire the datum lines, how useful it was in drilling holes with awkward distances from said datum lines, and its utility in removing material that would have been impossible/impractical to remove with a saw or file.

As far as teamwork and time management goes, I feel like our team did well in both aspects. We were fairly cohesive and worked together well, and planned work sessions far ahead of deadlines. There wasn’t much to be improved or learned in this section – by now all of us have well-developed time management and interpersonal skills.

I feel that the layout of ME250 could be significantly improved. I believe that lab time and lecture time did not synergize, and as a result the class felt unstructured and valuable time was wasted. For the scope of the project in this class, especially given the fact that students are not yet expected to have any design or manufacturing experience this early in the curriculum, the structure and cohesion of the class is a critical part of how the competition turns out. There are several key points of improvement that stand out to me:

1. Assign groups first; this class is 99.99% about group work and to delay assigning groups is to delay the progression of the class. Group members can still come up with individual concepts and strategies if asked to.
2. The initial reviewing of concept and strategy selection should include a section on the practicality and likelihood of success of the strategy. For example, if a team proposes blocking the flipper with a screw-driven arm, the GSI should note that screw drives are typically very slow (especially with the motors given) and that another team will absolutely beat them to the flipper. This is not to necessarily discourage creativity on the part of the teams- merely to make sure teams know that their design is probably impractical. (As a side note, they might point out that robots with four wheels do not turn if the wheels don’t have a steering mechanism.) Ideally this step of critiquing designs would occur very early on in the class, giving teams time to come up with different concepts and designs. To encourage risk-free creativity, the concept should not be graded until it has gone through several revisions.
3. Lecture should be more focused on the design principles relevant to the Slotbots game. For example, lecture topics should encompass topics such as “drivetrains” instead of just “bearings.” A lecture about drivetrains could address specific difficulties in designing one- including overconstraint, manufacturing components, where one might use different types of bearings in a drivetrain, how one would go about getting the most power out of the drivetrain, etc. This would help teams create more robust designs, and make the class more competitive as a whole. Ideally concepts in these lectures would be looked for (not required) by the GSI’s when reviewing concept and strategy selections.
4. Because a significant amount of the learning in this class is based on hands-on application of engineering concepts, the times for lecture and lab should be reversed. Lab should be 1.5 hours and lecture should only be 1. This gives more time to learn how to use CAD software, how to machine (especially the finer points of machining which can’t really be addressed in lecture), etc. This would also potentially leave time for teams to get manufacturing practice in after training and before the robot construction phase begins.
5. Rules and the design of the arena should be set in stone before the class begins. Every change of the rules affects the validity of a team’s design. To have to keep changing the concept to accommodate rule changes halfway through the semester is a particularly frustrating experience.
6. As far as the kit goes, changing the wheels to a material which actually grips the arena surface readily would save a lot of teams a lot of hassle. Furthermore, having square angle stock versus rounded angle stock would make designing pillow blocks significantly easier- teams would not be so tempted to cut up their square/rectangular tubing as material for pillow blocks. Also, issuing 3/8” bushings would be nice, considering that 3/8” round stock is provided.
7. On the arena, making the ball-holders flush with the carpeted surface would have made the game significantly easier. From what I saw at the competition, no robots could consistently make it over the centerline using the stock wheels and the stock motors- the robots would inevitably get caught or simply lack the power the drive over the ball holding plate in the middle.
8. In any discussion of drivetrains, I would make a bigger deal of using the caster/including different types of casters in the kit. These robots are supposed to be simple, and there are only a few ways to make a robot turn using two powered wheels. These include a steering system (impractical at this level of complexity), a 6-wheeled arrangement where the middle wheels are offset (complex), and a tricycle-like arrangement where one "wheel" is actually a caster/slip plate. There may be other options for surface-based robots to move around, but these seem like the most successful options to me. Four wheeled designs simply cannot turn effectively without a steering system - these should be discouraged for this project.

 As far as my own performance in this class goes, I could have put more time and effort into the design section of the class. I went to our team-scheduled design meetings and contributed ideas, but a large section of the CAD was done by other members of the group. I feel that I made up for this by spending a significant amount of time and effort during the manufacturing section, but I would have learned more about using Solidworks had I spent more time in the design phase.

All in all, my ME250 experience was a stressful and time consuming one, but rewarding as well. Would I go through it again? Not willingly. Do I appreciate everything I learned? Definitely.