The halfway point of last semester was last Wednesday. It's on the academic calendar, marked as the 33rd day. For freshmen, though, the real halfway point was not Wednesday but Thursday: Hopper Demo Day.
For several days, the studios had been set up for the demos. There was a table in the middle of the room with a black Delrin (plastic) square on it, off of which we would launch our hoppers. A tape measure ran from the table to the ceiling so that on Demo Day we'd be able to measure how high our hopper jumped. There was another tape measure on the table to find horizontal displacement of hoppers.
Before demo-ing, we all had to turn in two spreadsheets. The first kept track of all the Delrin we had used through our prototypes. The second and more interesting spreadsheet asked for the mass of our hopper, its name, and the amount of energy it stored. (Project specs said that energy could be at most three joules). That information was all put into a different spreadsheet which the NINJAs filled out as we demo-ed our hoppers. That spreadsheet had columns for whether or not the trigger worked, whether or not the hopper stayed in one piece, vertical displacement, horizontal displacement, and any other notes.
All we needed in class for Demo Day was our hopper, a print-out of our original design goal, something to write on, and a pen. The instructor handed out a sheet with everyone's names for us to evaluate each others' hoppers. First, we evaluated the challenge level of a person's design goal -- how ambitious were they? Then, we evaluated the extent to which the hopper met the design goal, ranging from "Not at all" to "Nailed it!"
Each person had three minutes to demo. In that time, we had to read our design goals, set up our hopper, and have it hop (if it could). Ideally, a student would read his design goal, set up his hopper, step back, and after at least 0.5 seconds but not too long (that's the time-delayed trigger), the toy would hop in a manner completely fulfilling the design goal. The trickiest part of that is actually the trigger. Some triggers, particularly ones based on suction cups, tended to not release very easily or quickly, some triggers were very finicky, and other toys (mine included) simply didn't have functioning triggers. People who didn't have functioning triggers or who couldn't get their trigger to behave properly for the demo manually triggered the hopping mechanism.
Most people had hoppers that did hop, even if they didn't completely fulfill the design goal. There were lots that did meet non-trivial design goals -- flipped multiple times in the air, hopped then rolled, hopped then glided, or made a noise as they hopped. There were several, though -- once again, mine included -- that just didn't hop, and I think everyone without a functioning hopper could tell you what was wrong with their designs.
So, my hopper. It didn't hop because it fell apart. It wasn't the first time that had happened, but I thought I'd done enough to prevent the hopper falling to pieces. I'm still not entirely sure of the physics of what happened, since it only happened some of the time. Basically, though, all my pieces were on a piece of latex tubing, and as I released the tubing after stretching it out as far as I could, the pieces along one half of the tubing didn't move back in towards the center with the tubing. Instead, they went flying off the tubing.
Part of the problem was that I was using too short a piece of tubing -- it wasn't that far to the end of the tubing if the pieces slid. I had thought originally that I had to use a short piece to stay under three joules of stored energy, but once I did the calculation for the spreadsheet, I should have changed to a longer piece of tubing. I stored less than a joule of energy, and with a longer piece of tubing I could have had a more stable hopper and significantly more energy. I still wouldn't have had a trigger, and my hopper still might not have hopped, but it would have stayed in one piece, and it would have had a better chance at hopping.
The design itself was very flawed. The idea was that I had a central body and then two smaller ramp-like bodies, one on each side. The two ramps would be pulled back, and then they would slam into the central body, sending it up and forward. The problem is that collisions like that are inelastic; most of the energy was going into the pieces slamming together, not into the hopping motion. My instructor warned me of this pretty early on in the design process -- when we were first doing CAD, right after making sketch models. I really wanted to see just how much the idea was capable of doing, though, because it wasn't a mechanism anyone else was using (for good reason). I was curious.
My instructor said that curiosity and wanting to explore were good reasons to go on (as opposed to not wanting to back up and do more sketches and sketch models), so I just kept going with the design. I knew it wasn't going to get impressive distance or height, but based on rough calculations and a CAD motion simulation, I expected it to hop about 20 cm.
With the design at the time, the CAD simulation was a simplification but probably mostly correct, though the ramps in the simulation were more massive than the real ones, which definitely made a difference. However, as I changed my design, even in fairly small ways, I didn't keep track of the energy I was storing, with the result that in my final design, I wasn't storing anywhere near enough energy to actually make my toy hop. With so much wasted energy inherent in the design, I couldn't afford to not have somewhere near three joules of energy stored, but I redid my calculations too late to fix that.
I'm not sure at what point I should have decided that I had done enough exploring and this design wasn't going anywhere. The first time I had something that I thought should have worked but didn't was the experimental prototype -- a week before the final prototype was due. It felt like it was too late to go back and start again, especially because I had learned enough to know that none of my initial sketches were particularly feasible ideas. From all of this, my conclusion is that I worked too linearly and didn't do enough in parallel. Once I admitted that I was continuing with this design out of curiosity, I should have also been developing some new ideas, doing new sketches, maybe even building other sketch models, since I knew that this design had such a large possibility of not working.
I made a couple of other large mistakes, and those were at the very beginning. First of all, my design goal was simply to have a toy that hopped forward and up (at an angle). That wasn't very ambitious, and as a goal it didn't really help shape my ideas at all. A more specific goal is more challenging, but it also gives guidance for the initial ideas. My other beginning mistake is also related to the idea generation process. I thought too much in terms of making sketch models, which was the next step, and not enough in terms of the materials I would actually be using to build my prototype, and I didn't think enough about the project specs. My initial ideas had no trigger mechanisms, and they were based on the idea of using rubber bands and tacky glue, not latex tubing, press fits, and heat stakes.
The hopper project was really frustrating for me, and there is very little I would do the same way were I to start back at the beginning. I learned a lot, but that's not always very comforting, particularly right after I watched my hopper explode as I tried to demo. At this point, I'm relieved to be done with hoppers, and I'm ready to move on to the next project.
No comments:
Post a Comment