The challenge for episode 4 was to build a robot that could compete in 3 different Olympic events: the 100 yard dash, javelin toss, and the long jump. This might sound like an easy challenge, but there are some competing parameters. To have enough power for the javelin toss, this requires a bit of weight. However, for the 100 yard dash and the long jump, extra weight might hinder you there.
The design I chose for my robot was heavily based on a cockroach robot I saw at Carnegie Mellon University. Instead of wheels, it has flailing legs that are shaped like the letter "C." I chose this design because I felt that the walking system fit all three challenges: running, jumping, and throwing.
RED TEAM’S STRATEGY
Just like every other challenge, we did a lot of brainstorming that you didn’t get to see (which is probably good because it would be super boring to watch!). One idea that we, the Red Team, all agreed on was using an RC car as our chassis since it is already optimized for speed. One hundred yard dash: done! Now for the javelin toss and long jump. For the javelin toss, we were really limited by the size of the RC car chassis, so we decided that it would be our weakest event. We saw the javelin toss and the long jump as two very similar events, however for the long jump it would be more advantageous to have a light system, which is what we were limited to anyway with our RC car chassis.
Dan was actually a fairly decent leader, although ill-focused. I appreciated his positive attitude, and he did actually keep his cool during the whole build. One major thing that we did wrong was spend too much time on testing our systems. When we got the RC car, Dan tried to build a dynamometer. I still to this day don’t really understand why. It was a weird idea because there wasn’t really any way/time to hook it up to any sensors. There was no way to get any data. Also, even if there was, we never intended to alter the vehicle in any way. So that whole thing was pointless. Also, we all thought the long bow was a good idea until the guy at the bow shop said that launching anything bigger than an arrow would not go well. We should have ditched that idea right then and there, but we marched forward any. Dan spent a lot of time setting up a mount for that system and I spent at least a few hours machining a custom adapter for the bow string to accommodate the javelin. All in all, a lot of time was wasting creating and testing systems that ultimately did not need testing, which ultimately lead to running out of time on the jumping bot. It’s really sad because the time we wasting on these other things could have been spent making and testing the jumping bot.
There were a lot of things that went wrong with the jumping bot. First of all, Dan simply did not have enough time to work on it. We originally wanted to do a pneumatic jumper (see below), but by the time we realized that wasn’t possible, we had one day left. Also, Dan went into fabrication without first having a good design; it seemed like he was cutting things off just to have Joel weld them back on again. That’s an engineering lesson you learn with time: you should always start with a general sketch or 3D design before you start fabricating, otherwise you end up wasting a lot of material and time. Next, Dan does not understand the “in’s and out’s” of fabrication. The design required two sets of tubes, one set sliding inside the other. The outer tubes needed to be slotted for the spring mounts on the inner tubes. Anyone who has spent any time machining knows that cutting into metal creates burrs (these are raised edges created by the cutting process). The slots that Dan needed would create burrs inside that outer pipe that would be really difficult to remove. It would have been possible, but it would have taken time (like probably half an hour). Also, the clearance (the gap) between the inner and outer pipes was so small that any bur would have caused interference (jamming). Finally, since Dan is inexperienced in metal working, he wasn’t even cutting the metal, he was MELTING the metal so that what would have been small burrs were giant blobs of hardened steel that were even harder to get rid of. All in all, the jumping bot failed because of poor time management, poor design, and poor fabrication.
As I previously mentioned, Dan was actually melting through the steel tubes while trying to cut slots for the spring mounts. When I saw what was going on, I couldn’t believe my eyes!! There was molten steel. On the mill. And he just kept going, like nothing was wrong! But there were several things going wrong. First, he was not using any coolant whatsoever. When cutting steel, this is really important!! Unlike aluminum that conducts heat easily, steel retains the heat, so you have to add coolant (SO IT DOESN’T START MELTING!!). One indication that you are getting too hot is when your chips (the metal shavings you are creating) are coming off blue. But Dan was way past having blue chips. Also, his RPM’s were entirely too high. This will invite more heat from friction. Finally, he was using the wrong tool: a two “flute” end mill instead of a four “flute” end mill. The “flutes” are the number of cutting edges on the tool. You want to use more flutes when cutting steel because it distributes the heat and wear across more cutting surfaces SO YOU WON’T DESTROY YOUR TOOL AND MELT YOUR PART!! Basically Dan had worn the tool down to about a ¼” diameter, and I was just sure that it was going to snap and explode. Looking back, I probably overreacted (although it was evident he had no clue what he was doing) because the tool would have just fallen apart and not exploded, but I was just so worried about how negligent Dan was being with such a potentially dangerous machine, I think I just wanted to say something to get him to stop.
One thing you really didn’t get to see during this episode is the friendship that developed between Gui and myself. Although Dan was being positive and making decisions for the team (something I struggle with as a leader), we were wasting a lot of time, which was really frusterating for Gui and myself. Overall we came together with our mutual “enemy” and have gotten along ever since!
Originally, we wanted to go with a pneumatic jumper bot, like a smaller version of the Blue Team’s jumper. Gui has a lot of expertise in pneumatics. He played a major role in building Big Dog, a project from Boston Dynamics (a super awesome robotics company – look it up!). Dan assigned Gui the task of designing this system. The problem was that the parts he needed were not stock items in McMaster or Grainger. When Gui told Dan this, Dan was like, “Man, I really need you to get that working.” So Gui continued to look for parts and work on the problem. Again, Gui told Dan, “Look, I can’t get the parts, we need to look at another design.” Again, Dan said, “I need you to get this working!” It was as if Dan didn’t believe that Gui was giving his all, that all Dan needed to do was push Gui a little harder to get what he wanted, which wasn’t true at all. Ultimately pushing Gui to work on a task that wasn't acheivable cost our team a lot of time, time that could have been spent on a funcational jumper bot.
Some of the funniest things that happened during the show occurred while the cameras were not rolling! There was one point during the challenge day that we couldn’t remember the guest judge’s name, and Gui, being the clever person that he is, started talking to the audio guys through his mic (we were mic’ed pretty much anytime we were not sleeping). He was like “Hey… hey… audio guys… hey… what’s the guest judge’s name… hey… hey.. audio guys..” We could see one of the audio guys smiling from across the set but started shaking his head, but then our favorite camera guys (Loy) came over and saved us: “His name is Jason.” Okay, might not be the funniest thing you ever heard, but it’s one of my favorite memories of the show!