With the MCM robot done and running, I moved onto building the Rampbot.
This was the part of the concept 9 design that deployed outside the starting box and allowed the MCM robot to climb the ramp.
This ramp needed to start inside the starting box (blue lines), and then extend to the correct position as soon as the match started.
So, I needed the ramp to move about an inch to the left, as soon as the match started. I thought about this for a little while and realized that the best way to do this was with a four-bar linkage. This would be the optimal way to do this motion without fail. The stable base would also allow many pounds of weight to be added to the lower base, ensuring that it would never move even if hit by a falling cannonball.
I threw in a random 4 bar thing into the CAD.
The building of the CAD went much smoother with this than with the ramprunner robot, probably because I had been thinking about it for a while. The one small challenge was adapting the four-bar linkage base to the curved ramp. Basically, I just drew in random connectors until the geometry worked. It also became symmetrical to operate on both sides of the field.
The motion was driven by two big servos in parallel (for redundancy). The servos were placed on the upper half and dragged themselves over to one side or the other based on which side the robot was running on.
Back view illustrating the linkage in the vertical position. It fit in the starting box by about 1/4″.
The building of the rampbot went farily smooth. The robot went together extremely quickly, as it consisted of only about four machined parts. Some parts had to be glued together though, so the process took a few days.
The parts after laser cutting.
aaaaannnnddddd…. Somewhat assembled!
One half assembled with Rampbot C9.1V2 onboard. Picture taken in MITERS.
I tested this and it worked on the first try (dam sun dis guud). It set a score of 418 points, twice what anyone else had scored at that time. With this running, I decided to take a brief break and sidetrack myself for a bit.
The Saga of Switched Reluctance
With the ramp-bot almost done and over two months before competition, I decided it was time to try something really far out there.
Many 2.007 alums I talked to had experienced problems with their motors either burning out or stripping their gearboxes. After thinking a bit (3 seconds or so) I realized it might be worth the time to actually try and build my own motor, with the parts available. Hopefully, I could create something both more powerful and more reliable than the BO-P5 motors.
Being a brushless hipster I knew the motor would be some brushless outrunner contraption. The most important component of this type of motor is neodymium magnets, which are not specifically provided in the kit. I harangued the lab instructors and dr. Wendell about magnets a bunch, with varying degrees of success, but none of them promised the 14 magnets I would need to build a nice 12n14p motor. The ultimate decision was that we couldn’t use this many magnets, so conventional BLDC is a no-go. Crap. Well, dang, looks like that’s the end of that.
So I made a final hail-Mary, and just googled “motors without magnets.”
Surprisingly I found tons of results. The object most prominent was the induction motor, a magic black box I had heard of before, followed closely by the switched reluctance motor which I’d never heard of at all. I did a bunch more searching and paper reading and realized that not only were switched reluctance motors a thing, but that they were quite simple and probably even possible to build with the 2.007 kit. Some nice features of switched reluctance motors:
Similar to in concept to BLDC motors which I had experience with
Only need low-side fets
I couldn’t find quite as much literature as I wanted, or pictures of motors others had built, so I just read a lot of papers. They proved very interesting and held concepts pertaining not only to switched reluctance but to motors in general.
Here was the first one I cadded.
Initial designs featured a fairly conventional switched reluctance motor, in which the flux path circles the entire outside of the motor. Blue lines are flux loops.
I then switched to a different design which makes use of adjacently wound teeth to create smaller flux loops. Theoretically, this design features a lower Kv and therefore more torque. I hoped to use as few years as possible on the robot as the gears provided in the kit are plastic.
A switched reluctance motor is similar to a BLDC motor in that it requires commutation, and thus some sensors to determine rotor position. Initially I planned for some HFI scheme but Ben talked me out of this because it would be hard. Thus I had to devise some sort of sensor scheme. 2.007 provides small IR reflectance sensors, so I used them as they were in infinite supply. Initial designs featured a timing plate with holes, while the final design moved onto a circular plate with metallic foil tape attached. This was ghetto but worked great.
The final motor in CAD.
In the electronics regime, I decided to take the absolute simplest approach and wire the sensors to a quad comparator. This would directly wiggle some nice logic level fets given to me by Mike (gate drivers get fukt). This was dirt simple but it worked.
Some build photos and video:
I tried winding with thin black wire but it was extremely ineffective. I then switched to real magnet wire.
Lowkey photo with The Dentist
In terms of actually using this on the robot, I gave myself a deadline of three weeks prior to impound to have this fully working. I did have it working by that time, but “working” is a very relative term. It didn’t have quite as much torque as I wanted with the initial winding scheme, and was definitely still very much a “test object” rather than an actual motor system. I decided to keep the keep the competition robot brushed, and entered the switched reluctance motor in the DeFlorez Competition, where it took home an Honorable Mention.
Read on about the competition robot in part 3!!