To the teams who opted to build a “sideways” elevator this season (an elevator in which the mechanism mounted parallel to the carriage rather than perpendicular), what were your takeaways from that design process? What aspects of the design were trickier than anticipated? What challenges did you have to design around and how did you solve them? What were some unanticipated issues? How did these vary from a “typical” elevator design? Would you chose to make a sideways elevator again?
Any examples of teams that I could go have a look at?
FTC teams usually do that
3647, The Millenium Falcons had a really cool looking sideways elevator this year (that was also very effective).
4414 and 2102 are the ones that come to mind besides 3647.
5460 and 3309 this year also come to mind. The first one I knew of was 3538 from 2018.
Also 1577 has a very effective sideways elevator
1577 from Israel is also a notable sideways elevator team, and don’t forget 971, 2767 as well
Edit: Haha @AriMB beat me to it
While playing against 5460 we noticed that pushing on one side of their robot vs the other had different effects defense wise. The bumper side pin was much more effective on their elevator side. I suspect this is because that’s where a large chunk of their weight was so any bumper lift interaction gave the pinning robot a lot of traction and gave 5460 a lot less traction.
The reason 3538 built an ambidextrous robot last year was because there were several areas on the field with very short cycles (contested cubes to scale, pyramid to vault, feeder station to opponents switch) that we felt cutting out a turn around would be very beneficial. Especially in auto attempting to do a 3 cube to scale auto (which we successfully did on Curie last year)
Due to team limitations in sensor and software integration we wanted something that could be manually controlled effectively which ruled out a pink arm and pass through mechanisms like 179 or 1114. We had position holding on the elevator and eventually on the arm but never used presets or auto sequences in teleoperated mode.
The decision came down to the sideways elevator and what many teams had in a normal elevator with an end effector that could reach over the back to score on the scale (1747, 1678, 195). We decided if the sideways elevator went south we could just turn it 90 degrees and do the normal thing.
Mechanical design obstacles were the fact that it was so far off to one side it was more difficult to support. It was always more wobbly than I liked.
We drove the arm with a gear reduction which was risky from the beginning. Thankfully it managed the season without too much trouble (a few lost teeth) but it really got out of hand during the off season.
Because the arm could be basically anywhere above the drive base, packaging to keep those areas clear was quite difficult. We consider this a major obstacle to us adding a buddy climb which was so crucial to ranking high last year. (and ultimately why we decided to forgoe a similar mechanism this year in favor of focusing on the end game climb, we planned a partner climb but dropped it due to schedule issues)
Another thing to consider is that driving that thing was HARD. I am still not sure how our operator managed it and her explanations only leave me more perplexed. It took quite a bit of deliberate practice to get our drivers used to utilizing the ambidexterity and not favor one side or be out of sync where the bot would turn, the arm would flip, the bot would turn back, and the arm would flip again and so on.
I would love to use this design again but I do not regret our decision to not use it this year. I love seeing these designs and hope to continue to see this concept iterated and spread.
We did a sideways elevator this year, primarily based off 118 in 2011. We figured it was going to be easier to package a slide when turning the elevator sideways vs. making a large cutout in the elevator similar to 4003 last year. We discussed at one point doing a cargo on one side and hatch on the other but figured we wouldnt have the weight to have separate mechanisms. I believe we may be the only sideways round tube elevator this year, which itself caused some unique challenges, but also seemed to fix some of the round tube elevator issues.
The issue with our sideways elevator is that its positioned in a way that makes it impossible to really change our claw easily to something similar to 1684, unless we do it vertically which we talked about but would be introducing other variables.
3663 in 2018 did sideways elevator. The main thing we learned was don’t use 20x20 mm extrusion and make sure it is attached to the frame solidly. Even when it was retracted all the way, our elevator would flex ± 6" when we accelerated. This year we went to a welded elevator made from 1/16" 2x1 tube. Much more rigid and a lot lighter.
What helped me personally was developing a “playbook” of sorts for the types of situations the ambidexterity was useful:
- Contested to scale cycles
- Portal to scale cycles
By having this explicit list of moves I could use, I was able to more naturally find places and ways to integrate them into my normal driving style. The more standardized move-set also probably helped my co-driver get a more intuitive feeling for when I was expecting an arm flip.
Communication was also critical for us in 2018 for cube placement. Because we were able to reach variable heights with the manual control and flip-over, there was a lot of communication between her and I for where and how to place cubes. This resulted in a playbook of sorts being made for cube placement as well, which simplified the communication and intuition development for that too.
I believe 2481 Roboteers had one
I don’t think so. This year they had a “standard” elevator with their manipulators able to move in the X-axis on the carriage. Last year they did a pink arm.
Ok I must have interpreted the question wrong
What is a pink arm?
A telescoping arm on a 180-degree pivot. Take a look at 27 and 610 this year, or 2481 last year.
And so on and so forth.
Making a sideways elevator wasn’t really different from standard elevator design, other than the fact that everything is mounted/rotated 90 deg from normal. Mounting was definitely one of the harder things involved with sideways elevators. Conventional elevators transfer the forward and backwards forces of the drive train direction changes and/or slamming into objects into some type of elevator support bars. With a sideways elevators, the forces (in an extreme case), want to cause the frame of the elevator stages to twist and flex. The “rectangle” shape frame of the stage wants to become a parallelogram. We avoided this problem by having plenty of stage overlap (this is why we have a 3 stage elevator) and with our gusset design. 5460 really enjoys their sideways elevator. It has allowed us to have extremely fast cycles since we don’t have to rotate the robot for scoring. We really enjoy our robot this year, but having a sideways elevator makes other parts of the robot more difficult to integrate. It took a really long time for our team to develop a lvl 3 climb because of the weird center of gravity and few mounting areas available. In the end, we would definitely consider building another sideways elevator again.
Here are some good pictures