Development of 1678's Can Grabber

Having been intimately involved with all phases of development and testing of 1678’s can grabber, I thought the FIRST community might be interested in seeing how things progressed throughout the season.

After kick-off, our team agreed to implement a “chokehold” strategy of acquiring all four cans from the step in autonomous. Our belief was that, if we accomplished this, it would be impossible for our opponents to match our score. This topic has already been well debated in other threads so I won’t go into the details.

Our original design involved a foldable wing that could deploy in under a second. We prototyped it using two windsurfing masts (I’m an avid windsurfer) and the following videos show how that went:

After a lot of discussion and consultation with engineers who knew a lot more than us about light wingspans (think airplanes) we decided to design the wing using thin sheet folded into a rectangular section and articulated with 4 hinges. The design was completely fabricated and assembled a few weeks before the Central Valley Regional. Here are a couple of videos of testing:

We were getting pretty excited at this point. We had a design that was within transport configuration, could be set up in less than a minute, reached the cans in about a second and released cleanly. We were working on the pulley system to reel the entire contraption back in when we saw the reveal videos showing 118’s and 3310’s can grabbers. We thought, “what is the point in going for all four cans when someone is going to beat us to at least 2 and possibly all 4?” We immediately abandoned this design and began working on our own 2-can grabber with just a couple of weeks to go before the Central Valley Regional. We decided on a motor-driven system using extendable hooks similar to 118’s. Here are a couple of videos of us testing the hook and pull back:

Here’s video of one of the first tests on the full system:

One of the things we worked on quite a bit was preventing bounce-back and also the addition of a shield to prevent opponents getting their hooks into the can. Here’s a video of the shield testing:

We ran this system at the Central Valley Regional, Sacramento Regional and Silicon Valley Regional. At CVR, we built a cheesecake version from scratch and mounted it on 1323’s robot. At Sacramento, we mounted the same cheesecake on 5458’s robot. At SVR we mounted it on 5027’s robot and then gave it to them to take to CMP. 5027 was very successful in the can war on Newton, including a great battle with 3310.

Even though our 2-can grabber was fast (around 250 ms) we knew that teams would be faster at champs so, even before SVR, we started working on a lighter, faster version. We decided we needed to power the system with surgical tubing to make it faster so we came up with a four-bar linkage that allowed us to develop a lot of power with a system that could be placed behind the arms. The design team came up with a super light triangular frame design for the arms made out of 1/2 - inch carbon tubing but it just seemed to weak to me. I got my windsuring masts out again and we prototyped the four bar linkage hammering the mast onto the can. It was super fast and almost destroyed the top of the can. We were convinced we should just use a single carbon tube. This was about a week and a half before champs. We hastily ordered some 1-inch carbon tubes and began designing a new hook using 1-inch aluminum tubing for the prototyping. The biggest problem we were having was bounce back. The hook would bounce more than a foot above the lid before settling back in and we knew we could lose the can in that instant. The solution came by accident when we were testing cushioning materials to soften the impact and reduce bounce back. During one test we saw the cushioning material push the can forward, locking the hook in which resulted in minimal bounce back. Here’s a video of that revelation:

We quickly modified our hooks to incorporate a wedge behind the hook to push the can forward. This turned out to be the crucial piece of the puzzle that made our can grabbers so effective. We felt that it accomplished three different objectives: (1) it locked in the hook so there was no bounce-back, (2) it moved the target for our opponents grabber making us think we didn’t need the shield anymore, and (3) it absorbed the direct impact of the arm reducing the risk of breaking the carbon tube under the extreme force we were exerting on the arms. We tested our final arms with 12 wraps of surgical tubing per arm and achieved times or around 130-150 ms. in the shop but decided to go with just 8 wraps at champs to reduce the strain on the system. We feel good about that decision:

Throughout this entire process, our students worked relentlessly and never stopped pushing to make our can grabber faster and more reliable. I am so proud of what they accomplished.

Your team’s continuous improvement spirit is inspiring. Enjoy the Einstein victory!

Great job, I enjoyed your post. Would you supply pictures of the 4 bar drive and surgical tubing? We didn’t get to see it in St Louis

Your story is very inspiring. You worked the hardest on the most important part of the game, and you now have a championship victory for your efforts. Congratulations.

When I was watching Einstein, I was in the nose bleed seats. I found this disappointing, because I knew I wouldn’t be able to see that can battles very well. My dad then handed me a cheap 7x monocular. Although this was nice, I still wanted a better view…

So I captured this video with a Galaxy S5 and the monocular at 240fps:

Not quite field side quality, but it was still entertaining.

(If you’re interested, I have some of the other Einstein matches, but not all. It was very hard to hold the phone and monocular aimed correctly)

the amount of thought that went into your grabber in order to turn the advantage even a tiny bit more in your favor is astonishing. A full competition year of improvement is hard to do with any system, let alone something that is technically a secondary skill

How much did the final mechanism end up weighting? Very awesome stuff guys.

The engineering process is showing so greatly, and it’s just amazing seeing how all of that development turned out for the team. I can’t wait for what the team does next.

Congrats on Einstein!

Congrats on winning!
We were on Hopper so I got to see you canburglar a couple times before Saturday. They were truly amazing, I thought you were going to break the lids on the cans it was so powerful.

It’s fascinating for me to see this layout of the whole process, because I was involved with prototyping from day one. There was a team of a few students who only worked on the “chokehold” mechanism for all or most of build season. It was amazing being part of that team, and analyzing the geometry of the cans and landfill and robot and different materials we could use and different ways to bring the cans in afterwards.

My favorite part was designing the hooks themselves, in the early stages. One thing Mr. Harvey didn’t show as much is how the hook design changed. When we were working on a 4-recon grabber, we had one design of the hooks that looked like this:
They would go into the center of the hole, and the two “fingers” would bend in and snap out because of torsional springs around a bolt at the pivot point. Those strings held the two fingers in tension against the spring, and in the final design with the bent sheet metal, there were pneumatic pistons that these strings were wrapped around. To release the cans, the pneumatic pistons would release, allowing the strings to fall and release the fingers, which would release the cans. We realized that we wanted a passive release system in case of a tug-of-war, so we experimented with looping a zip tie over the pistons, and tying the strings to the other end of the loop. If there was too much force, the zip ties would break without the pistons needing to release, and the strings would release. This was one of our final designs of the hooks for the 4-recon grabber. I think we did some CADing to make the design more sleek and lighter, but I don’t think the mechanism changed beyond that.

I was not as involved with the 2-recon grabber, because I focused more on the chairman’s presentation at that point in the season, but the hook design changed to something without any moving parts or active systems, which is what we used for CVR, Sac, and SVR, and Mr. Harvey talked about the Champs hook design!

There was indeed an incredible amount of work done on this by a majority of the mechanical subteam at one point or another, as well as Mr. Harvey and Mike! I learned so much about the engineering process, and I’m so glad I got to work on it because it made me really excited to see other teams’ recon grabbers!

Very impressive development! I have two questions -

1- earlier in the season you had a release video with all 4 cans being taken - and people were guessing you were using some sort of tape measure system.

Was that in fact, the original 4 can wind-surfing system?

2- In this video

It looks like you have a motorized pivot for the pole - if that’s correct can you say a bit more about that? And you eventually switched to surgical tubing? And, it looks like the poles are extending as they pivot and retract, can you describe that mechanism a bit more?

Thanks, and congratulations!

In the video Can Grab 4 the system mounted to the robot is the teaser video system that we used.

As for 2 it was just directly attached to an output shaft that cranked hard of the arm to accelerate it quickly. The switch to surgical tubing was only mad for our Champs grabbers. The extension was done by anchoring a string to the body of our robot and to the INSIDE of the arm at the bottom, as the arm comes down it pulls on the string and causes it to extend. The top of the square tubing has bearings in place as does the bottom inside of the round tubing so it slides properly and does not rotate.