FANTASTIC design! There is something special about how elegant this design is. I’m not well-versed with linkages, why couldn’t you just power the pivot for the arm directly? I’d imagine that’d be simpler.
Doing this would require us to put a large amount of torque through a sprocket on the pivot point at the back of the robot. This would require a large sprocket that would push us to move the pivot point forward, not allowing the arm to reach as high. On the 2018 Everybot we did this and teams had issues with the chain popping off the sprocket. We also get mechanical advantage out of the linkage and maybe most importantly the linkage over-centers when the arm is up so no motor power is needed to hold the arm up.
Great job as usual! I look forward to seeing many of these on the field this year.
I just made a category on the REV website for the Everybot, so teams who want to build this robot can find all the REV parts in 1 place easily. https://www.revrobotics.com/frc/everybot-2022/
*note: I also added a few items that are 1:1 with other items on the BOM
Here’s the YouTube archive of the unveil show from tonight, where you can hear a bunch of awesome insights about the design from the Everybot crew.
(47:46 is when it starts, in case the timestamp link didn’t work)
Okay that makes a lot of sense now. Well done!
Thanks for doing this! We appreciate anything the FRC vendors do to make it easier for teams to buy parts and build their robot.
We are also super thankful to REV Robotics for including the Spark Max motor controllers in the KOP and rookie KOP and for including a Neo in the Rookie KOP. This plus the REV voucher made it to where we could include a brushless motor on this years Everybot and gives teams a path to upgrade their drivetrain to brushless by purchasing just 4 Neos.
I kind of figured once they said that, but it is even more refined than I thought. I especially love having the direction you drive pick the bar.
My more complex climb just used a thrifty (or other telescope) climber and a strike plate connected to the release pin that when hitting the low bar released it so the hook was at the right height to passive climb. This is somewhat simpler and yet very effective looking.
edit: One small concern I did have was the climber being deployed early. I guess in test driving you didn’t have any issues with that?
I think this a great resource for the FRC teams out there. Thanks to 118 for taking the time to put this together!
The hangar mechanism is fantastic!
My favorite design solutions are the ones that are simple, effective, and elegant.
I wasn’t able to shake them loose while practicing. The springs and geometry keep some preload pulling the hooks down until you push them pretty hard under the low rung. If they were to flip up accidentally, they could be flipped back down by just driving under the low rung.
so if you are wondering how to get those hooks made if you don’t have tools, I would recommend looking at SendCutSend.com * You can pay to get them directly cut out of Aluminum, wood, plastic or other materials and thicknesses.
One thing that might be worth looking at if you are on a budget is to get a template cut out of thin aluminum, and use that for cutting other ones with a jigsaw or router
*note: sendcutsend sponsors my Battlebot Switchback, but I am a customer and big fan even if they didn’t do that. They make awesome parts really fast, and I hope they get more involved in FIRST in the future.
Lots of people are talking about how clever the fully passive climber is and we want to give credit to someone who helped make the passive climber a reality. Ender Kerr, alum of team 1339, AngelBotics. Ender took some time away from his day job as a thermal engineer at SpaceX to help out with Everybot this year, and the passive level 2 climber mechanism was his creation. He wrote the below to capture the design process that led to the climb we built.
- Initial concepts involved a nearly identical climbing mechanism to the 2020 Everybot, with a motor driven hook climb.
- Taking inspiration from the extremely simple “ramp and divot” drive on climbs from the 2013 low bar, we made a prototype to answer if a passive structure could successfully drive onto the medium rung. This test showed that it was possible to climb with passive structure, but that it was important to have high enough ground clearance to prevent jamming the rear structure/bumpers against the ground (this jamming is not physically possible with the Everybot config).
- While we could have made a motorized deployable structure, we really wanted to make it entirely passive, with no additional motors to add, wire and program, in addition to the painful cost penalty associated with adding a controller and a motor to the BOM.
- The challenge lay mostly in getting the passive structure into place, because the mid rung lives above the normal field height restriction. However, the low rung is short enough that a robot can interface with it while still staying below the field height limit.
- This means that you could use the low rung itself as a trigger to deploy the structure by driving under it. This sacrifices the ability to drive under without deploying, but is super beneficial in the reduction in cost and complexity.
- We were initially trading a few deployment concepts:
- A “unlatching” deploy, where driving under the bar rotated the arm down and unlatched a linkage that then allowed the hook to deploy fully
- A “pivot deploy” where the arm had a vertical bar that struck the low rung and then was rotated up and into position.
- Went with #2 for several reasons:
- It allows for full speed driving across the field, slamming into the low rung and then the mid rung for a sub 5 second climb consistently.
- It was possible to set up the geometry such that the hook is high enough to engage before the tip up bar disengages from the low rung, giving good confidence that the hooks will always be high enough to engage the mid rung, so long as they don’t bounce off the lower hard stop. This means you can drive over as fast as you can and it still hits.
- The tip up bar could have a nearly identical hook mounted to the rear, to enable strategic flexibility in climbing either the mid or low run in <3s
- Springs could be mounted such that they have near constant stretch, and act in an “over center” configuration, where they held the hooks in the down position until they were rotated up far enough, where they then helped deploy and hold the tailhook arms upright.
- Made a wood prototype climber to test deployment geometry and kinematics and dial in allowable tolerances. Showed that clearance to the structure/rung was particularly important when coming in from an angle, and that connecting the hooks in any way would jam the robot entirely at an angle. Also lead to some adjustments in the height of the climb to lower the height and make engagement more consistent. Showed that we needed more spring force upwards and less to keep hooks down on the field.
- The final version refined the design in the above ways and made it lighter and thinner where we thought was appropriate. Borrowed the bronze bushing hinge we were using on the arm assembly for smooth movement, shifted to a outer plate and backing plate assembly to give adequate clearance to the raised intake, added frame and correct spring mounting location, and pulled the hang cones of acceptable CG as close as we could to where our best guess at final robot CG was (Still needed ~16lbs of counterweight to the rear of the bot to get the low hang to close without resting on a wheel). The CG issue was particularly driven by wanting to keep the main bar vertical to make both it and the tailhook arms easier to manufacture, and being limited on how close to center we could put it with our preferred mounting method to the main frame.
- Testing on the real chassis showed that further softening of the hooks was required to prevent tipping backwards during the climb. This was achieved by keeping the “climb distance” the same, but by softening the forward edge of the final divot to require less overall height increase during the climb. This sacrifices some retention ability of the hooks, but seems to have been a good trade.
- Despite being a pretty heavily constrained problem, there is certainly still room for adjustment and changes, and we look forward to seeing what improvements teams come up with.
SendCutSend sponsors at least half of Battlebots, $40 credit in the Virtual KOP (team pays shipping) would be under $200k to get stickers on thousands of robots and extremely functional for many teams… Or put vouchers in First Choice if they were nervous about the total price tag?
Of course, knocking out their real production capacity with $0 revenue FIRST orders for the month of Jan and Feb might be a bad business move and is probably the real issue with my idea.
I’d be so down though, the stickers are awesome.
Also they make good parts fast.
We are working on creating 1:1 templates for cutting the hooks out with a jigsaw/scroll saw/coping saw. So if teams need to make them by hand they will be able to.
In my experience Send Cut Send will likely be the cheapest vendor to get the hooks cut out from. Another option if you have issues getting them as an approved vendor through your school or organization is to use the waterjetting service Andymark runs with 221 Robotics.
Another option is asking a local team who has a router to cut a set out for you. If you are a team who has a router and the means to do so cutting out some sets of hooks and giving them to your local teams would be a really awesome thing to do.
Thanks for the write up. I appreciate all the small details that went into making this incredible part.
Also, somewhat related the cargo grabber reminds me a lot of watching 1011 play Stronghold, putting lots of boulders in the low goal.
2022 Everybot now available in Onshape here or from the MKCad App!
Holy snap that’s smart
We used SendCutSend.com this spring when we couldn’t get into our shop as a group and we still wanted to compete with our robot. They were super easy to work with and delivered a great product. Definitely would recommend!
Looks awesome 118, well done. Since it looks like you have a working field, do you have an estimation of how many CARGO it can score during teleop?
Huge fan of the Everybot this year Ryan, congrats to you and the team for making it happen! Brilliantly clever!
If anyone needs a set of hooks and is close enough to Davis that they can pick them up, I’d be more than happy to cut some for y’all, shoot me a DM!
Great resource Everybot team.
As you know, we built an Everybot in 2022. We’ll be having freshman build an Everybot and sophomores build a separate Everybot this year to get as many students some hands on experience to make up for the lack of it during the pandemic. I’ve had a couple of underclassmen approach me concerned of “how will we learn how to design our own robot if we only ever build from instructions” (which: side note is very easy to answer because our underclassmen don’t even comprehend how much there is to learn before stepping into the “design from scratch” part) and I can’t wait to point to the linkage in the arm and passive climbers as great examples of ways they’re going to be learning the design part by “just following instructions”.
Can’t wait to see how long it is that our underclassmen Everybots outperform what our upperclassmen are working on. I suspect it’ll be quite awhile!