Robot in 3 Days at Dayton Final Writeup

Hello!

This is the page for the Robot in 3 Days team at Dayton. We work out of team 5811 BONDS’ build shop, and this could not have been done without them. Additionally, we would like to thank the University of Dayton for providing us with housing for what little sleep we got.

This was our first year doing a robot in 3 days, and what an experience it was! While it was very grueling, with less sleep than we really should have gotten, I know that I learned a lot. This is a writeup of all of the things we learned about where we went right and wrong on our prototypes. Once we have a chance to get our CAD a little more organized we will post that here. Please feel free to post any questions you have, and one of our members will be here to answer it.

Finally we have social media if you want to see how our robot does when we get to competition on Feb 2, or for next year and a reveal video

twitch: https://www.twitch.tv/udri3d/
Instagram: https://www.instagram.com/first_dri3d/
youtube: https://youtu.be/qHI_AS_Q1No

Drivetrain

Our host team, 5811, is sponsored by Foxlite Plastic for free polycarbonate. As a result, our build materials consisted primarily of plastic. A unique portion of this is our entirely polycarbonate chassis. This is something that we entirely expected would function poorly on its first iteration. The basic design of this frame was modeled off the Andymark KOP Chassis modified to make bending the polycarbonate easier.

The resulting chassis was a surprisingly sturdy chassis without any supports or belly pan. While not as rigid as the KOP chassis members of UD Ri3D have used in the past, it retained its shape surprisingly well

Construction

We machined this chassis using an Inventables X-Carve 1000 x 1000mm. with dimensions of 30x26, the side plates were slightly longer than the X-Carves 750mm cut bed, which we solved by machining parts diagonally in the cut bed. This method of cutting yields a max robot size in any dimension of 3’

All of our parts were designed in Solidworks, converted to GCode in Carbide Create, Imported into Inventable Easel, and machined out of .175” polycarbonate sheets. These parts were machined in one pass with a depth of the full.175”

After being cut out, pieces were bent over the edge of a table along the bend line by a heat gun. These pieces were assembled with .5” grip rivets into a final assembly. The wheels, gearboxes, and belts were put into the chassis in the exact same way as the KOP Chassis

End Result

This chassis handled the same way as a KOP Chassis with a similar wheel structure (2 6” omni wheels on each end). As Ri3D ended before we were able to test with more variety of wheels, we cannot speak to the chassis use with other wheels. The chassis had definite issues with flex across the width of the robot as the weight of the motors were more than the one cross beam could handle.

Testing

We tried to test the polycarbonate chassis in a variety of ways the chassis would’ve during a competition. This involved driving over a 1.5” pipe, being rammed repeatedly, and dropping the robot from ~2 feet.

|301x534.5054945054945

The robot suffered no visible wear over the course of around 3 dozen runs over a 1.5” square pipe.

Over 1 hour of constant impacts from a KOP chassis weighted down with ~60lb of iron strapped to the top led to some minor cracking across the front of the robot. These cracks only appeared at the end of testing and didn’t make it through the material (outlined in yellow)

The last test we did was dropping from 2 feet which was the maximum height we thought the robot chassis would be during the endgame. We dropped the robot twice, one to see the result and once to film it. After the first drop, we were only able to find one major crack in the front of the robot and one minor crack along the left side of the robot.

The second time we dropped the robot was much more dramatic resulting in both the front and rear plates snapping entirely along the mount to the superstructure.

Improvements

Once the chassis is repaired, more tests will be possible with the chassis being used on Feb 2 for a full competition test. Based on the testing we have done we have been able to complete so far, the chassis seems to be viable – with a few minor but important changes. The superstructure is currently only attached in the middle of the front and back plates - the weakest parts of the robot and so when falling the force is as perfect as is possible for breaking the robot with the weight pressing in the center of the beams and the ground pushing up from the outside. By adding additional support across the center of the robot and mounting the superstructure more on the side rails the chassis should be much more able to resist the falls we subjected the robot too.

Drivetrain Conclusion:
Is a polycarbonate chassis viable in FIRST Robotics Competition? Absolutely, 1714 is famous for their entirely polycarbonate robots. 2576 posted this in 2014. However as far as 5811 is concerned we have been unable to find documentation on building these chassis, and our polycarbonate tends to be ~80% of the thickness of other examples. The goal of this chassis was to start our own development of a chassis we can manufacture and use in-house, and more importantly begin investigating the use of pocketing to further reduce weight below aluminum alternatives.

Intake:

Strategy: An important part of our strategy was the ability to quickly and effectively obtain fuel cells for scoring with our shooter. Our design for this was a series of horizontal rollers coming out of our main superstructure. The design and method of deployment went through many iterations over the course of a 3-day build.

Construction: The intake was routered out of .175” polycarbonate on the Inventable X-Carve. Geometry was designed in 2d drawings in Solidworks before routing.

Testing: Our First Prototype was just a proof of concept to prove that horizontal rollers could effectively pull the fuel cell into the robot.
intake1

Our next many prototypes were variations on geometry to get a fuel cell over the frame of the robot. In order to intake the ball into the robot, we used 2 belts which fell down to constantly provide pressure into the robot

intake2 intake3

This made it very easy for the robot to intake balls in the center of the robot because of the constant force in. however, this design quickly stopped functioning as the balls were on the outside of the intake.

In these situations, the ball would get stuck in the plates holding the intake and refuse to move. While the above video didn’t have bumpers with a bumper gap, the exact same result happened with bumpers. Additionally, the mecanum wheels on the intake didn’t center the ball

Conclusion

Better testing to allow the intake to work outside of the center 8 inches would make this intake work much better. As it currently works the intake is too small for use in actual competition and we would recommend significant changes and prototyping before use.

Indexing : The outer structure of the indexing system worked great for providing structure and stability to the entire bot, however I think we should have thought through how to attach it to the drive train before getting to that point. I also noticed that since the CAD didn’t match the bot in terms of spacing between the structure, other systems such as the intake on rails suffered. In terms of the actual indexing system itself I thought the single belt system worked well and I liked where the top pulley was placed, the bottom one could have been lower, but it still worked well. The rails that the ball sat on could have been a little closer to the belt so we wouldn’t have to use the pool noodles, but I think that was a CAD vs build thing and the pulleys were a little higher than anticipated, and overall this was a minor issue that was quickly solved. I also think I could have been quicker getting the indexing design from CAD to build, but that’s more of a me thing.

Shooter : The shooter was probably the most successful part of the robot. During strategy discussions, the team identified the ability to score balls as a necessary quality in an eliminations team. We also decided that shooting high goal is worth the added complexity. While we did not see the inner port as necessity, we continuously made design decisions that aided our ability to make the center port: high release point and a line drive trajectory.

We went with a 4in Fairlane wheel powered 1:1 off of 2 Neos. The hood shot at 75deg, and did not adjust. We found that 6.25 inches for the ball, plus 2.5in pool noodles cut in half gave the ideal amount of compression. We only had access to a single ball so we did not have the ability to test shooting multiple balls in rapid succession, however, the shooter was very capable of hitting most shots on the field. We could shoot from just behind the safe zone in front of the goal, as well as anywhere behind it to the trench zone with some adjustment of wheel speed. While the geometry may not be ideal for a short bot, we have no major complaints about this hood design.

Climbing Arm : The idea of the climb was a carabiner that would hook around the bar, linked to a rope that would winch the robot up. The goal for this was to simply be able to climb up ourselves, and not any sort of buddy climb.

The hook needs to be more stable to be reliable, the velcro and U-bracket used to hold it in place don’t hold it well enough to prevent it from twisting or falling off when the robot hits a bump or defense. The arm also can extend past the 12in outside the back of the frame currently which could prove to be an issue since without that extension we can only reach one height of the bar. I think with some work this system could be great, but right now it could use some help. In terms of stability, the lexan holds its position really well without significant wobble when swinging up towards the bar, due to the 90 degree bend on it.
For future work, we need to work on how the hook is being held and what we want to do about the reach issue.

The winch that we we used a 775pro with a 49:1 reduction through versaplanetary, then a 2:1 through gears, all to a 1in hex. Although a little slow, it was fully functional. The only complaint we had was that there was not very much space between the winch bar and the motor just under it, I am concerned if the rope spools in the wrong spot the motor could get torn off.

Control Systems : The goal of our controls system was to: make an operating control system, make it easy to troubleshoot, and to make it compact. Our team made “wiring box” with a sheet of lexan with small hexagons cut into it to allow for attachments of the RoboRIO, Power Distribution Panel, and 8 Victor SPX motor controllers via zip tie.

The sheet was bent twice to form 2 walls to attach to and a base. Shooter motor controllers, VRM, and radio were mounted in other locations. It succeeded in being very compact, and it functioned properly. However, it was not very easy to troubleshoot. It was hard to follow PWM wires from Motor Controllers to RIO, and it was hard to manage wires between MC and Rio. What we would modify for future use: Increase distance between Motor Controllers, increase distance between Motor Controllers and back of RIO, and spend more time mounting wires.

Generally: The team was exceeded our own expectations, with very little budget and this being our first time doing a Ri3D. If we had just a little more time, we definitely had the space and weight to add a control panel mechanism, but we never got around to it. We have a lot of things to improve on plannning wise for next year, Ri3D is not as easy when you do not have a solid plan of what to do going into it. Nevertheless, we are happy with our robot, Lex

-Isaac, and the rest of Ri3D at Dayton

5 Likes
2 Likes

How many balls can you fit in your feeder?

We had the room for 5, but we never able to confirm it because we only had access to a single ball