955 Preseason CAD Release

Like this past year, team 955 is releasing preseason CAD. In the fall, @Radicalthief developed a west coast drive, and I developed an elevator.

Elevator

Onshape link

Design Goals and Constraints:

  • Customizable size: Without knowing what the game is, I wanted to be able to easily adjust the width and height of the elevator. This meant maintaining solid design practices that facilitate adjustment. The as designed size of the elevator is roughly the size of a 2018 or 2019 elevator, however, I didn’t spend much time determining overall dimensions as it’s meant to be changed.
  • Easily manufactured: Our team has access to vertical mills, a cnc router, lathes, and basic 3d printers. All the aluminum parts on the elevator only require holes drilled, there are no parts requiring milling.
  • Capable of lifting game elements and climbing: The elevator is designed such that it is fast enough to lift game elements and can provide enough force to climb on the first stage. Without knowing the game, this gives us the most flexibility in our designs.

Overall Design:

  • Box tube construction: By far the most common and easy to work with (compared to round tube).

  • 2 stages: This seems to be a solid balance between total travel and complexity. The design should readily scale to a third stage if more reach is needed. However, I would prefer to add a virtual 4 bar or similar on the carriage (like code orange 2018) rather than extending it to a third stage.

  • WCP clamping bearing blocks: These bearing blocks are easy to assemble and maintain without having to take the whole stage apart. They require 0.5” between stages, which isn’t as compact as other options. However, the larger spacing between stages allows for things such as climber hooks to be mounted between stages.

  • WCP tube plugs: These tube plugs take the place of gussets, and are much simpler to design around, as there is no potential for interference with the bearings hitting gussets.

  • Weight: Estimated weight of 20.6 lbs, probably closer to 25 lbs in real life. Weight could likely be shaved by switching to thin wall max tube, but we have a lot of 1/16” 2x1.

Rigging:

  • Cascade: The first stage in a cascade elevator can provide more force than the carriage, which helps accomplish the climbing goal without requiring shifting gearboxes. Cascade also allows for simpler rigging and is less finicky.

  • Chain driven 1st stage: Because of cascade rigging, the first stage can be chain driven while the second stage is cable driven. This eliminates the need for more complicated machining and design for a cable spool. A single #25h chain is more than enough to carry the weight of the whole robot. A series of 4-40 bolts through angle brackets is used to attach the chain to the 1st stage. We had success using this method for our mid climb last year.

  • Cable driven 2nd stage: The second stage is rigged with a pair of UHMWPE cables, very similar to the thriftybot or greyt elevators. Turnbuckles are used to tension both the chain and the cables.

  • Counterbalancing: Mounts for 2 10.6 lb constant force springs are available. If we use this elevator during season, I doubt we will use the springs because they don’t contribute significantly to performance and come with extra safety precautions.

Gearbox:

  • Dual neo gearbox: With the goal of climbing and lifting on the same elevator, 2 neos were required to provide enough power. The final ratio is 8.2:1, which allows the carriage to travel at 7.3 ft/s while lifting game elements. While climbing, the first stage moves at 2.8 ft/s while drawing 28 amps per motor.

  • Gearbox Design: 8.2:1 would be unrealistic to accomplish in a single stage, so the gearbox has 2 stages. The first stage is accomplished through a pair of versaplanetary gearboxes, geared 5:1. If I was building this from scratch, I would choose maxplanetaries, but 955 has lots of VPs on hand. The second stage is a gear reduction (36:22). The combination of planetary and spur gear reductions is compact and easy to build.

Wiring and Sensors:

  • Sensors: A pair of rev throughbore encoders, geared together 30:31 are used for determining elevator position. This setup ensures that the elevator position can be calculated without zeroing the encoders at the beginning of each match (credit goes to @Nick_Coussens for giving me the idea). The downside to this setup is that the code to calculate position is a bit more complicated, and removing the encoders requires resetting the zero location.

  • Wiring: Energy chain is routed on the front left side of the elevator and attaches to the carriage. However, if I were to run this during season, I think running the energy chain next to the chain run on the back of the elevator would be a better location.

West Coast Drive:

Onshape Link

Design Goals and Constraints:

  • Customizability: As with the elevator, the drivebase size and design should be easily adjusted to fit the needs of the game.

  • Ground Clearance: 955 is comfortable running swerve drive this upcoming season if the game allows for it. To do this, we required a bit higher ground clearance. We still chose to do a rather standard WCD (rather than 8 wheels for example) because at the time the decision was made to design WCD, we weren’t completely comfortable with swerve yet. At this point before the season, we likely won’t run this drivebase in its stock configuration because we will either use swerve or need some modifications for a terrain game.

  • Bumper Mounting: Last year’s kit drive chassis relied on wingnuts to attach the bumpers. While reliable, changing bumpers was slow and wingnuts were easily lost. One of the main goals in developing our wcd was to improve our bumper mounting methods.

Drivebase:

  • 6” REV Wheels: 6” wheels seem to provide a reasonable amount of ground clearance without being overkill (we are absolutely not repeating 955 2020’s drivebase). REV wheels are reasonably cheap and should hold up well throughout the season.

  • 4+2 Wheel Setup: Our stock kit bot drivebase with 6 traction wheels had numerous issues with turning last year. For this drivebase we are choosing to have omnis on 2 corners to hopefully improve our turning. We chose 2 omnis over 4 because of the increased traction. There is still a 1/16” center drop to ensure the center wheels are always in contact with the ground.

  • Bearing Blocks: Our 2020 drivebase had numerous issues with chain tension with bearings mounted directly in the tube. We hope to avoid these issues this year by using clamping bearing blocks and the WCP cams for tensioning. However, this does come with the tradeoff of decreased space on the top of the drive rails for mounting mechanisms.

  • Flipped WCP SS Gearbox: A fairly standard gearbox for west coast drives. We chose this over the new REV drivebase gearboxes because of increased ground clearance and ease of mounting. While we could manufacture our own custom gearboxes, it is absolutely not worth the time and design effort for our team.

  • Rear Mounted Gearbox: Rather than mounting the gearbox in the center, which is standard, our design has the gearbox mounted at the back/front of the drivebase. This allows for more room to mount mechanisms in the center and is easier to maintain if mechanisms are located over the center of the robot.

  • #25H Chain Drive: Our past west coast drives used #35 chain, which is rather heavy and overkill. Instead, this design uses #25H chain, which provides a solid middle ground between 25 and #35 chain.

  • Speed: 15.2 ft/s seems like a reasonable compromise between acceleration and top speed without knowing the game. This could be fairly easily changed by simply swapping gears in the gearbox.

Bumpers:

  • 2 Piece Bumpers: Split bumpers are easier to put on and take off the robot because they can be slid from the sides rather than coming down from the top, especially with tall robots.

  • Toggle Clamp Mounts: Brackets on the bumpers slide down over studs mounted to the drivebase. The studs prevent the bumpers from sliding off while the bumpers are retained vertically by a pair of toggle clamps (per half) which hold down the brackets. Inspired by this post.

As always, feedback is appreciated!

5 Likes

Those bumper mounts are awesome. We were looking at doing something almost exactly like that but we ran out of time this preseason with all our other projects. See you guys at Clackams once competition season starts.

3 Likes

How does the dual encoder setup works?
Do you think it is necessary since the elevator will probably start collapsed anyway due to height restrictions?

1 Like

It’s definitely not strictly necessary, however, it does eliminate one potential problem in setup. If it does turn out to be unnecessary or difficult, I’ve also designed in limit switch mounts (currently suppressed in the part studio).

With the encoder setup, each encoder reads an absolute position. Because the encoders are geared together 31:30, the encoders will read unique values up to 31 revolutions of the elevator shaft.

The software works by generating lists of possible positions of the elevator for each encoder. For example, if the 30T encoder reads 90 degrees, then it must have rotated n + 90 degrees, where n is any integer from 0 to 30 rotations. With a bit of conversion, this can be turned into a list of possible elevator positions (using the elevator travel per rev). If you repeat this for the second encoder, you get 2 lists of potential positions. The position that matches in both lists is the elevator position.

4 Likes

Sounds interesting, Thanks!

If you’d like to read more, this is essentially what the Chinese remainder theorem describes.

2 Likes

Seems like this would be a good use for REV-21-2099

Takes two NEOS and combines their input into one, and its output can be directly connected to the input of a MAXPlanetary. Should come out a good bit lighter than two VP setups.

1 Like

I did consider using that gearbox, but we just have a lot of versaplanetaries and gears laying around. It’s a lot more economical for us to just build a simple custom gearbox. It does weigh more, but it’s low at the bottom of the robot, which I’m ok with.

1 Like

An update on drivebase and bumpers:

After some thought and research, I’m not so sure I trust the bumper mounting method we designed. Because only 2 clamps retain each bumper half vertically, it’s possible that upward force on the bumper could make it rotate and break. Furthermore, the bumper mounting plates take up a lot of space on the top of the drivebase rail.

Another solution we’re looking at is the 2910 method, which relies on bolts driven into standoffs. The method is proven to be durable and effective, and should be quick to change. The primary downside is that it requires a drill and driver to change bumpers.

To test it out, I designed a quick swerve chassis with this method.
Onshape link

After kickoff, and some discussion with the rest of the team, we will make the final decision on what method to use.

2 Likes

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.