As a part of our off-season preparations, we have started to plan out our build season and I wanted to ask how long you guys aim spend in the CAD stages? I also wanted to ask whether or not people found “Crayola CAD” effective, and how you guys would typically iterate though the CAD process?
12 Likes
Crayola CAD, I call it Cartoon CAD, is incredibly effective and important. To me, CAD throughout the season never stops, we are constantly iterating the design which means we never stop doing CAD.
(Durations) below are estimates and often many of these steps happen in parallel. The drivebase is done first and quickly (<1 week) while other subsystems are in concepting phase.
Overall CAD steps:
- Whiteboarding, robot concepting (1 afternoon)
- This includes researching other robots or mechanisms used for similar tasks in past games
- Robot Architecture Sketches (2 days)
- Most often a Side View, illustrates the frame, bumper, intake, superstructure, and different scoring positions
- Often a separate sketch done by someone else for the endgame mechanism they are exploring in parallel
- Example from 2019 with intake, climber, arm on elevator, etc
- 3D Cartoon (Crayola) CAD (2-5 days)
- Extrude shapes to make a multi-body master model that has the Drivebase, Bumpers, Superstructure, Intake, Endgame mechanisms
- Sometimes separate those bodies into individual parts and remerge into an assembly so you can show the intake stowing, scoring positions, turret swept area etc.
- As the designed, I often find this takes me about an hour to split all this out, and honestly isn’t necessary for myself who can understand my own sketches, but is immensely useful for other team members to visualize the architecture, potential interferences, get excited about owning a certain subsystem, etc.
- Examples below from 2022 and 2023
- Prototype Subsystem Design (2 weeks)
- Always intakes and shooters, often feeders/serializers, sometimes drivebase if terrain on field
- Lasercut plywood, planetary motors or Falcon (more powerful than a handheld drill), adjustability like moveable roller position / compression amounts
- Lots of testing of different materials for optimal grippiness, etc
- Examples in 2020 Technical Presentation on prototyping serializers
- Initial Subsystem Design (1 week)
- Usually start a fresh Master Sketch that outlines the subsystem: motor placement, stroke length / range of motion, compression amounts / roller size, belt/gear powertrains
- Often these are multiple sketches stacked ontop of each other illustrating structure, powertrain, stow pose, top view, etc.
- I gave a meh workshop on Master Modeling that discusses some of these, especially the multi-body methods used more for robot architecture.
- Sometimes import Master Sketch into individual parts or just model parts as standalone pieces
- Start integrating subsystem together into a TLA (top level assembly) to ensure no collisions and lock in mounting provisions and other space claims on the drivebase that is being manufactured at this time
- Example, 2020 drive gearbox with layout sketch
- Detailed Subsystem Design (2 weeks)
- Add all the holes, screws, electronics mounting, bits and bobs
- Proper filenames with Cheesy Parts numbers
- Part drawings for manually machined parts
- Part dxfs / stls for CNC router / 3D printer
- Student subsystem owner is owner of procurement and kitting of parts for their subsystem, they see it through to completion
- Subsystem Iteration (5 weeks)
- Ongoing throughout the season, redesigning subsystems or tweaking as we see flaws, get inspired by other teams, have a new project for worlds, etc
- Much easier to do now with no bag
96 Likes
I’m curious what other designers / teams do or how long they spend on each phase of CAD.
2 Likes
I follow a similar structure to you, but instead of any Cartoon/Crayola CAD I just put more work into the 2D sketches. I see the value cartoon cad can have, I’ve just never found it as necessary (probably because the cad teams I work with tend to be small).
16 Likes
I put a large focus and devote a lot of time to 2d sketches when designing as well. Most typically it’s side views, but front and top down views can also be quite helpful at times.
In a pinch/early on with 3d modelling, drawing 1x2 tubing structures as a single solid extrusion can be immensely helpful too. Lots of time saved by being able to modify a simple sketch/extrusion rather than update individual parts multiple times over throughout the initial design stage.
4 Likes
We have very similar steps but timeframes can drastically change depending on the mechanism. Because of this we find ourselves iterating a final design for one mechanisms while in parallel in early prototyping stages another.
2046 also has a preliminary design review and and critical design review sometime in week 3 and 4 but have been trying to have smaller peer reviews before manufacturing any mechanism/prototype.
How does 254 integrate design reviews into there CAD process to catch errors? Do you aim for a smaller reviews for individual mechanisms, larger reviews for the entire robot, or do you find yourself with a quick enough manufacturing process you can find the errors after assembly and testing with little wasted time.
3 Likes
Generally the most experienced design student works alongside a mentor on overall system architecture. We usually have a pretty good idea about subsystem space claims and interdependencies from that so we rarely run into truly full-stop blockers as we develop subsystems.
Subsystem design reviews are conducted informally in the lab, Zoom call screenshare co-annotates, and a final one is done during the BOM review. The early ones catch things like powertrain design, packaging/integration-points, and potential risks.
Once more mature we do a BOM review on a zoom call. We export the BOM to a spreadsheet that we import into Airtable for tracking order and manufacturing status. On the call, the designer screenshare annd and thens starts hiding component by component from the subsystem until everything is hidden (it’s a way for checking things off the list). It’s here that we check all the implementation decisions like fastener selection, “is it strong enough” gut-checks, and catch big errors.
Finally, before parts go to manufacturing they need a part drawing made and that is when a mentor will sit with the student(s) to review their drawing and also ask questions about things like tolerances.
34 Likes
This is fairly close to my preferred method as well. In my experience the keys to a successful design are winning robot concepts (we usually spend two or three days brainstorming these because we want to get it right the first time, and also smaller design team) and well-refined 2D/master sketches (again, smaller team, making less mistakes at the beginning of the design process regarding interferences and such makes life a lot easier when we’re in week 4/5 doing design reviews and making tweaks).
2 Likes
@Torrance Can you elaborate on 254’s process for tracking manufacturing/assembly status? We’re looking into a few different processes and Airtable came up as one potential solution. Thanks!
2 Likes
We use Airtable for Program Management, tracking parts in subsystems and making sure they get ordered, manufactured, and assembled onto robot.
Each item, COTS or custom, is entered as a row. Often the initial upload of these comes from a Solidworks BOM export → CSV → Airtable import, but we will also add parts individually throughout the season. We have a variety of columns which combined with Airtable’s great View / Filter / Sort / Group commands allow us to organize the data in many different ways.
We still use Cheesy Parts, but really only for the creation of Part Numbers per subsystem and making sure PNs dont overlap. We do not use the status or other fields in Cheesy Parts.
As parts get finished, we put them in labeled cardboard bins. Cardboard bins are nice because if they get gross (oil/grease/dust), we can just recycle them and have new bins for the next year. The color coded dots allow us to label very large parts, like superstructure tubes, leaving them on top of the rack or in other areas of the lab.
35 Likes
This post added an immense amount of clarity to an issue we have had in the past. Organizing what happens after the design is done to ensure nothing is forgotten and nothing is misplaced is a gap we have been trying to close.
Between the BOM export to that table and your bin sorting strategy I can see how your team implements a robot beyond the cad design and for most teams that is actually the hard part. This is something we could implement easily and I thank you for sharing this!
1 Like
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.