How do teams integrate CAD into their design process

So the title says most of it. My main interest is what does your design schedule look like. Do you start building prior to the CAD being done? Do you wait for it to be completed, and just prototype, making changes to the CAD as you learn from your experimentation?

I’m learning CAD right now in order to help our team take a step forward to becoming less jank, and more planned out, but that won’t happen if we don’t utilize CAD to it’s the fullest extent, and that is why I ask.

1 Like

-Week 1
Strategy decision and design guidelines for the robot (like height, capacity etc) and prototypes the CAD team CAD the prototypes to give the manufacturing team precise drawings to have better quality prototypes

  • Week 2 and 3
    After design guidelines are established the CAD team works full throttle while the manuf team makes more refine prototypes trying to forsee problems, week 3 the bot is fully CADed we send the drawings to sponsors and start manufature in house, CAD is essencial to save time and to encouter problems early on, and you need to CAD everything even bolts and cables, to find interferences, and to plan and start the manufacture of the cabling of the right size even before the robot is built

Disclaimer: my team has a decent amount of in-house machining capability

Our design schedule (philosophy? timeline?) is mostly driven by the fact that we only meet 3 days a week. It’s really convenient to mock up some prototype, go home and CAD a better version based on what you learned, come back and make that, repeat.

As we start to get more finalized designs, or at least better ideas of overall robot design, we start integrating. I think this year it was around the end of week 2 that it started to be a focus? The design team puts everything into the main robot CAD and makes sure there’s no conflicts, there’s space for electronics, etc.

After we “solve” integration issues, it’s mostly up to the component designers what happens next. In most cases, we kept iterating for another week or two, still on the cycle of build-test-improve. Things get put on the robot as they get the kinks worked out, and the electrical team (programming in disguise) starts wiring stuff as it gets put on.

TLDR:

  • Do you start building prior to the CAD being done?
    • Yes. Prototypes start immediately. Sometimes we make a “final” version of a part (shooter this year), sometimes the part that goes to competition is just another iteration of a prototype dressed up to look prettier (intake/conveyor this year).

Thank you! On your statement about needing to CAD wires, how would I go about that (I think we’ll using Onshape) It seems like way too much work to create a bunch of sweeps in just the right way, so do you know of any solutions that don’t involve that?

1 Like

Disclaimer: my team has a lot of manpower and decent in-house machining capability, so what works for us might not work for everyone else.

What my team does is after kickoff, where we make all our big strategy decisions (ex. tall vs short, priority list, etc.) we have “mechanism day” the next day. Here, we come up with every mechanism we can think of that could potentially be viable for the game, and decide which ones we want to prototype. We try and prototype everything that we don’t already have experience with and sometimes things we know how to build but need to fine-tune. For example, this year we came up with a bunch of manipulator prototypes, a shooter prototype because we needed to tune it, but only a handful of climber prototypes because for us, climbing on a bar is more or less a solved problem.

From there, we separate into groups to prototype each mechanism we decide is worth prototyping. We prototype until week 3, where we come back together as a team to figure out what ultimately goes on the robot. Once everyone’s reached a consensus, we start to CAD. CADing the robot and working out integration issues usually takes about a week, and we only start to order/machine parts once the entire robot CAD has been more or less finalized.

From there, which is usually around week 4, we start building. We build and wire the drivetrain first while mechanism leads work out the kinks in their mechanisms and get their stuff on the robot, which usually happens around week 5 or 6. Once we have a functional robot, we hand it off to programming and our drivers, who do their thing until our first competition. After competition, we iterate based on what we saw both during testing and at competition, and try to make all of our changes in time for our next event.

TLDR: Prototyping => CAD => build competition robot => competition => iteration

1 Like

I’ve never been in a situation where I felt I needed to CAD wires. Fasteners go in if I’m fitting things into a tight spot and those fractions of an inch start to matter. I suppose the same could be said for wires, but that seems like overkill for the average FRC team.

We CAD some early prototypes and we also just build some early prototypes. The more experimental a prototype is the more we might try a few rough ideas quickly without CAD to get feedback as soon as possible. We do like to CAD prototypes in many cases, though, because it allows us to parallelize fabrication as much as possible (we’re probably 98% in-house). If a mentor or build lead needs to be describing what needs to be fabricated or how it needs to be assembled on a prototype without CAD, it’s hard to involve many newer build team members. If everyone can see the CAD, it’s pretty reasonable to have as many as five team members working in parallel to fabricate parts or assemble things. There are also some prototypes that really benefit from one or more CNC’d parts and those are going to have to be CAD’ed even if the rest of the prototype does not go through CAD at the start.

Our build is prototype-driven. We do not CAD very far ahead of the prototypes. We need to see things working (or not working) in real life. Maybe some teams have the ability to just know that a CAD design is going to work but we consider that too risky.

After the kickoff we do some strategy so that we know exactly which systems are gonna be on the robot, then we set some constraints / goals for each sub-system (e.g shifting drivetrain gearbox, translating climber, etc) and then we CAD our prototypes (for most subsystems we do not make a prototype version, for others we may prototype a handful of different prototype versions) after that we make the final cad of the robot (we try our best to cad everything - we made the mistake of not doing that in the past and it lead to many mistakes that could have been avoided)
Good luck with CAD it will be a great tool for you team!

We use solidworks, during the first weeks the electronics team doesn’t have that much work load, so that’s why we do it, and in SW we do the cables with just 3D sketches with some sliplines

Week 4 of build season, a featurescript does a deep analysis of all open alliance and public onshape models and designs my robots in full.

11 Likes

My team used to do basically all of our CAD after finishing the robot (aside from volumetric boxes), with most of the design being done on paper or on whiteboards. Once we got a CNC router for a measly $3000 or something like that we lost all ability to machine things by hand but we do CAD absolutely everything down before making it, since we sort of have to. It feels a little like it might be bad how we’re so dependent on our CNC but it has made us make substantially better robots and has freed up our design restrictions and also our rookies’ time, since we don’t have gussets for them to make anymore.

That being said, we go straight into prototyping using 8020 or plywood and when we think we have something worth machining as a sort of V1 of the mechanism we use the geometry we learned from the prototypes, CAD a full mechanism, and throw it on the CNC. Generally this leads to us having a robot with a V1 of most of our high-priority mechanisms within 2 weeks, which is great for driver practice. We then continue iterating, cycling between testing, thinking up adjustments, implementing the adjustments in CAD, and then cranking out a new iteration on our CNC.