To CAD or not to CAD?

I’m wondering what to do and I’d love to hear what other groups are doing.
What level of CAD detail do you have?
Do you put screws and nuts in CAD? Do you put chains in CAD?
Do you put all electrical (including special connections) in the CAD?
If so, what are your reasons? And where do you find their files (screws, chains, etc.)?

There’s usually very little reason to CAD nuts and bolts… the only reason I think you could justify doing so would be if you’re worried about clearance in a particularly tight area. Similarly there’s little reason to CAD wires as they can generally be routed to fit wherever they’re needed.

We use CAD to enhance our robot design, not to create busy work for students. We’ll do the minimum amount of detail up front to create the necessary components and have a high level of confidence in our manufacturing processes. Sometimes if we have time some students will go back and fill in some unnecessary details like hardware, but they just do that for fun.

Electrical components are a different story… they’re larger and can throw a bigger wrench in your design plans if you forget to allocate space for them. We always download the CAD files of the major electrical components and fit them into our top level designs to make sure they fit.


The main reason to CAD is to better prepare for the build process.

Creating CAD models can help you solve problems with packaging, geometry, and integration before a hole is drilled in material. Virtual holes in material are free…

CAD everything you need to. We don’t add our electronics to our 3D models yet, but we do plan out the layout of the electrical board in Visio. In the future when we have more resources in the CAD team we will likely add everything to our model.

I have never asked my kids to CAD the chain/belt runs… but it is important to calculate the center distances for whichever you end up using as that is of critical importance.

I do not also enforce 100% accurate fasteners in the models we make. We do however discuss at length and notate what fasteners we will use.

I’m not sure about the nuts and bolts, but I do know that we CADed our entire electrical system this year, including the wiring. There were multiple reasons behind this, the main reason being that we didn’t want a messy electrical board, and wanted it to look intentional.

1 Like

My opinion has always been that we need to CAD with whatever level of detail we need to make sure it will work, no more, no less. So we don’t CAD chains link-by-link, but we’ll put in a single chain part with the same dimensions to make sure it doesn’t accidentally intersect anything. We don’t insert the complicated CAD models of electronics parts, but we use boxes with the same dimensions and mounting holes to make sure everything fits properly on the electronics board. We don’t CAD nuts, bolts, or rivets in wide open areas, but we’ll add them in places where we think they might interfere with other parts.

This is what we do for our CAD design during the build season between prototyping and manufacturing. After everything is manufactured and assembled when the mechanical team has a bit less work, we’ll go back and add more detail to the CAD to make nice renders and sketches for t-shirts and promotional materials.


Do as much CAD as your team and schedule can afford.

Sometimes that means skipping modeling of hardware or chain. This can definitely burn you later, it is a risk, but if you are careful and short on time you can definitely skip it. I’ve certainly skipped it before.

I basically never model electrical in CAD; I’m sure the electrical teams I have worked with would prefer if I did. I instead just try to leave a lot of space for them, and I at least account for the battery.


Typically, we CAD model our electronics (atleast last year) to allow us to have our belly pan CNCed for faster assembly. It also helps ensure we keep our electronics packed tight without causing issues.

Belt and chain runs sometimes. It isnt a very detailed run if we do, it’s just clearance to ensure we wont be rubbing on something or intersecting something we should.

Bolts sometimes. I have recently but that’s just to make smaller models like swerve modules or gearboxes look complete. On an entire robot, very rarely will I put fasteners.

Mostly aligned with these comments. We would only do fasteners if we are concerned in a particularly tight area. Somewhat disagree about wiring. We don’t CAD all the individual wires, but we try and do both route planning and block out cableways to ensure we can efficiently wire the robot. In the past we have placed all the mechanical bits and then discovered afterwards it was very problematic to get wiring in. This was especially true with elevators, where we had a large cable bundle to route up to the elevator.

We do model and place all the main electrical elements and integrate the electrical board into the mechanical model. As Ginger_Power said, failing to do this can make assembly way more fun than it needs to be. This also allows the controls team to prebuild the electrical board and do a lot of the wiring which speeds up final assembly. One tip is to strip out/simplify the electrical models as some of them have a truly absurd level of detail.

Once the design has stabilized, we try and do a design review for manufacturability, identify potential issues and try and fix them before we start making the parts. The review is both from a part manufacturability and from an assembly perspective. Sometimes very small changes can significantly simplify things.

Since my team has a CNC that we use on a lot of parts and a team structure where we assign one designer as the mechanical lead for a whole subsystem (we have a big design team) we cad almost everything down to the screws and nuts, just to be safe. Better to spend a few seconds in CAD doing it then trying to fix it irl

Should you have detailed models of belts, chains, fasteners? No. Where I work I’ve rarely ever seen threads on models of fasteners. It’s just too many faces that bog down the computers. Instead of having every link in a chain or every tooth in a belt you can put in a swept profile of the space they will take with fewer faces and it will load quick and even better, will keep you from putting another part 1mm from a moving chain/belt because you didn’t see it in the assembly.

You should have enough CAD data so you don’t make a mistake that causes you to do a major rework of a part that’s already made or one that causes you to take too long to fix/replace a broken part at a competition because you left the CAD data of a belt, chain or fastener out of your assembly.

The more parts you leave off your CAD the higher the chance is you’ll make a design error. Every design error will cost you time.

If you don’t have time to do the CAD will you have time to fix the mistake?

1 Like

I’ve got just the tool for fixing mistakes.

It’s called a hammer. Often referred to on our team as a “persuader”. Because sometimes parts just need a little bit of encouragement.

1 Like

Not that I think it needs to be in CAD, but folks should keep in mind the difficulty fastener placement can present in terms of assembly and maintenance. You don’t want to design something that can only be fixed by that one kid with tiny hands with that one weird specially shaped hex key. You’ve all been there.


Two things I recommend for CAD in FRC
My assumption here is since you’re asking, your team doesn’t have a lot of experience/buy-in yet.

  1. Start with Onshape - No setup other than a login and creating a team environment. Not AS capable as the others, but more than enough for FRC. This will increase the buy-in tremendously cuz students can try it at home just to fool around and may end up liking it. Don’t get hung up on this package is better than that. I’ve been to three different companies and used four different CAD packages.

    1. If this is your first year in CAD, focus on mainly/only using it for brainstorming this year. Students without experience tend to have a hard time with size comparisons. They can see the shapes in their head, but then I have them describe it or draw it on the whiteboard. Sometimes I wonder if they’re gunna someday be that person trying to shove a ded yak into the overhead bin of an airplane. Simpler shapes with added dimensions go a loooong way in the brainstorming phase.
    2. If you have 1-3 students that can’t get enough CAD, have them focus on key components that fit within the simple brainstorming shapes this year. I’m thinking things like wrist/shoulder of an arm or moving electronics around to see how they can fit better. Or another good use is laying out the hole pattern for large sheets of plastic. Students can spend 40+ hours a week on CAD and still not have it finished before parts need to be ordered and cut, so don’t let it hold up the team for the first couple of years while you get settled in.

OK, three
If you HAVE to use something other than Onshape (or Fusion 360 since it has cloud storage) use GrabCAD. They are like Github for CAD and have plugins directly for Solidworks and Inventor.

This is exactly true in the real world. The thing is, after HS, the people that get hired KNOW CAD. In HS unfortunately, it is sometimes faster to cut twice rather than wait on the CAD. This is certainly not ideal, but the current position some teams are in. Higher tier teams (big and small) recognize this and do put a big effort into reversing this issue.

1 Like

I usually do for FRC, and always do for real world projects. They are really useful for checking clearances and having an accurate weight estimate of the mechanism.

These are fairly quick to add in if you make a single lightweight CAD model for each type of screw. E.g. #10-32 Socket Head Cap Screw, with every length option being a configuration of that base part. Suppress all threads to make the parts easier on lower end computers. This configuration approach is really useful to quickly change the screw length in assemblies, without having to redo mates.

For chains or belts in FRC, I usually model this as a simplified part at the maximum possible dimensions and manually set weight, like this:


SolidWorks does have a chain pattern feature to model actual roller chain, but these simplified chain runs are easier on your computer when you have massive assemblies open.

For FRC: components (and mounting points) = yes, wiring runs (except for zip tie holes) = no.

For real world projects: everything gets modeled.

If you use SolidWorks, McMaster is a great source to download editable part models of hardware. (They offer other file formats as well, I just can’t directly confirm if they are editable). The editable files really help when you want to suppress threads to improve assembly performance, or make a series of length-specific configurations in a single .SLDPRT file.


One last item:

Even if you decide to omit modeling fasteners in your assembly to speed up the process: ALAWYS add pneumatic fittings onto your pneumatic cylinders!

I’ve lost track at the number of times various FRC mechanism ideas or prototypes have failed to work in the real world, simply because the mechanism tries to rotate through pneumatic fittings and.or tubing.


Can you tell me more about this grabcad plugin for solidworks? Currently our team just uses the desktop grabcad application as a simple github for all our cad files. Just upload your changes and download someone else’s changes before making your own.

This is our CAD model from this year. We won an award on it. I don’t know if we modeled nuts and bolts, but I do know we didn’t model our wiring.

1 Like

Our team has been putting a lot of work in the past few years to improve CAD fidelity; we now model all hardware and electrical components with the exception of rivets. The trick to this is by putting in a little work in the offseason to streamline the in-season process - all of our electrical components have been converted to work neatly with SolidWorks native files, while we’ve also been building up a library of configurable screws and hardware spanning every screw from #4-40 through 1/4"-20. The configurable screws part is the most important - you can combine these screws with feature-driven patterns to make hardware modeling a total pushover. It takes only a few minutes to detail every piece of hardware in a subassembly when all the tools are available.

Modeling wires is inconvenient and not really worth your time, but it very much can be worth your time to model energy chain and the necessary rigs/mounting features to make it work. Using patterns, this is actually pretty straightforward. In my opinion, the hardest part is getting the models for the links, which require you to dig through Igus’ website for at least twenty minutes.

The GrabCAD plugin is essentially identical to the standalone application, it just means you can upload/download without having to open another application. The GrabCAD dev team seems much more interested in developing and supporting their desktop application than the plugin, so our team uses the desktop app.

I’m playing around with wire routing using Solidworks and it’s taking a long time just to get one wire routed in 3D. I still don’t have the routing like it should be and I have some more things to learn before I get to the point where I can know the outcome before I click the buttons. The good news is that the component library and wire library is very small for FRC, so that shouldn’t take long to finish.

I played around with SW routing last summer and frankly, I wasn’t impressed. At the very least for FRC scale, it’s faster get your wiring lengths just by using a tape measure. As long as you leave enough room for a wire bend radius, you can pretty much always count on being OK with the amount of room you need between components.