Making CAD Models Faster

[MikeFinn]

Our team always tries to make a full CAD model before we start fabricating most of the robot. Unfortunately, it took us way too long to come up a full CAD model this past season and the robot was thrown together at the last minute. Now I'm the new lead of the design team and am trying to figure out ways to improve from last season.

I tried to brainstorm different ways of getting a final design done faster, but I feel like I'm missing some things. It seems like many teams are creating the final design while the prototyping groups are still testing out different mechanisms, and I'm wondering how that works. (Robot in 3 Days seems like a good example)

Here are different ideas I had come up with that might be beneficial to finishing a model faster. We plan on / already have implemented all of these ideas.

1: Keep the design simple
2: Use COTS with a downloadable CAD file as often as possible
3: Increase the size and knowledge of your design team during the offseason
4: Spend a lot of time at home making the CAD model, even if it means staying up all night
5. Make it easy to share files between members
6. Take advantage of R14 by using parts and assemblies from previous designs

R14 Software and mechanical/electrical designs created before Kickoff are only permitted if the source
files (complete information sufficient to produce the design) are available publicly prior to Kickoff.

Any input or criticism is welcome and appreciated.

[Ty Tremblay]

I would strongly recommend against #4.

Break your robot into parts. After you've finished the drivetrain and figured out how to mount things to it, start building your drivetrain (or use the kit chassis and save a ton of time that way).

Parallel processing is your friend. If you can find a way to build and CAD at the same time it's a big help.

[Cothron Theiss]

Thankfully, you're already taking the first step which is planning and preparing. But I'll go through your points and add my advice.

Here are different ideas I had come up with that might be beneficial to finishing a model faster. We plan on / already have implemented all of these ideas.

1: Keep the design simple

Simple, easy to CAD, and easy to manufacture. It's easier said than done, but it's a great start.

2: Use COTS with a downloadable CAD file as often as possible

Go ahead and download part files of stuff you'll want to or think you'll want to use now. Download the KOP when it comes out. Download fasteners from McMaster Carr if you plan on including fasteners in your CAD. Download these finals before Kick-Off so you are just a little more prepared when it comes to that time. One thing I also did was, after converting the STEP files to SolidWorks files, I went back and added the weight, material, and any mates or constraints to assemblies. It's an oft overlooked step (Get it, STEP?), but it can save you a ton of time and headache later down the road.

3: Increase the size and knowledge of your design team during the offseason

Yup. I can say from experience that one student cannot do all of the CAD and design work and be fully successful. And practicing CAD is far more important than people realize. Your familiarity and skill with CAD determines your speed more than anything else, so use it well and use it often.

4: Spend a lot of time at home making the CAD model, even if it means staying up all night

No.

Will you have to spend time at home to polish and clean up your CAD models? Quite possibly. But should you be doing heavy CAD at home? No. For one, it's always good to do heavy design and CAD work in communication with the rest of your team. But you should NOT be sacrificing sleep, your social life, or your grades for robotics. On top of that, you'll quickly find your productivity plummet as your hours stack up. Sleepy CAD is sloppy CAD and sloppy CAD is a waste of time.

5. Make it easy to share files between members

Yep.

6. Take advantage of R14 by using parts and assemblies from previous designs

Yep. Similar to downloading COTS files, any work you can do before Kick-Off just makes your life during Build Season that much easier.

One of the best ways to speed up your CAD work is to mitigate and reduce mistakes. At some point if not many points in the season, you will change something you had not planned on changing. Whether this is because of mistake or because designs changed, it doesn't matter. But an extra 15 minutes of work when you first create a part can end up saving an hour when you have to go back and change it.
When I do CAD, I almost never leave sketches undefined. It doesn't matter how insignificant a line might be, I'm gonna define it just so I know it won't do funny things when I start doing funny things with parts in an assembly.
Some people swear by keyboard shortcuts and specific optimization techniques. I agree that they can help, but I found myself using computers other than my own so often that I've grown pretty accustomed to just using the defaults mostly.
Develop naming and storing standards within your team. Do it. Agree on it. Adhere to it. Make sure everyone knows how to upload or transfer parts and assemblies correctly. There's nothing more frustrating and more common than opening an assembly and not being able to view all of it or the mates not working because some parts are in some other folder that you don't have.

Good luck, and happy CADding!

[Michael Hill]

I do a lot of things in the offseason that helps to CAD faster during the season. I've built up a decent library of COTS components, all with proper material and mass details (gears, bearings, sprockets, etc) and corrected features That way I don't need to waste time going to vendor sites to download step files, fix them and look at bunch of boring grey components. Second, I do a lot of CAD practice in the off season making drivebase a, gearboxes, or other components. That way, all (or most) of the problems are already solved for the season for a particular design. Lastly, I watch a good amount of YouTube videos on new SolidWorks features and look out for things that can save time. Another good thing you can do is download other teams' CAD, not just to look at, but examine why they made the design choices they did so that you can apply them to your own design (but don't outright copy, that's not cool). A lot of problems in FRC that you'll face have already been solved, it's just a matter of building up a knowledge base that you can pull these solutions from.

[MikeFinn]

Thanks for all the feedback!

I was hesitant to put on #4, because of the issues described here. However, I know our design lead the last couple years has been forced to stay up late CADing, and I just assumed other teams did it too. Of course, our team has done it in the past because we basically had one person CADing the robot the past few years. Because we spent our offseason training several more people, I'm guessing that working at home will be minimal.

Parallel processing is your friend. If you can find a way to build and CAD at the same time it's a big help.

Could you describe how this could be done? It's hard to make a final design while the prototyping teams are still figuring out the best way to shoot a ball or lift an object.
The closest thing I could think of is to have the mechanism designers "work part-time" as members of the prototyping team. That way they would understand what it is they're designing and possibly even start on the design work while its still being prototyped.

One of the best ways to speed up your CAD work is to mitigate and reduce mistakes. At some point if not many points in the season, you will change something you had not planned on changing. Whether this is because of mistake or because designs changed, it doesn't matter. But an extra 15 minutes of work when you first create a part can end up saving an hour when you have to go back and change it.
When I do CAD, I almost never leave sketches undefined. It doesn't matter how insignificant a line might be, I'm gonna define it just so I know it won't do funny things when I start doing funny things with parts in an assembly.

Thanks for pointing this out. I know I've adopted this practice, but I really should get the other students to do this.
Example: Our team is 90% sure that we are using a WCD drive base for next season, so I made an 8WD out of VersaFrame and COTS gearboxes (we aren't at a level to make sheet metal chassis or custom gearboxes) where we can easily change the dimensions, wheel spacing and gearbox placement just by changing a few numbers. I'll have to put it on GrabCAD later, but if we decide to go with this design, we could easily have the drive base designed in Week 1.

I've built up a decent library of COTS components, all with proper material and mass details (gears, bearings, sprockets, etc) and corrected features

We're building up an Autodesk Vault library of COTS components, but I never thought to change the mass and material before we upload it. This is a really good idea, but I don't think we'll have the time to do it for this build season. Definitely something to try next offseason.

Again, thanks for all of your help!

[Ty Tremblay]

Originally Posted by MikeFinn

Could you describe how this could be done? It's hard to make a final design while the prototyping teams are still figuring out the best way to shoot a ball or lift an object.
The closest thing I could think of is to have the mechanism designers "work part-time" as members of the prototyping team. That way they would understand what it is they're designing and possibly even start on the design work while its still being prototyped.

I absolutely recommend having your mechanism designers be part of the prototyping team. In my opinion, teams should work on prototypes until they have at least one that they feel performs as it should at a competition. (For example, if your 2016 prototype is sinking 9/10 balls from the same spot, you have a really nice prototype.) After the prototype is "finished" is when the CAD should start. This will make CAD much faster since those doing the CAD will understand the design.

Where most of your parallel processing advantages will come, however, is NOT from building prototypes while designing the final bot, but from building your bot while finalizing the CAD. One possible solution could be:

• Design your drivetrain first, keeping mounting points, clearance, and interference in mind.
• Once your drivetrain design is done, build it while designing your manipulator(s).
• Once your manipulator(s) design is done, build it while finishing your drivetrain.
• Integrate the subsystems.