CAD?

Well the devastators are planning on starting a CAD team so what are some tips that you guys have? How many people should be in it? How long does it take? What software do you guys use? Can we use inventor? Anything else that could help us start this team would be great thanks!

Practice is a huge thing. The more experience the members have working with the programs the faster the modeling will go. As far as modeling software is concerned, there aren’t significant differences between Autodesk Inventor, Solidworks or PRC Creo. You get all three for free through FIRST (I think), along with Inventor being free for all students regardless of FIRST affiliation.

With two people modeling (one person with a semester of experience and another with 2 years) it took is 2-3 weeks to get a full model done, but with more experienced members and probably 1-2 more people modeling we’re hoping to get it down to 1-2 weeks

Definitely start learning as soon as possible. You do not want ro start learning during build season. You can use any program you want to. We have used inventor in the past however we switched to CREO.

One thing that I have found is that knowing how to use CAD software, and how to use CAD to design is a big difference. I have seen many students model something in CAD that would be very difficult or impossible to actually make. Be sure that the people on your CAD team know the manufacturing capabilities of your team (i.e. sheetmetal, 3D-printing, extrusion frames, etc.).
Also, make sure that you have a computer that can run the software. For best results, you will want a full workstation built specifically for CAD, which can get expensive. However, for FRC type stuff, you can get away with lower cost consumer computers. Fast processor, dedicated graphics, and a good amount of system RAM are important.

Like what was said earlier, start practicing now! It will benefit you enormously!
My team uses PTC Creo because we have mentors who know it very well and you can get it for free from PTC. They also have pre-modeled parts including the entire KOP drivetrain which was super helpful for us this year because we could immediately start designing ball mechanisms.
My team had me and one other person on CAD this year. It didn’t work out to well because I also do programming so I had to be in 2 places at once most of the build season. I would recommend having about 3-4 people who know CAD very well, maybe more depending on your team’s size. It’s very important that they work very closely with your robot design people.

Hope this helps, and if you have anymore questions feel free to ask!

Things I would like to emphasize, through personal experience:

  1. Practice is important, as other people have mentioned. Make sure students are familiar with all of the modeling functions within whatever program you are using (I am just going to use Autodesk products for the sake of simplicity). Extrusion, Hole, Sketching, Constraining, all of that jazz. It will make sure students do not become overwhelmed with seemingly difficult parts to model.

  2. Organization is CRITICAL in order to avoid problems. Inventor really likes when you keep files where they were and not moved around over the course of the build season. If your students need to move a file, have them save a copy as a part or assembly and save it to wherever necessary; even then you are at the risk of having path issues (Inventor loads a part from a certain file, but cannot find the file because it has moved).

If you are using the Autodesk Vault program (not sure if it is free or not), all of what I said can be avoided. Loading files and creating files in the vault, Inventor will not care about pathing.

  1. Efficiency is important, and this one I would like to emphasize because I wished I could have when I was on the team. Within designing a machine, changes will happen constantly, and you want to change the model as quickly as possible. It is not efficient for it to take a half an hour to change a mounting plate. If you can simplify the way something is modeled, I high suggest doing it (situation permitting). There is a really big difference between being able to CAD and being able to CAD well. Here are some ways that I have found out to be efficient.

One way to do this is set as many dimensions as functions as you can. Minimize the amount of hand changed dimensions as you can. Here is a simplified example: If you make a square, you would have to dimension all four sides. Instead, you can dimension two sides (x and y), and then dimension the opposing sides as functions of the already dimensioned. That way, changing one dimension will change two instead.

Take what you want from this post. It is a lot of information, so take it one step at a time. The last thing you want to do is overwhelm/stress your sub team. This should be a fun experience for them (and you) to learn about modeling. I wish you luck on your endeavor. If I can think of anything else, I will edit my post.

tl;dr: Be prepared, be familiar with your program, be organized with your parts, be efficient with your modeling, and most importantly, have fun and make sure you are learning. :slight_smile:

Ideally your design people are the ones CADing. After all, CAD is a tool for designing.

I’m glad you guys are doing this! As said practice is a must. Going on to programs, Inventor and SolidWorks in my opinion are the better programs, but SolidWorks is used in industry much more. Since you are starting from scratch I would advise to use SolidWorks. Once you have that, I would also recommend sectioning off members. One good idea is to have people to design and other people for doing prints for manufacturing to split up the work load. Only have two or three people max actually DESIGNING the robot though. Mentors will always be with them discussing design solutions, which can make working with more than 3 students very hectic :slight_smile:

I would go as far as to say that you have to make sure your design people are invested in using CAD before you start using it for the build season. In 2012 and 2013, my team’s design and CAD groups were separate, and it caused nothing but pain for both parties.

The end of the season is the perfect time to start building your CAD/D team-- you have plenty of time before next season. CAD is not something that can be learned effectively overnight-- it requires patience and effort, like any other skill. A lot of kids are intimidated byt the steep learning curve, but I personally think it’s one of the best technical skills one can learn on a robotics team.

The other thing would be, as Eric said, efficient parts design. Your designers should be intimately familiar with the contraint tools of your program of choice-- both sketch constraints and assembly constraints. Good “model intent”-- ie, designing for changes-- is key to having an efficient CAD team.

If your team does not have good model intent practices, to quote the immortal words of South Park, “You’re gonna have a bad time.”

Communication with the build team is key. Otherwise, there will be two very different robots, one physical and one virtual.

Develop a system for part numbers and revisions and stick to it.

Never use names like Intake Left Arm Outer Side Plate 2.sldprt as this makes it very difficult for BOM and PDF drawings when you send parts to sponsor machine shops. Names like 2014-110-001_Rev1.sldprt or 2014-110_Rev8.sldasm are much better and better prepare you for how companies in the real world work.

Don’t be sloppy on your modeling. Constrain ever sketch entity - if you open a Solidworks sketch there must never be any blue lines.

Don’t use project/convert geometry from an assembly into a part. It’s great and fast when it works the first time but when you iterate and change designs it will be a minefield of broken references that will take for ever to fix.

We had this issue this year. Can you explain a bit more about how you organize the part numbers so that the team can handle them?

If you are starting fresh, I would recommend trying out Solidworks. I like it better than Inventor and, at least where I live, there’s a lot more real world demand for it.

Look online for tutorials. Many FRC teams have put up FRC specific cad tutorials that are very helpful (973 RAMP being the first one to come to mind).

Another thing to keep in mind, is that a team sould never restrict students to just the “CAD team”. Make sure that the people deciding how “it” fits together have experience putting “it” together. The best practice you can get is sitting down with your machinists/builders (they don’t necessarily have to be the ones cadding but it will benefit you if they know how) and design and build a prototype robot in the off - season. Anyone can put up a CAD image on cd, it takes skill to design a robot effectively with your resources in mind. Doing it in the off season lets you learn from your mistakes with out hurting your team’s productivity.

If you have any solidworks specific questions feel free to pm me, and of course others who post cad specific details on cheif.

I used a similar standard when I was on 228 as we currently use on 148: YEAR-SUB_SYSTEM-PART_ID. The top level assembly is 000, and each major subsystem gets a number like 100 for drive train, 200 for mechanism 1, 300 for mechanism 2, etc. Sub assemblies within that mechanism get numbers like 110 or 320. Each individual part is assigned a unique ID in the lowest assembly it is used in. If parts are COTS, you use the the manufacturer part number.

2014-000 Rev1 (main robot assembly)

  • 2014-100 Rev1 (entire drivetrain assembly)
    – 2014-100-001 Rev1 (drive train part 1)
    – 2014-100-002 Rev1 (drive train part 2)
    – 2014-110 Rev1 (gearbox)
    — 2014-110-001 (gearbox part 1)
    — McMaster 91251A342 (#10-32 x 0.5in SHCS)
  • 2014-200 Rev1 (mechanism 1 assembly)
    – 2014-200-001 Rev1 (mechanism 1 part 1)
    etc

This is also really useful because in Solidworks you can set who has write access and who has read-only access, so you can have multiple people working on the CAD at once (one person could be editing -100 assembly and sub parts, another editing -200 assembly and sub parts, etc). The key to making all of that work is by opening all referenced parts as read only. To do that, check this box:

[img2]http://www.chiefdelphi.com/forums/attachment.php?attachmentid=16814&stc=1&d=1397394568[/img2]





Take the summer to CAD something like last years robot or a battery cart or something like that, it really helped me to prepare for Build Season this year. You also don’t necessarily want too much of a good thing, I’ve found that more than a few people working on one CAD can get to be tricky if you’re not all on the exact same page. Also please save to multiple places, you don’t want to lose all that work. Good luck with CAD!

Is there any link whatsoever between the programming language we use on our robot and the CAD software we might use to model it?

Not at all. They are separate programs and there is no connection. The type of software your team uses really depends on prior experience.

Thanks!

Another thing to do in the off season is to make COTS parts in Solidworks (or whatever your team decides to use) your team uses a lot. It will save you time during the season. You can adjust the masses of certain items and even apply materials to make the look better for a render if you choose…stuff you shouldn’t waste time on during the season that costs nothing to do now. I’d suggest CADing up a drive base or two in the off season. Not only will it make you more proficient in CAD, but it can provide some insight on drive base design that will help during the season (common pitfalls, difficult to CAD parts, etc.).

Honestly, this ties into the “constrain everything” practice, and while I agree it’s an ideal to strive towards, it’s not always feasible.

Certainly, any sketch you construct from scratch should be fully constrained, but there comes a point (complex mounting patterns transferred from .STEP files that you had to get from the internet, for example) where constraining everything will do nothing but give you a headache. Any time you need to incorporate geometry that you yourself did not actually design, you’re going to run into this. It seems clear to me that the best solution is not to re-create all of the geometry from scratch.

Fortunately, those aren’t things that you tend to make iterative changes to, as they’re not something that you are actually designing, so you’re not going to lose much by being “sloppy” there.