Time to make a model
 

Hi Everyone,

Here is a question for teams that fully CAD out their robot before building it: How many hours does the model take (it would be helpful to include the number of people and distribution of hours)? Can this time be decreased with more CADers?

Thanks for your time,

- Lemiant

Re: Time to make a model
 

I have never done the entire robot in full detail myself. However, to help get a better answer, you might answer some questions. How detailed does it need to be?(down to the metal, screw, wire...?) Also, how complex is your robot? Some are much more complicated then others. A picture might help approximate difficulty and time. As far as more people making less work, I've not had the best experience with this. It works great if you have a specific method of doing it and work closely together. It can definitely cut back on the time if done well. If done poorly, it can take just as long because parts will need redrawn if the dimensions don't line up or communication isn't kept.

Re: Time to make a model
 

Yes, I am thinking full detail, enough that sheet. Metal parts will be waterjetted with holes for rivets. I am wondering about both a full FRC competition robot, and also a robot which is just a chassis with a single simple manipulator (also the chassis will be heavily based on CADs from sheet metal teams).

Re: Time to make a model
 

I have done the complete CAD for the past 2 years. I have to tell you, it takes a *lot* of work. I suggest spending preseason teaching 2 to 3 other team members who you trust how to use Autodesk Inventor, so that they will be prepared for season. The way I generally do it is CAD everything needed for waterjet cutting/ manufacturing. We send all of this out as soon as possible (end of 2nd week to beginning of 3rd). This tends to require about 25 hours of work. It includes the assembly, all part files, file structure, part drawings (most important aspect of CAD. Make sure you are really good at these). This starts with a chassis, btw.
After we meet our deadline, I begin to CAD everything else. First comes E Boards, with all of the modules on them. Then comes chain, wheels, pneumatic hardware, part coloring (I can get obsessive with that). Eventually I will do the pneumatic runs, and last the E Wiring (which I only do an overview of. Theres no reason to get meticulous there).
All in all, I would say that CAD takes upwards of 100 hours, if you wish to do EVERYTHING, and prepare a submission to the Excellence in Design Award. You'll see this year's CAD below, for reference.


[/url]

Re: Time to make a model
 

If you know your CAD software really well and already have a design with zero issues memorized, you can probably model it in a day or two. However this is rarely the case in reality, as the reason to use CAD is to find problems and fix them before spending money making physical prototypes. So the time it takes to model the robot is proportional to exactly how much you design everything ahead of time and fix problems.

On 228, we only model what we need to build the robot, which is primarily everything mechanical (and large control system components for mounting locations and weight/volume considerations). Adding screws, wiring, pneumatics tubing, etc. doesn't necessarily help us, is very time consuming, and in an assembly already with several hundred parts, adding these really start to bog down your computer.

Instead, what we do is assign a part number (such as 11-DRV-012) to every custom or modified part on our robot, draw up a 2D machine drawing for said part, and either send that (along with the CAD model) to our machine shop sponsors or print it out tobe machined on the manual mill or lathe in our shop. We then put all of these part drawings into a binder, which theoretically could be used to build an exact clone of our robot.

Here's the CAD model of our 2011 robot. This was probably about 80-120 man-hours of work, with the average probably around 4-6 man-hours of work a day for the first three or four weeks.

Adding more people can help or can hinder (it all depends on communication and organization); you will need a robust revision control scheme to keep all changes to the CAD models updated and propagated to other group members. Things being designed in parallel also need careful attention paid to interfaces between the subsystems, and making sure they do not try to occupy the same space, and that the still work as a system.