In previous years, when I quickly prototyped our robots for spacial constraints, I usually started with large 'blocks' that represent the drive train, mechanisms, etc. Then I extruded out holes from the blocks in order to represent the Aluminum Bar. Then I added holes where necessary. I could get a good concept of everything in about an hour and a half. Then I'd start a new assembly using a bottom-up approach. This is how we did our '08 bot. The frame was one of the more successful and robust frames we've done (2nd to our '11 bot). The methodology was useful for answering questions to our "can we do X" questions with regards to spacial constraints.
An example of this for our 2010 bot is
here. The blue volume was eventually cut into for frame rails and other constraint testing, then I made new parts for the lower frame and attached them to the purple frame in that photo.
Another edit -- I'll also point out that I made that blue volume into a shell, which I then put around the [semi] detailed CAD as we went along through the design. That allowed me to verify that our kicker (for example) would also clear during a bump traversal.