Cary L. Brown wrote:
I haven't used AutoCAD in ages, but I'm curious about something. I seem to remember that the "3D capability in AutoCAD was fairly limited... mainly just linear extrusion of 2D sketches, no "complex curve" capabilities. At least in the early implementation... and more recently, AutoDeck treats "Inventor" as their preferred 3D solution.
So... for those "complex curved" shapes (ie, hull walls, etc)... have you figured out how best to do those using your toolset? Or, am I mistaken about how this works, and it has that ability well-implemented already?
Complex curves are a problem. Especially for modeling. My approach is to use the 3-d edge-surf to create a high-density mesh for say a warped or twisted sheet of paper (or metal, or ...), and then explode that into it's constituent parts, and reassemble as a solid. It is not great, but it works. I have long since gotten used to the limitations on the complex 3D curves, and simply work around as best I can. For this project, if I have any such curves, I will likely create them as basic meshes, and leave it at that. I cannot think of anything on this project though that will be that complicated. The Secondary Hull is essentially a 2-d curve rotated about the axis, with parts cut out and other parts added on. The primary hull is essentially a 2-d curve spun about a different axis, but is far simpler. I can only think of the three "bumps" around the deflector (Port, Starboard, and Below) that might have issues - and I doubt I'll have much problem with them.
Now Excelsior, on the other hand, might be a bit more challenging. A Galaxy Class even moreso. My solution for those though is that the mechanical implementation of such ships will still follow a basic mathematical approximation, with the ribs of the hull defining (rather than describing) the contours. Hull plates are basically flat or slightly warped pieces attached to the ribs. a "linear" breakdown of a complex curve is acceptable, if the edges line up with the underlying structure.
AutoCAD before R12 was very
limited in 3D capability. It was as you say, extrapolation of 2D lines. Some basic meshes could also be generated. R12 introduced the solid-modeling capability, and R13 improved upon it. There are still limitations, and certainly it is not comparable to a modern version, but it does 98% of what I need it to, and I can "fill in the gaps" on the remainder manually.