12 Replies Latest reply on Sep 5, 2012 11:43 AM by Steve Ostrovsky

    Modeling at astronomical scales?

    Jay Andrews

      I'm experimenting with animations.  I work in aerospace design, so my animations are of spacecraft in situ.  Problem is, part models can't be over 1km diameter, but I'm trying to show a model of the earth in the background approximately to scale, and worse, to wrap a star field around the model as well.  Obviously 1km does not cover this.  Theoretically I could scale everything down, but my spacecraft are modeled 1:1, and lots of reasons to leave them that way.

       

      Is there a way to increase the allowable model geometric size?

       

      I read in another thread that sw11 has ability to do larger scale.  How much larger?

       

      Any other advice?

       

      Pic just for fun.  Scale is way low on this earth model.  And the background is just a background, doesn't rotate around with the rest of the model, which is why I want to model a backdrop, so the stars will move with the camera.

        • Re: Modeling at astronomical scales?
          Alin Vargatu

          As per Knowledgebase:

           

          In order to maintain a high degree of accuracy, SolidWorks limits the area in which geometry can be created in parts and assemblies. The total space available in SolidWorks (part or assembly*) is a 1000 meter cube, centered on the origin. Also, any linear or diameter dimension value must be less than or equal to 1000 meters. Radial dimensions must be less than or equal to 500 meters.

           

          *For SolidWorks 2009 and later versions, the total space available for assemblies has been increased to a 100km cube.

          • Re: Modeling at astronomical scales?
            John Burrill

            Jay,  all vectory drawing packages have limits of accuracy based on the scale of the scene.  In MicroStation you can't go larger than 2billion positional units.  In AutoCAD, you are limited to a total of 16 significant digits, so when you start modeling 10's of thousands of feet, it affects the precision of your decimal places. 3DSMax employs a system similar to AutoCAD's and Solidworks tries to hold accuracy to 1.0e-8 meters and limit the total size.

            What animators usually do is render out their scenes in layers.  for example, you'd have one animation of a rocket pitching for re-entry, another for the rising curve of the earth an another for the starfield that the camera rotates across and use an editing package like Avid or Combustion or Aftereffects to composite them together.  Trying to do everything in one camera is really difficult when there are great differences in scale involved.  Also in animation, there's a little trade-off between realism and dramatic effect that's much easier to achieve manipulating footage, than it is manipulating a scene.

            You might take a look at a free program called Blender.  It comes with a decent editor.

              • Re: Modeling at astronomical scales?
                Jay Andrews

                Thanks for the advice, I'll look into that.

                 

                Sounds a bit cumbersome.  I'm trying to imagine how you could possibly get the camera motions to match in the multiple animations before compositing, and also how to blend them seamlessly.

                 

                I'd still like to pursue it from the other direction of all in one animation, so anybody else with some feedback on my other questions would be helpful, and any other expounding on what John wrote also could be helpful.

                  • Re: Modeling at astronomical scales?
                    Jerry Steiger

                    Jay,

                     

                    How far away from the Earth are your craft going to be? It seems like the sphere of stars would have to be a lot farther away than that to "move" realistically.

                     

                    How small a detail on your craft do you want to be able to see? That would establish about how small you can scale the craft. (Just make another configuration with the Scale as the only feature.) Does the maximum size star sphere make sense with that scale?

                     

                    One other trick is that the maximum radius for a part is still limited by the 1 km cube. So the segments of your star sphere will be fairly flat. It will probably not be worth the effort to make them with any curvature.

                     

                    Jerry Steiger

                      • Re: Modeling at astronomical scales?
                        Jay Andrews

                        Geo orbit, which is not as far as the moon, but much farther than low earth orbit.

                         

                        I'm still considering scaling exactly how you describe, or I may just leave it as I have it, which is the earth is much smaller to scale than the s/c, but looks pretty good, and the stars in the background don't pan with the camera, I can maybe live with that too, but would rather have them pan and would give up a little quality for it.

                         

                        Not too big of a deal at this point, it's just an exercize, to see how well I can make an animation using solidworks.  I'm really interested in how much improvement SW11 would be in the large scale.  I'll have to try to find that.  Tough to get software upgraded around here though, even though we subscribe, I have the best ssystem in the co., and I'm still on '10.

                          • Re: Modeling at astronomical scales?
                            Jay Andrews

                            FYI, I found something kind of interesting.

                             

                            I went forward with scaling everything down to see how that would work out.  Not well at all!  Even with all of the image qualities cranked way up in parts and assembly, having the parts scaled down by 100,000x makes the mates extremely sloppy.  For example, concentric mates bounce around like crazy, say a 1" shaft scaled down, the concentric mate bounces around about three inches radially.  Also, even with HLR, all the hidden lines are showing through.  Not usable at all for presentation quality. 

                             

                            I'm not surprised at this result, but just sharing it.

                             

                            It looks like I'll have to either be out of scale, and/or do composite animation with something like blender, as described above.

                    • Re: Modeling at astronomical scales?
                      Steve Ostrovsky

                      If you really need a model of the solar system and our planets with their orbits, just use this fun little model.