9 Replies Latest reply on Feb 11, 2008 7:20 AM by Coen de Winter

    Need some Help: surface won't thicken properly

    Coen de Winter
      Hi guys,

      This is my first post on this forum. I seem to have a little problem. I'm drawing a 1936 hydroplane in solidworks 2008. Just because it would be a nice challenge. I have a complete surface made according to my table of offsets. This surface represents the outside of the planking of the hull. The deck and sides are 6mm thick and the bottom is 8mm thick.

      The problem is Solid won't thicken the surface with a simple Thicken feature.

      If I mirror my surfaces to get the whole boat, and check the form solid box, it will form a nice solid. A shell at this point will actually succeed, but the inside surface will be no good at all. I cannot even copy the resulting inside surface because Solidworks tells me the surface is self-intersecting :S.

      I even tried to copy the original surface, than scaling it down. The problem here is that it's pretty inaccurate. I can't really get the position of the inside surface spot on.

      Eventually I found that the problem is in the surface for the deck of the boat. Solidwork won't offset this surface inside or out.

      I have attached the part if you would like to see what I'm talking about. As you can see in the part, I do manage to get the hull looking good. The problem here is the inside surface again.

      This one has really got me stomped now. Does anyone have a clue about what to do about this? Any help would be greatly appreciated.

      Greetings,
      Coen de Winter
        • Need some Help: surface won't thicken properly
          Jeff Mowry
          Perhaps separate the hull from the rest of it and work them as separate bodies. You can join them later. By the way, can you post a pic for thos of us who don't yet use 2008?

            • Need some Help: surface won't thicken properly
              Coen de Winter

              I tried to work with the surfaces separately as well. Solidworks just seems to be unable to thicken any of the separate surfaces as well.

              As you can see in the last picture I do manage to get the hull to where I want it to be. The problem is with the way the inside surfaces are created when I use the shell. I attached an eDrawings file, so you can see...

              Thanks for the fast response btw...
                • Need some Help: surface won't thicken properly
                  Ricky Jordan
                  Coen,

                  The reason that the inside surfaces of the shell are wacky is that the minimum radius of curvature of the surface body is less than the offset amount in your shell. It doesn't look like I'll have time tonight to look at this much more but I'm pretty certain this is the cause.

                  Best Regards,
                    • Need some Help: surface won't thicken properly
                      Ricky Jordan
                      Coen,

                      I found some time and solved the problem. I've attached a version which fixes the wacky surfaces on the interior. The root cause of the problem was in two areas. I've attached images of the areas below. The minimum radius of curvature in both of these areas was WELL below the thickness of your shell. You can use Tools->Check to check the minimum radius of curvature.

                      The cause of the minimum radius being so low was the use of a Boundary Surface feature with Direction Curve pairs adjacent to one another. (Meet at a common point - See image). When two curves are adjacent and in the same direction grouping, the Boundary Surface feature just doesn't do that well. I even tried the "Linear" option which is new in SW2008 that is supposed to address this. (See the What's New Manual - Page 4-2). The quickest way to fix this was to use a Surface Loft instead. You have to drop the "Normal" to profile end condition at the center plane.

                      I left the original Boundary Surfaces in there and created the Surface Lofts as replacements along with updating all references in child features.

                      Nice model! The only suggestion I have is to check the smoothness of some of the construction splines using the curvature comb tool. I can tell you put quite a bit of work into this. Excellent work!

                      Best Regards,
                        • Need some Help: surface won't thicken properly
                          Coen de Winter
                          Hi Ricky,

                          Thanks for the compliment and helping me out.
                          For some reason the pictures won't show and I can't download the file. But I did as you told, using the lofted surface tool, and it works great.

                          The design is by Willard Crandall, back in 1936. I'm making the 3D model because I thought it would be a nice exercise/challenge. I'll drop in the frames this weekend. Maybe I'll have it finished by the end of next week or so.. I could upload the model if anyone is interested in it...

                          Anyway thanks again!

                          Greetings,
                          Coen
                          • Need some Help: surface won't thicken properly
                            Just to add some clarification here, the boundary vs. Loft is not the problem. The problem is that by creating a loft or boundary suraface with the profiles (or 1st direction curves) sharing a common vertex, you will create a de-generate or singularity in the surface (look at the converging mesh lines that meet at the bow and stern) these types of surfaces will not offset or shell. Now in the case of Loft it will because the tolerance of the profiles (or guide curves) are more tolerant than boundary (which is more precise than loft) and in some case, even singularities will get pass the sheller - and in this case they do (We've enhanced our shell to do this over the last few releases.) Also, if Ricky had made the first profile "Normal to profile" (normal to mirror plane) it would have failed the sheller because of this singularity problem and it looks like Coen needs to have it Normal-to-Profile, at lease for part of the hull, and that's why you had it on and that's why it fail.

                            It's unclear whether Coen wants a distinct edge down the center or not, but from the likes of the 2nd direction curves, he does - but there must be some transition to sharp to make this work geometrically at the bow and the stern
                            If you want to these types of surfaces, its best to use fill which does not suffer from singularities (when optimize option is unchecked) Either that, or do it as a loft or boundary surface and then trim the ends where the singularities are and put in a Fill surface. (see attached.)

                            Also, the Linear option has nothing to do with make the Boundary "non-singular"; it is meant for providing a alternate solution between curves (or profiles) that have dissimilar entities.