6 Replies Latest reply on Jul 27, 2008 8:29 PM by John Kreutzberger

    Unexpected result with planar surface

    John Kreutzberger
      I saw something interesting yesterday when working on a fairly complex parting line for a plastic mold design. I use the mold tools to a certain extent, but almost always create my parting surfaces manually and that's when this came up. In one area, I indentified a portion of my p/l that would be planar. Therefore I created a plane and opened up a sketch on it. This sketch had some converted edges from this swoopy part , some converted edges from adjacent surfaces and some lines. I noticed later that this surface ends up being not absolutely parallel to the plane that I sketched on to create it. It's usable anyway because the angle is 0.00000035°. Being pretty O.C.D.-this bugged me for a while, but I finally got over it and moved on. I only noticed that at all because I tried using that surface for mating when placing it in assembly. The mate failed when I used the surface, but was fine when I used the plane that the surface was created on.

      Knitting was not the problem because I rolled back to before I knit my p/l together and it was still not perfectly parallel. Any thoughts?

      (I cannot post the file because it is for Boeing, and they are touchy about that sort of thing.)
        • Unexpected result with planar surface
          Charles Culp
          Did you create a 2D sketch on the plane? I am completely boggled how a 2D sketch on a plane, with planar surface, wouldn't be planar with itself.

          Also, I have found that "filled surface" often works slightly differently, so you could use that instead and see if it works.
            • Unexpected result with planar surface
              John Kreutzberger
              YEP-It was a 2-d sketch on a plane I had created specifically for that purpose. Then after using the sketch to create the planar surface I can measure the surface against the sketch plane it was created on and SW tells me that the angle is 0.00 degress-NOT that they are parallel. After expanding the angle precision to the max, I learn that my `planar' surface is at an angle of .00000035° from the sketch plane it was created on.

              Boggled my mind too-that's why I posted it. I ended up re-doing it with an extruded surface on a plane perpindicular to the first plane that gave me the expected result. All I can think of is the converted edge from the model may have had some influence on the planar surface.
            • Unexpected result with planar surface
              Matt Lombard
              JK,

              My first thought was that a knit caused it to tweak a little bit, but you said it was measured that way before the knit.

              Second thought is to maybe not wonder why, but to knit in the questionable surface, then use another truly planar surface with a Replace Face to fix it.

              Sometimes converted edges can get squirrelly. Is there anything you can do to simplify the converted edges? I have seen surfaces sometimes go "black hole" on you, when they seem to suck down into a single point, leaving really bizarre geometry that you never intended. Maybe convert the edge as a spline and use the Minimum Curvature spline gizmo to see if there is a place where it is kinked with tiny curvature. I'm not sure what you'd do if you found it, but at least you'd know where the problem was coming from.

              Stuff like this leads to sloppy workarounds. If you're working on all native geometry, you can sometimes go back and clean things up so they are less questionable, but you mold guys work from imported junk a lot, and that can be the worst. If it's imported data, maybe go back and run an import diagnosis, or heal edges or something. Just grasping at straws here, really, but these things sometimes have helped me. I'm sure you've thought of all of this already.
                • Unexpected result with planar surface
                  John Kreutzberger
                  Hi Matt,

                  This was imported from CatiaV5 and was a total mess. ImportDiagnostics cleaned it up so that it was a valid solid. Then I had several surfaces that needed to be parallel or perpindicular for us to be able to build a mold that were not quite parallel or perpindicular. Therefore I used replace face to clean all of that up.

                  The situation that prompted my initial post was due to using the converted edge in the planar surface-I am sure of it but cannot say why it happened. (there are enough surfacing tools available now that I am learning to take different approaches when my first idea yields what I call here `unexpected results')

                  Knitting does indeed tweak the surfaces sometimes but that was not the case here. I was in a session with Mark Biasotti in San Diego and he seemed to say that this was due to tolerances being loosened enough to allow things to happen that would normally error out. Fine-I'll buy that and it does come in handy sometimes. Sometimes that does also inspire what you call sloppy workarounds. Hacking and whacking is still a valuable `skill' to have, but I try to explore more elegant solutions first.
                • Unexpected result with planar surface
                  Jerry Steiger
                  John,

                  I just have to ask. What is O.C.D.? I have a pretty strong feeling that I am whatever it is, since stuff that is off by 35 parts per 100 million will drive me nuts too.