5 Replies Latest reply on Jan 26, 2010 2:34 PM by Andrew Troup

    Thinking of using Face Fillet w 2 hold lines for surface model?  My advice: Don't!

    Andrew Troup

      I've just comprehensively proved to myself that this is strictly a notional implementation. You can generally get it to work, if you jump through enough hoops, but it's just not worth it.


      The functionality, and prerequisites, appear so closely predicated on the solid case, that it's massively easier to build a solid 'rail', fillet it, then convert it back to surfaces.

       

      If you choose to disregard this recommendation, it will help to know what is meant by a 'laminar edge' (cryptically raised by the software to thwart you at every opportunity, and not mentioned in 2008 Help. Back in 2003 it rated a brief mention, which is how I knew what it meant).

      It means that (in this particular case) both the junction of the two face sets you choose, AND the far edges which define the two hold lines, cannot be open edges. What's more, they must be knitted to the adjacent face, outboard of the hold line. It is laborious to provide the necessary "skirt" surfaces if the topology is even slightly interesting, and everything then has to be trimmed and knitted together. Surfaces which are simply a continuation of the face sets for filleting, (which will probably be available in the model) cannot be used, because when you knit to them, your hold lines will vanish....(This behaviour may be different in SW2010?)

      These skirts are automatically provided if you create a solid of rectangular cross section, per the graphic ... which shows very simple geometry.Naturally, the difficulties with surfaces compound considerably with more interesting shapes.

       

      Other tips: make sure you select 'Curvature continuity' if you want 2 hold lines. Otherwise there's no toggle to select to tell it you want two hold lines; it's just a matter of picking enough edges to define the two lines. In the solid case, this often requires only one edge from each side.

       

      Finally a plea to Solidworks: PLEASE include the terminology raised by error messages in Help.

      Also it would be great if there was some control over the spline profile for a 2-rail fillet, even if it was just an 'elliptical' option.

       

      Message was edited by: Andrew Troup  added clarification that edges must be knitted to avoid the "laminar edge" error, added rider re 2010 change

        • Re: Thinking of using Face Fillet w 2 hold lines for surface model?  My advice: Don't!
          Christopher Thompson

          I tried modeling this as surfaces in SW2009, and zebra strips confirm the curvature is continous across the double hold line of the face round. In the fillet options, "curvature continuous" needs to be selected. As you mentioned, the ruled surface skirt needed to be knitted to the main surface body. I could see where having this option available inside the ruled surface faeture could be useful, although I typically knit several surfaces together at once.

           

          Hold-Line_Surf-1_SW2009.JPG

           

          Hold-Line_Surf-2_zebra_SW2009.JPG

           

          Surface features are typically time consuming (even in Pro/Engineer) than solid features, but offer more modeling control. Sometimes the least time consuming option is to use the hydrid modeling (combination of solid & surface features) approach, and I do this mostly myself when it is possible.

          • Re: Thinking of using Face Fillet w 2 hold lines for surface model?  My advice: Don't!
            Christopher Thompson

            Here is a more interesting model using a sweep and multiple face fillets (double hold). This concept did not require a knit feature as in the previous example.

             

            Hold-Line_Surf-3_SW2009.JPG

             

            Hold-Line_Surf-4_deviation_SW2009.JPG

             

            One issue that concerns me, is why is there two faces along the fillet after the body is mirrored? Although it can be fixed by a delete face then fill (contact) feature, but I do not see why the extra steps are neccessary after mirroring the body.

            • Re: Thinking of using Face Fillet w 2 hold lines for surface model?  My advice: Don't!
              Matt Lombard

              Andrew, I'm having a little difficulty following exactly what you are saying. I agree that the functionality and the error messages are not represented well in the help.

               

              I just wanted to make one comment. Face fillet with 2 hold lines is really a special case. The software doesn't tell you this anywhere, you kind of have to figure it out on your own. If you use a single hold line with a fillet, it is like a variable radius fillet, where the radius size is determined by the distance of the hold line from the edge that is removed by the fillet. If you use 2 hold lines, this concept does not work. In fact, with 2 hold lines, you cannot have arc-based fillets at all unless both hold lines are the same distance from the filleted edge. The only way a fillet with 2 hold lines can work at all is if the fillet is spline based, and the only way to do that is to use the Continuous Curvature option.

               

              So the first thing that should be said in the Help about 2 hold line fillets is that you must use the CC option. In fact, if you choose the 2 hold line option, CC should also be chosen automatically by the software.

               

              Aside from that, I think "laminar edges" means that you have two open edges that lie on top of one another, as you would if you had a cube made from 6 un-knitted faces. Knitting the faces will solve that problem, and doing face fillets avoids it because you are picking faces rather than edges.

                • Re: Thinking of using Face Fillet w 2 hold lines for surface model?  My advice: Don't!
                  Andrew Troup

                  Matt

                   

                  Sorry about the poor explanation, in spite of which it seems you got the main thrust of what I was faltering to put across...

                   

                  Your point is well taken that if we follow the convention that a fillet has an arc-based section, most 'fillets' with 2 hold lines are no longer technically a fillet,and this is not well signalled by the interface design.

                   

                  This is particularly so in 2008 and previous versions, in which there is NO explicit "2 hold line" toggle, which I'm inferring from your words " if you choose the 2 hold line option" may have been added subsequently.


                  If not, then it seems to me that your proposal for curvature continuity to be invoked automatically might appear a bit mystifying and capricious, because it would flicker on and off as the user added and subtracted edges from the hold line selection set.

                   

                  To clarify my point about 'laminar edges', the software does not allow a user specifying a face fillet on a set of surfaces to pick hold lines which fit this description.

                  This description does not appear anywhere in Help, hence the error message is mystifying.

                  What's worse, it is not evident that they are talking about the hold lines.

                   

                  Furthermore, split lines create laminar edges, and hence (in the case of a surface model) cannot AFAIK be used as hold lines.

                   

                  Really the main thrust of my post was to share my realisation that it will generally be considerably quicker, and the software more cooperative, if you switch to solids in order to add a 2 hold line fillet to a surface model.