8 Replies Latest reply on Aug 6, 2015 11:21 AM by Mike Price

    Help with non-standard fillet features

    John Bodine

      Hello,

       

      I am currently working with a surface model that was giving me fits while trying to use the Solidworks fillet feature. It would not successfully propagate through all the faces that I needed a fillet between. So I changed my POV on the situation and came up with an idea that I thought was pretty slick to avoid the mess of the fillet feature. I trimmed the two faces back and then created a boundary surface that had the right contour between the two faces, effectively creating a constant length fillet. This seemed to work beautifully through my part. Now for the issues... I am having to offset the original part and thicken it (using thicken feature and other methods such as offset + lofting to form solid). Then I have to offset from the thickened face. This is where the problems start to show up.

       

      For some reason, my "fillet" splits into two separate faces after enough offsetting/thickening. For instance, there is a forward and an aft section to the fillet (roughly parallel to the faces needing the fillet). This is fine for the thickened feature, but it is causing other issues down the road. What are some ways that I can make my "fillet" more robust? After an extreme enough distance, the fillet actually breaks down an leaves a gap between the original two surfaces. Is there anything I can do to improve this without creating more surfaces to fill the gaps?

       

      Thanks,

      John

        • Re: Help with non-standard fillet features
          Mike Price

          Hi John,

           

          Can you show a screen shot?  What is the curvature of your boundary surface and to which value are your trying to offset?

           

          I would use this method as a last resort and try to get the fillet to take.

           

          Were you able to fillet any edge and if so, would SW let you offset that fillet?

           

          Are you seeing the split of the boundary surface in the preview only or actually when you create the boundary?  Sometimes it will preview as split, but will actually create the unbroken offset.

           

          Finally, if I were on a time crunch, you could just recreate your boundary by offsetting your surfaces from which you created your original boundary.

            • Re: Help with non-standard fillet features
              John Bodine

              Hi Mike,

               

              Thanks for the quick reply! The curvature of my boundary surface is huge compared to the offset that I am trying to produce (100:1 in most areas).

               

              I was able to create a fillet on the opposite side of the model which has similar but not identical geometry. When I offset an actual fillet feature, it's beautiful on the offset surface. I guess one of the hidden benefits of the fillet feature is that is can change its definition based on context.

               

              I did not realize that the preview would have that effect. I will double check to see if this works. Thanks for the tip.

               

              Adding another patched surface between the two original surfaces after the offset is my last resort, mostly because more parts have to fit to this piece, hence more offsets and more patching.

               

              Here are some screen caps from the model. First up, you can see the fake fillet between the two surfaces (the thin strip running from top left), not an especially informative shot, but gives a bit of context.

               

              ScreenGrab3.PNG

               

              Second, here is the curvature at the "fillet." Something I had not noticed before was the not-so-nice change in curvature at the forward edge of the boundary surface (the red line). I don't think that I have seen this before because the image quality setting keeps resetting, even with the macro that I run after most feature creations. Love that bug.

               

              HighImageQuality_Curvature1.PNG

               

              Here is the same shot, with the curvature I usually see. Notice that with the normal image settings, this looks really close to normal fillet curvature.

               

              PostMacroCurvature.PNG

               

               

              Lastly, this is the underside of the thickened surface, highlighted to show the multiple faces generated after the offsets (left-most side of boundary). After taking a new look at this model this morning, I think that the change in curvature discussed above may be the cause of this.

               

              UndersideView_SplitFillet.PNG

               

               

              I realize this is a large wall of text and info, thank you for your time.

               

              Best,

              John

                • Re: Help with non-standard fillet features
                  John Bodine

                  I guess I should also note that I am more than willing to abandon this funky fillet method if there is a better way. Time is a concern, but fixing this now will save me a lot of time down the road.

                    • Re: Help with non-standard fillet features
                      Mike Price

                      So were you able to get the boundary to offset without splitting?

                       

                      Regarding getting an actual fillet to take, you'd have to try many different things (tangent propagation on/off, face filleting, manual selection of edges,etc) or you could post your file and usually somebody can get a fillet to take.  If you post a 2015, I can't help you there, but could with a parasolid.....

                        • Re: Help with non-standard fillet features
                          John Bodine

                          Hi Mike,

                           

                          I was not able to get the surface to offset by itself. I went back to the original surface, tidied some things up, and messed with the fillet feature for a while, and eventually got the surface to take an actual fillet. The variable radius fillet with identical radii seemed to work for a lot of the model. Thanks for your help!

                           

                          -John

                            • Re: Help with non-standard fillet features
                              Mike Price

                              Hi John,

                               

                              Glad you were able to work it out.  Filleting is one of those things that sometimes you don't know why some take and others do not.  Seems to me that a variable radius with identical radii is a constant radii fillet, but what do I know...

                               

                              There was a recent post concerning fillets.  Your method of creating your own was something I saw years ago on the forum, but couldn't find the post.....However, I still had the part.  You might want to take a look at how they did it and compare notes.

                               

                              I am still wondering why your boundary wouldn't offset.  On a side, I have had instability issues with boundary surfaces that I don't get with lofted surfaces.  Therefore, I always try a loft first and if that won't go, I then try a boundary.  Of course, boundary surfaces are needed when using only one direction 1 curve(s) and are also nice in that you can slide connectors.  You could try a loft and see if it would offset.

                               

                              Why wont this part fillet

                                • Re: Help with non-standard fillet features
                                  John Bodine

                                  Regarding the variable radius fillet, I read somewhere that the constant radius fillet and the VR fillet use different equation sets to define them. Sometimes the VR fillet works where the constant fillet fails. I'm in the same boat as you, I'm not sure about the distinction between VR fillet with identical radii and a constant radius fillet.

                                   

                                  I would be very interested in taking a look at that part. I think there is a viable non-standard fillet operation using different surface features for eccentric geometry where a normal fillet just won't take. I assume the part is the lock file part in that thread?

                                   

                                  I'm not entirely sure why the surface would not offset either. It was having issues in sections where the surface was very long compared to its width (0.2" wide x 50+" long). The shorter sections with relatively large curvature changes along the length of the fillet worked very well. Maybe that gives a bit of insight. My thinking is that the constraints and feature properties that I was using forced the part to take on unintended geometry along the length of the fillet. I checked the old model last night and found a couple of sections that had abnormally low radius of curvature. That seems to be the most likely cause of the issue.

                                   

                                  -John