Fill surface tangency

gauthier_östervall

I am trying to understand how to deal with continuity of fills. Here is the document.

Here is my general shape, see the guides used to make the fill:




The problem is that between the guides, the fill surface is not tangent to the bottom cylinder:



I've tried to use this plane to define tangency (since this should make the fill use the plane normal as tangent, I thought?):




And for good measure, I tried with the cylinder:


None of this worked. How do I tell the fill to start continuous to the cylinder?

S1mon

[edit: I missed your document link]

If you can share a public document I’m sure we can help. In general if you’re not picking the open edge of a surface you want it to be tangent to, you will need to explicitly tell the fill command the reference surface for that sketch or face. 

martin_kopplow

I don't see the need to make the fill tangent to the horizontal plane, when we have the cylinder. It can be forced to be tangent, after splitting the cylinder into dedicated tangent regions to refer to, though the resulting surface isn't nessecarily beautiful. The takeoff curvature of the conics could be too much different and force a bulge to form.

Here's my copy of the document

Just a quick try. There are possibly more elagant ways to do it.

Edit: Playing with the rho values and adding one more guide improves things quite some:

gauthier_östervall

@martin_kopplow Thanks a lot, very helpful! Was I expecting too much, when I thought the tangency with the cylinder as a whole should work?

I didn't quite understand what you meant about bulges, but saw them for myself trying to reproduce what you did. Maybe it would help if the guides weren't on parallel planes, but in planes going through the middle point? Or if the guides weren't conics but pairs of conics?

I really didn't think this project would be this complicated, am I thinking this the wrong way? This is the terminating shoe of a gutter.

martin_kopplow

@gauthier_östervall Maybe sometimes we just have to think much simpler. What about this approach, with splines rather than conics, and a cutaway? 

At least it gets all the tangency right. Of course, I can't know what the actual requirements for this job actually are.

gauthier_östervall

@martin_kopplow Oh, that's awesome, thanks! I'm very new at conics, my understanding was that they would make further operations lower down in the feature list more stable. As in easier to project, for example.

For this very part, I will go for a solution similar to what you provided here. But I still want to learn conics and tangency of fills

martin_kopplow

@gauthier_östervall Well, conics provide a nice approach for controlling tangency at their take-off and landing point and clean math, but not much control in-beween, and stacking them may lead to a difficult curvature situation in the middle, especially when stacked and single conics are combined into a single surface. Splines, on the other hand, are more flexible, but need some attention so they don't get out of shape when editing model history. We need to decide what to use best in any given setup.

If it is an easyly configurable model you're after, you would e.g. want a setup with very few features. A simple loft could do the trick, then. It depends ... ;0)

No matter what, it is always important to avoid too many tangent edges meeting at any one point.