Loft strategy - straight or curved guide?

Hi all,
Every now and then I continue trying to make this hopelessly difficult fender geometry. But now I wonder what strategy is best as to avoid future problems with the geometry pictured below. I have two alternatives:

  1. Extending the loft curves only to the curved boundary surface, as to follow the wheel opening (which has an arc with greater radius than this surface, and so it will be cut out later)

  2. Align to the top surface if the fender, extending the loft curves to a flat boundary
    surface, as pictured.

The isocurves of the final lofted surface are affected accordingly. So again, which strategy is preferred? Pros and cons with each?

// Rolf

Hi Ril - if you are asking about the pink surface in the foreground, I’d make that from a clean new curve, the one you show looks like it is generated from a trimmed edge or something - kind of messy. But I am not sure I understand the actual question…


Yes, they are messy, but the pink surfaces (one arc and one flat) are just “cutting srf’s”, and, boundaries for extending the loft curves.

In any case, I decided to go for the arc, since the curvature of the isocurves stay smother.


But now I got another problem: The surface isn’t really smooth, using Loft I get small small wrinkles on the surfaces in some places using Loft. (Fig.1, 2)

But if I ExtractWireframe and NetworkSrf on top of those wireframes, then the smoothness is much better. (Fig 3.) Why is it so?

I can’t reduce the number of Control Points (in the Loft dialog) too much, since then the surface won’t follow the curves close enough.

// Rolf

Fig 1.

Fig 2.

Fig. 3.

I’m not sure what all the inputs are, but I think I’d approach this as a ‘soft box’ and trim the wheel arch into it rather than try to build the surface to the wheel arch.

@RIL - something like this - mine is not very nice but just to indicate the layout.

FenderLayout.3dm (190.1 KB)


As you can see from Fig 2, the loft curves are extended beyond the wheel arc, to a smaller diameter (down to the pink boundary surface in the first post).

The final wheel arc pictured in Fig 1,2 was cut out after the surface was in place (except for Fig 3 which was done with NetworkSrf).

The only input curves used for Lofting were the “selected” ones in this picture:

Edit: All sides, except for towards the CPlane in the middle, are lofted far outside the final edges, which are cut out afterwards (using curved surfaces)

// Rolf

Yes, @pascal, I’ve been working a lot with a similar cage, but the geometry has so many more curvatures than only the “cube sides”, both in the front and on top, and at the lower edges (pictured below). For this reason I made a complex surface “manually”, although ending up having glitches between the complex patches. But I used that surface to make “intersection” curves to be used with Loft as to make the final surface in “one go” instead, in order to get rid of the glitches.

So, the loft curves you see in the last picture are very complex, but the spot where the surface gets wrinkled isn’t very complex at all. < Scratching head >

I’m using Loft / Normal / Refine 12 CPs, which follows the curves the best, but wrinkles show up here and there in a way I find puzzling.

// Rolf

Edit: Fig. Surface has some complex curvatures :