loft feature creating edges

[matheuswolffm]

Hey everyone!

So i'm making a Shell Eco-Marathon prototype-class monocoque shape, and lofting many transversal sections into a finished longitudinal profile, like a fuselage. Sketches used are 29, 45, 48, 51. I'm much more used to Inventor, kind of regretting the shift as of now.

I've done coincident points with longitudinal profiles to mold it, and most of the sections at the front are fine. However, there is a specific transition area where no matter the combination of sections, an edge shows up, and i'm not sure why.

I've lofted a couple bodies for demonstration (39, 37, 38, 40, 41), so if anyone could help me out i'd be very grateful. If you need me to clean the file further let me know. I think i've narrowed the problem to sketches 51 and 54.

[TrippyLighting]

---

@matheuswolffm wrote:

I'm much more used to Inventor, kind of regretting the shift as of now.

This mostly a matter of understanding how lofting works, not a matter of which CAD tool you use.

In the end both, Autodesk Inventor and Fusion use ASM (Autodesk Shape Manager) the same geometric modeling kernel.

kernel. 

The longitudinal splines have an enormous amount of fit points. Less is more in this case.
I would trace these curves manually with single-span control point splines.

[matheuswolffm]

Thanks for the reply!

I agree, but i'm not implying that the loft feature worked any different, just that the workflow change is making me question whether I'm using proper practices for fusion. 

On the subject matter, I've used a lot of points to make the control splines detailed and with lots of points for me to create transversal sections, but when making the actual bodies i'm not using more than 4 or 5. Are you sure this is a problem? 

And, even then, the loft works fine until the tail end of the model, which makes me think there is a particular issue with how I defined those specific sketches (51 and 54) and not the longitudinal profiles.

[TrippyLighting]

While inspecting your model a bit further I notices that thee are yellow icons in the timeline.

I ran a "Modify->compute all" and there are 24 warnings in you timeline.

I would fix all of those before proceeding.

[TrippyLighting]

Your profile sketches are under constrained, and under dimensioned.  Fr example, the center lines are not constrained to the sketch origin.

Mirroring fit-point splines is not a recommended workflow for Fusion specific reasons I am not going to explain here.

Patching together loft profiles from more than one curve is not a good workflow. 

All profiles should use the same number of fit points., or control points.
Use the simples curve you can get away with to reflect the shape. I prefer single-span control point splines where applicable. Single-span means that you have one more control point then the degree of the spline.

Use as few profiles and as few rails as you can get away with for smooth lofts.

The loft in your model is already pretty "wavy".

For lofting, more is less. For complex shapes more than one loft is needed. 

[matheuswolffm]

Thanks yet again. I'm working on the references and warnings as you suggested, but most don't interfere with the model. I used another parametric CAD last time I worked on this and more sections+guide lines worked well, but I guess I'll try using these guidelines. But to be sure, it should be possible for me to obtain a full profile with no edges, correct?

[TrippyLighting · Accepted solution]

---

@matheuswolffm wrote:

... But to be sure, it should be possible for me to obtain a full profile with no edges, correct?

---

I need to better understand what you are referring to ss "lines" .

This?
 

or this?

Those cannot be removed but are not problematic. NURBS surfaces in most mainstream CAD software are non cyclic. They have an and and a beginning and the "line" is the seam where the end and beginning join. The surface transition at that seam is usually curvature continuous.

[matheuswolffm]

I used the zebra curvature analysis to check some of the lofts i did, but sometimes they had aberrations. I figured that making the seam as "straight" as possible (by dragging the points to the topmost edge as per your first figure) gave me the best results, so I guess that is precisely what you mean. I'll try to keep that in mind! Apologies for the lack of formal knowledge.

[matheuswolffm]

One last question: I was having issues with the leading edge before, and I've read someplaces that suggest using a "Loft patch" to solve this type of loft that goes towards a single point, but apparently this function does not exist by this name anymore. What should I do other than creating a small circle on the very tip?