How do I shell this? - RC Airplane Wing

[Anonymous]

Ok, so I run into this problem a lot when designing aircraft for 3D printing.   When trying to shell, or create an offset surface I get this typical non descriptive error. 

Error: The operation could not create a valid result.
Try adjusting the values or changing the inputs.

Now, I've beat my head against a wall enough times on this kind of thing to have worked out ways to shell things by chopping off the extreme curved edges in order to get the shell/tool I need.  However, I'd like to know the proper way to go about fixing this.  Is my entire design approach wrong?  I'm a relatively new 3D CAD user, so I've been going with lofted sketch profiles to create solids.  Anyone got any tips on what I should be doing differently to be able to more reliably shell complex curved objects?  Most airplanes are almost entirely complex curves, so finding the right way would be awesome.

I've quickly drawn up an aircraft wing to demonstrate the problem.  It's curves are perhaps more complex than most, but I picked it as I figured it was a good choice to highlight the issues I always seem to be running into.

Here's the file:

https://a360.co/2QyPgfb

I've also posted a couple pictures of the wing, the original plane (Hall Bulldog), and a curvature map for reference.

Curvature MapCan't make a .8mm thick shellHall Bulldog

[davebYYPCU]

Problem is with the sharp and clean TE.

Why the shell?  (I know  - but when it won't / doesn't work, go another way)

Slice the solid and offset the edge.

Loft with the Patch Surface system, top and bottom separately, can offset or thicken those. (My workflow)

Another way is to loft the internal rib, and shell outside.

Many ways to skin a cat.  

(I am currently flying one of them.)

Might help....

[TrippyLighting]

The problem is really where the sharp training edge blends into a round corner. It looks smooth but in mathematical reality you get a degenerate NURBS surface. I took this into ZW3D and even there I was only able two shell this to ,35mm thickness.

@davebYYPCU advice is spot on. When a simple Shell won't work then deconstruct the part and to this surface by surface.

I'd have to say though that his is one darned nice loft. Well done!

[Anonymous]

Thanks for the replies guys.

The shell is simply to create a tool that I can then use to offset the airframes internal structure from the outer surface of the wing.  It's all got to do with making 3D printer slicer software output the lightest and strongest part possible.  Here's a couple pictures of one of my other designs so you can see what I'm working towards. You can see the added lightweight internal structure. 

In order to work well with FDM printers, all the internal bits need to be offset from the outer surface of the airframe.  So all I really need is an efficient way to make a cutting tool to make sure that the plane internals are offset the correct amount from the outer skin, which for most applications is .8mm.  I use shells, or more often, offset surfaces for this application.

Northern Pike RC AirplaneNorthern Pike Print Example

Using an offset face to slice up internal spars with the correct offset amount

So, no transitions from sharp points to curves.  I think I've tried that in the past with similar results, but I'll change stuff in my file and give it another go.  As for lofting patches and stitching stuff together, I can do it that way, but for whatever reason, stuff doesn't seem to come out as clean.  If that's going to give more consistent and better results though, I may need to practice with it.

[Anonymous]

So how sharp is sharp?  I made the trailing edges smooth curves instead of points.  But this seems to have just made things worse.  Where as before, I could at least still offset the top on bottom separately or shell the object if I sliced away the leading edge and wingtip, now I have to cut away the trailing edge as well to be able to get the offset I need.

Here's the rounded trailing edge.Here's the edges cut away to make an offset work.Loft ends in a point. Is this a bad method?

Maybe I need to do away with ending a loft at a point?  Is that a possibility for avoiding problems?  I'm going to mess around with piecing stuff together rather than doing one loft and see If I can come up with anything that works out smoother.

[marc.frank]

it is good design practice to have a defined trailing edge thickness

but even then it might not shell in f360