Hello,

The current project I am in is an all modular rc plane design. So far, the
complete wing has been openscaded with the exception of the wingtip, which
was planned to be curved... and that's why it ended up being more
complicated than anticipated. Below is what the wing looks like right now
(still working on the servo mount and associated mechanism) :
http://forum.openscad.org/file/t1378/perspective_aile_en_travail.perspective_aile_en_travail

The challenge is in the curved winglet design, where the path it follows is
a quadratic function. To do so, I take the 2D section of the wing (airfoil)
and I put linear fragments (regular linear extrusions) to link the points
along the quadratic path. However, I realized that junctions between linear
fragments are not perfect at all when using a regular custom rotate_extrude
of wing cross-section with an angle specified, because the fragments
parameter ($fn) for an extrusion along axis only apply to a 360 degrees
angle, and not locally to the specified portion of extrusion with an angle.

Here are some pics to help visualize.
http://forum.openscad.org/file/t1378/vue_du_dessus_saumon_ModulAir.vue_du_dessus_saumon_modulair
http://forum.openscad.org/file/t1378/vue_de_surplomb_intersection.vue_de_surplomb_intersection
http://forum.openscad.org/file/t1378/intersection.intersection

I am afraid these bad junctions could make bad STL... and then low quality
G-CODE. Is it worth it? I was planning on getting onto a more advance
polyhedron manipulation, but I can't go anywhere, because I found no actual
way to get the coordinates from a given 2d shape. If it can help, the shape
I want to extract the coordinates from is an airfoil with a kind of inner
wall (inside its perimeter) and I use minkowski to do so. Just like in the
picture below :
http://forum.openscad.org/file/t1378/inner_contour.inner_contour

I was wondering if someone had a solution for this challenge, wether it is a
rotate_extrude with $fn controlling the number of fragements on the
extrusion angle (not the 360), a way to extract coordinates from a custom 2d
shape, or anything else.

Regards,

Arthur Légaré from Québec.

--
Sent from: http://forum.openscad.org/

Hi Arthur,

That's quite an ambitious project!

I don't know much about OpenSCAD yet, but If you have the contour, can't you
export a DXF, and read coordinates from it?

A thought, if you import into a CAD/CAM program, then you could even export
gcode, which would be a list of the XY coordinates you want.  (I come from
CNC).

Hope this helps,

Alan

--
Sent from: http://forum.openscad.org/

If I understand correctly I think you can hull() two adjacent slices, that
will cleanup any holes etc.

Admin - PM me if you need anything, or if I've done something stupid...

Unless specifically shown otherwise above, my contribution is in the Public Domain; to the extent possible under law, I have waived all copyright and related or neighbouring rights to this work. Obviously inclusion of works of previous authors is not included in the above.

The TPP is no simple “trade agreement.”  Fight it! http://www.ourfairdeal.org/  time is running out!

Sent from: http://forum.openscad.org/

I'm not clear on what you want to accomplish, but you should be able to get a
decent approximation of your quadratic function using circular sections.

--
Sent from: http://forum.openscad.org/

The proper way to do this is to utilize a parametrized airfoil generator
function that returns you the airfoil data in list representation (i.e. a
representation you can visualize with polygon() ).

In the next step you utilize a so called trajectory function that calls the
generator function with changing parameters and produces a series of
airfoils that defines your wing just - like a skeleton. Before it stuffs
these data into a list, it properly places each airfoil into 3D. To do this
you use affine functions that are the counterparts of translate() and
rotate() but operate over a series of 3D points.

The final step is to skin or sweep the skeleton into a polyhedron.
Have a look at: https://www.thingiverse.com/thing:900137 to see the
principle.

--
Sent from: http://forum.openscad.org/