So, I just learned something new.

First of all, both the $fn and slices parameters will do nothing (at least nothing visible) unless you’re also using the twist parameter. When using twist, slices determines the number of vertical steps using intermediate rotations between the top and bottom. What $fn does is allow you to subdivide the edges of the extruded shape to produce faces with smooth compound curvature between those steps.

After a little experimenting, it looks like the algorithm is that for each closed curve in the 2D shape, the straight edges will be subdivided until the whole closed curve has at least $fn segments. Longer edges are given priority, such that the resulting segments will be as close to equal length as possible.
On Jul 5, 2022, 14:03 -0700, Julien Jacquet jacquetjulien@gmail.com, wrote:

Minor correction: the scale parameter, if set to a vector, can also produce cases where slices and $fn have a visible effect, even if twist is not used.
On Jul 5, 2022, 14:54 -0700, Whosawhatsis whosawhatsis@gmail.com, wrote:

On 7/5/22 17:10, Julien Jacquet wrote:

nope

nope

OpenSCAD uses straight lines to draw a circle. $fn is how many
facets it takes to draw a full circle.

$fn=3 will draw a triangle, 4 will draw a square. 6 will draw a hex nut.
72 up to 720 will draw increasingly smoother circles. But there's point
of diminishing returns. 36000 will work, but will also take a long time
to render and eat memory like M&M's candies.

To borrow a phrase from an ex bro-in-law, use enough to make it work,
 "good enough for the girls you go with" ;o)>

Take care & stay well.

Cheers, Gene Heskett.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/

Hello to all,
Thank you for your answers.
So yes I forgot to specify that it was during a twist...

Gene, well understood thanks for a polygon, but it's less clear to me when
twisting.....

Whosawhatsis, the effect of $fn is to "subdivide the edges of the extruded
shape to produce faces with smooth compound curvature between those
steps.", right?

Thanks :)

Le mer. 6 juil. 2022 à 02:25, gene heskett gheskett@shentel.net a écrit :

On 7/6/22 04:47, Julien Jacquet wrote:

Well, I probably stepped outside the bounds of my knowledge a bit with
my answer.

One of the areas where I feel there isn't a real world solution is when
I'm laying out
a screw thread, in this case for a buttress bolt, I am limited by the
available sizes
of cutting tools as to how accurately I can reproduce a 6mm buttress
thread in a 2
start configuration, getting a 12mm pitch by setting the tools yz point,
cutting to
the other end of the thread, rotating b an additional 360 degrees at the
stop point,
then rotate back 180 degrees then cut coming back to make the second start.

This is not the exact profile for a buttress thread. So then to make the
nuts, which
because of the 2 start, are identical, trimming this polygon to match
what I just
carved in hard maple. By using the polygon I can match the two tooth shapes
pretty exactly. However to achieve the actual angle of the thread so
there's no
sawtooth to the surfaces, the whole extruded polygon is tilted for best
surface
matching.

It seems to me that having the ability to warp just one axis, leaving
the polygon
vertical but the X axis warped from perpendicular by this ramp angle
would make
an even better profile. But I've not figured out how to do that in OpenSCAD.
Might that be done with the linear_extrude() that makes that 2d polygon
into 3d,
1 mm wide?

Or have I not adequately explained it?

Maybe but an example doesn't seem to be in the docs.

Cheers, Gene Heskett.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/

Look at
https://en.m.wikibooks.org/wiki/OpenSCAD_User_Manual/Transformations#multmatrix
One of the examples  is "skew" or "tilting" a shape. This may match your
requirement

Michael, fra mobilen

ons. 6. jul. 2022 18.28 skrev gene heskett gheskett@shentel.net:

On 7/6/22 14:14, Michael Möller wrote:

One of the links from that might work, here is what I have, at:

http://geneslinuxbox.net:6309/6040-stf/vise-screw-pix/tooth-sideface.png

to show the side profile of a tooth, and:

http://geneslinuxbox.net:6309/6040-stf/vise-screw-pix/outer-tooth-face.png

to show the tilt, and what I want to do is skew only the z back to vertical,
leaving the horizontal tilts you can see

Or skew the original w/o tilting z from vertical since that may be
simpler and
faster.

That was generated by this module:

module threadshape() //version 5
{
    translate([-.008,.275,0])rotate([94.5,0,0])
    linear_extrude(height = .48, center = true, convexity = 10, twist =
0, slices = 1, scale=1.0 )
polygon(points=[[0,0],[0,1.3],[3.8,4.8],[4,5.15],[3.75,5.5],[0,6]]);
    //added a point to round  the tip to better match a BallNose tool
}  // polygon(points[[x,z],[x,z],...]]);

That is wrapped around a cylinder to form the screw, and is used in
a .5mm diameter oversize version to be arg 2 of a difference() to make
the nut.  The .5mm is nozzle size comp in radius with an extra .1 for
clearances subtracted if making the screw, added if making the nut.
The fit is quite precise considering the printers rather cavalier
treatment of nozzle comp. It doesn't have any.  So holes are always .4mm
undersized, and circle outer's are .4mm bigger than code. But you guys
already know that so I'm carrying coals a Newcastle. ;o)>

Thanks all.
Take care and stay well, everybody.

Cheers, Gene Heskett.

"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.

• Louis D. Brandeis
  Genes Web page http://geneslinuxbox.net:6309/