Problems with difference

> I'm helping a friend's daughter to construct a 120 sided solid by
> 3d printing a skeleton that she's going to cover in mylar.
>
> My attempt is at:
> https://github.com/rugyoga/catalan_solids/blob/main/disdyakis_triacontahedron.scad
>
> Each face is a triangle.
> To save some time, I wanted to print the parallelogram formed by four
> triangles
> so I only need 30 of them. Ideally I wanted them to snap together like lego
> bricks.
> So the edges that mate to the neighbouring parallelogram should be half
> cylinders.
> The code for the parallelogram works great.
> But I can't get the difference operator to shave the half cylinders.
> I'm subtracting four tetrahedron where face slices the rod in half.
>
> c0 = 3 * (15 + sqrt(5)) / 44;
> c1 = (5 - sqrt(5)) / 2;
> c2 = 3 * (5 + 4 * sqrt(5)) / 22;
> c3 = 3 * (5 + sqrt(5)) / 10;
> c4 = sqrt(5);
> c5 = (75 + 27 * sqrt(5)) / 44;
> c6 = (15 + 9 * sqrt(5)) / 10;
> c7 = (5 + sqrt(5)) / 2;
> c8 = 3 * (5 + 4 * sqrt(5)) / 11;
>
> function vertices() =
>    [[0.0, 0.0,  c8], // 0
>     [0.0, 0.0, -c8],
>     [ c8, 0.0, 0.0],
>     [-c8, 0.0, 0.0],
>     [0.0,  c8, 0.0],
>     [0.0, -c8, 0.0],
>     [0.0,  c1,  c7],
>     [0.0,  c1, -c7],
>     [0.0, -c1,  c7],
>     [0.0, -c1, -c7],
>     [ c7, 0.0,  c1], // 10
>     [ c7, 0.0, -c1],
>     [-c7, 0.0,  c1],
>     [-c7, 0.0, -c1],
>     [ c1,  c7, 0.0],
>     [ c1, -c7, 0.0],
>     [-c1,  c7, 0.0],
>     [-c1, -c7, 0.0],
>     [ c3, 0.0,  c6],
>     [ c3, 0.0, -c6],
>     [-c3, 0.0,  c6], // 20
>     [-c3, 0.0, -c6],
>     [ c6,  c3, 0.0],
>     [ c6, -c3, 0.0],
>     [-c6,  c3, 0.0],
>     [-c6, -c3, 0.0],
>     [0.0,  c6,  c3],
>     [0.0,  c6, -c3],
>     [0.0, -c6,  c3],
>     [0.0, -c6, -c3],
>     [ c0,  c2,  c5], // 30
>     [ c0,  c2, -c5],
>     [ c0, -c2,  c5],
>     [ c0, -c2, -c5],
>     [-c0,  c2,  c5],
>     [-c0,  c2, -c5],
>     [-c0, -c2,  c5],
>     [-c0, -c2, -c5],
>     [ c5,  c0,  c2],
>     [ c5,  c0, -c2],
>     [ c5, -c0,  c2], // 40
>     [ c5, -c0, -c2],
>     [-c5,  c0,  c2],
>     [-c5,  c0, -c2],
>     [-c5, -c0,  c2],
>     [-c5, -c0, -c2],
>     [ c2,  c5,  c0],
>     [ c2,  c5, -c0],
>     [ c2, -c5,  c0],
>     [ c2, -c5, -c0],
>     [-c2,  c5,  c0], // 50
>     [-c2,  c5, -c0],
>     [-c2, -c5,  c0],
>     [-c2, -c5, -c0],
>     [ c4,  c4,  c4],
>     [ c4,  c4, -c4],
>     [ c4, -c4,  c4],
>     [ c4, -c4, -c4],
>     [-c4,  c4,  c4],
>     [-c4,  c4, -c4],
>     [-c4, -c4,  c4], // 60
>     [-c4, -c4, -c4]];
>  
> v = vertices()
>
> // cribbed from
> http://forum.openscad.org/Rods-between-3D-points-tp13104p13115.html
>
> module rod(p1,p2,tk){ // draw ray between 2 specified points
>    translate(p1)
>       sphere(r=tk);
>    translate(p2)
>      sphere(r=tk);
>    translate((p1+p2)/2)
>      rotate([-acos((p2[2]-p1[2]) / norm(p1-p2)),0,
>              -atan2(p2[0]-p1[0],p2[1]-p1[1])])
>         cylinder(r1=tk, h=norm(p1-p2), center = true);
> }
>
> module diamond() {
>     rod(v[0], v[6], 1);
>     rod(v[0], v[18], 1);
>     rod(v[0], v[8], 1);
>     rod(v[0], v[20], 1);
>     rod(v[20], v[6], 1);
>     rod(v[6], v[18], 1);
>     rod(v[18], v[8], 1);
>     rod(v[8], v[20], 1);
> }
>
> module tetra(points) {
>      polyhedron( points, [[0, 1, 2], [0, 1, 3], [0, 2, 3], [1, 2, 3]] );
> }
>
> module flattened_diamond() {
>      difference() {
>          diamond();
>          tetra( [v[8], v[18], v[7], v[15]] * 2);
>          tetra( [v[6], v[18], v[9], v[14]] * 2);
>          tetra( [v[8], v[20], v[7], v[15]] * 2);
>          tetra( [v[20], v[6], v[19], v[1]] * 2);
>      }
> }
>
> Can you see anything obvious that I'm doing wrong?
> Cheers,
> Guy
>
>
>
> --
> Sent from: http://forum.openscad.org/
>
> _______________________________________________
> OpenSCAD mailing list
> [hidden email]
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

> Does she know how fragile mylar is???
>
>
> On 12/18/20 11:30 PM, rugyoga wrote:
> > I'm helping a friend's daughter to construct a 120 sided solid by
> > 3d printing a skeleton that she's going to cover in mylar.
> >
> > My attempt is at:
> > https://github.com/rugyoga/catalan_solids/blob/main/disdyakis_triacontahedron.scad
> >
> > Each face is a triangle.
> > To save some time, I wanted to print the parallelogram formed by four
> > triangles
> > so I only need 30 of them. Ideally I wanted them to snap together like lego
> > bricks.
> > So the edges that mate to the neighbouring parallelogram should be half
> > cylinders.
> > The code for the parallelogram works great.
> > But I can't get the difference operator to shave the half cylinders.
> > I'm subtracting four tetrahedron where face slices the rod in half.
> >
> > c0 = 3 * (15 + sqrt(5)) / 44;
> > c1 = (5 - sqrt(5)) / 2;
> > c2 = 3 * (5 + 4 * sqrt(5)) / 22;
> > c3 = 3 * (5 + sqrt(5)) / 10;
> > c4 = sqrt(5);
> > c5 = (75 + 27 * sqrt(5)) / 44;
> > c6 = (15 + 9 * sqrt(5)) / 10;
> > c7 = (5 + sqrt(5)) / 2;
> > c8 = 3 * (5 + 4 * sqrt(5)) / 11;
> >
> > function vertices() =
> >    [[0.0, 0.0,  c8], // 0
> >     [0.0, 0.0, -c8],
> >     [ c8, 0.0, 0.0],
> >     [-c8, 0.0, 0.0],
> >     [0.0,  c8, 0.0],
> >     [0.0, -c8, 0.0],
> >     [0.0,  c1,  c7],
> >     [0.0,  c1, -c7],
> >     [0.0, -c1,  c7],
> >     [0.0, -c1, -c7],
> >     [ c7, 0.0,  c1], // 10
> >     [ c7, 0.0, -c1],
> >     [-c7, 0.0,  c1],
> >     [-c7, 0.0, -c1],
> >     [ c1,  c7, 0.0],
> >     [ c1, -c7, 0.0],
> >     [-c1,  c7, 0.0],
> >     [-c1, -c7, 0.0],
> >     [ c3, 0.0,  c6],
> >     [ c3, 0.0, -c6],
> >     [-c3, 0.0,  c6], // 20
> >     [-c3, 0.0, -c6],
> >     [ c6,  c3, 0.0],
> >     [ c6, -c3, 0.0],
> >     [-c6,  c3, 0.0],
> >     [-c6, -c3, 0.0],
> >     [0.0,  c6,  c3],
> >     [0.0,  c6, -c3],
> >     [0.0, -c6,  c3],
> >     [0.0, -c6, -c3],
> >     [ c0,  c2,  c5], // 30
> >     [ c0,  c2, -c5],
> >     [ c0, -c2,  c5],
> >     [ c0, -c2, -c5],
> >     [-c0,  c2,  c5],
> >     [-c0,  c2, -c5],
> >     [-c0, -c2,  c5],
> >     [-c0, -c2, -c5],
> >     [ c5,  c0,  c2],
> >     [ c5,  c0, -c2],
> >     [ c5, -c0,  c2], // 40
> >     [ c5, -c0, -c2],
> >     [-c5,  c0,  c2],
> >     [-c5,  c0, -c2],
> >     [-c5, -c0,  c2],
> >     [-c5, -c0, -c2],
> >     [ c2,  c5,  c0],
> >     [ c2,  c5, -c0],
> >     [ c2, -c5,  c0],
> >     [ c2, -c5, -c0],
> >     [-c2,  c5,  c0], // 50
> >     [-c2,  c5, -c0],
> >     [-c2, -c5,  c0],
> >     [-c2, -c5, -c0],
> >     [ c4,  c4,  c4],
> >     [ c4,  c4, -c4],
> >     [ c4, -c4,  c4],
> >     [ c4, -c4, -c4],
> >     [-c4,  c4,  c4],
> >     [-c4,  c4, -c4],
> >     [-c4, -c4,  c4], // 60
> >     [-c4, -c4, -c4]];
> >  
> > v = vertices()
> >
> > // cribbed from
> > http://forum.openscad.org/Rods-between-3D-points-tp13104p13115.html
> >
> > module rod(p1,p2,tk){ // draw ray between 2 specified points
> >    translate(p1)
> >       sphere(r=tk);
> >    translate(p2)
> >      sphere(r=tk);
> >    translate((p1+p2)/2)
> >      rotate([-acos((p2[2]-p1[2]) / norm(p1-p2)),0,
> >              -atan2(p2[0]-p1[0],p2[1]-p1[1])])
> >         cylinder(r1=tk, h=norm(p1-p2), center = true);
> > }
> >
> > module diamond() {
> >     rod(v[0], v[6], 1);
> >     rod(v[0], v[18], 1);
> >     rod(v[0], v[8], 1);
> >     rod(v[0], v[20], 1);
> >     rod(v[20], v[6], 1);
> >     rod(v[6], v[18], 1);
> >     rod(v[18], v[8], 1);
> >     rod(v[8], v[20], 1);
> > }
> >
> > module tetra(points) {
> >      polyhedron( points, [[0, 1, 2], [0, 1, 3], [0, 2, 3], [1, 2, 3]] );
> > }
> >
> > module flattened_diamond() {
> >      difference() {
> >          diamond();
> >          tetra( [v[8], v[18], v[7], v[15]] * 2);
> >          tetra( [v[6], v[18], v[9], v[14]] * 2);
> >          tetra( [v[8], v[20], v[7], v[15]] * 2);
> >          tetra( [v[20], v[6], v[19], v[1]] * 2);
> >      }
> > }
> >
> > Can you see anything obvious that I'm doing wrong?
> > Cheers,
> > Guy
> >
> >
> >
> > --
> > Sent from: http://forum.openscad.org/
> >
> > _______________________________________________
> > OpenSCAD mailing list
> > [hidden email]
> > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>
> _______________________________________________
> OpenSCAD mailing list
> [hidden email]
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>

> You might find this approach to making polyhedra interesting:
>
> https://www.myminifactory.com/object/3d-print-folding-pentagonal-hexacontahedron-puzzle-85216
>
> I made the tetrahedron and it worked very nicely.  I'm not sure I fully
> understand your plan, but if I were trying to assemble 30 of those "diamond"
> modules I doubt they would align and close up once I was done.  If you
> haven't thought about it you might want to introduce more registration
> geometry (e.g. alignment groove down the edge) so that the pieces hopefully
> line up properly.  I'd also be worried about the labor of removing support
> material and the very small surface area for bed adhesion.
>
> Another observation:  I tried to render the full_skeleton model from the
> above example.  It's a pretty obnoxious model because every edge is in there
> twice and there are a bunch of repeated spheres at each vertex.  Anyhow, it
> crashed opencad RC3 to try to render it.
>
> I made a generic wireframe module that takes out repeated edges and vertices
> and was able to render this beast, though it did take 24 minutes.
>
> If you're interested in polyhedra you might also want to take a look at
> this:
>
> https://github.com/revarbat/BOSL2/wiki/polyhedra.scad
>
> It will give you vertices and faces for all the catalan solids (and others).
>
>
>
>
> --
> Sent from: http://forum.openscad.org/
>
> _______________________________________________
> OpenSCAD mailing list
> [hidden email]
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

> > wrote:
>
>> Does hull recalculate the edges and smooth out the math?
>>
>> David
>>
>>
>> On 12/19/20 9:34 AM, adrianv wrote:
>> > You might find this approach to making polyhedra interesting:
>> >
>> >
>> https://www.myminifactory.com/object/3d-print-folding-pentagonal-hexacontahedron-puzzle-85216
>> >
>> > I made the tetrahedron and it worked very nicely.  I'm not sure I fully
>> > understand your plan, but if I were trying to assemble 30 of those
>> "diamond"
>> > modules I doubt they would align and close up once I was done.  If you
>> > haven't thought about it you might want to introduce more registration
>> > geometry (e.g. alignment groove down the edge) so that the pieces
>> hopefully
>> > line up properly.  I'd also be worried about the labor of removing
>> support
>> > material and the very small surface area for bed adhesion.
>> >
>> > Another observation:  I tried to render the full_skeleton model from
>> the
>> > above example.  It's a pretty obnoxious model because every edge is in
>> there
>> > twice and there are a bunch of repeated spheres at each vertex.
>> Anyhow,
>> it
>> > crashed opencad RC3 to try to render it.
>> >
>> > I made a generic wireframe module that takes out repeated edges and
>> vertices
>> > and was able to render this beast, though it did take 24 minutes.
>> >
>> > If you're interested in polyhedra you might also want to take a look at
>> > this:
>> >
>> > https://github.com/revarbat/BOSL2/wiki/polyhedra.scad
>> >
>> > It will give you vertices and faces for all the catalan solids (and
>> others).
>> >
>> >
>> >
>> >
>> > --
>> > Sent from: http://forum.openscad.org/
>> >
>> > _______________________________________________
>> > OpenSCAD mailing list
>> >

> adrianv:
>
> Cool finds!
> But the folding polyhedra doesn't work for our purpose because
> the net we would need would be too large for my 3d printer.
>
> That's also why we rejected printing the skeleton.
> (Also very wasteful of filament).
>
> Yup, the skeleton has a known bug.
> Hence the comment:
>
> // avoid this - it duplicates rods
> module skeleton(items) {
>
> Your comment about the registration geometry is spot on.
> My next step was to code pips - spheres evenly distributed along the
> semi-rods.
>
> The diamond solution was our favourite because:
> - low filament waste
> - easily scaled
> - only needed 30 of them
>
>
>
>
>
> --
> Sent from: http://forum.openscad.org/
>
> _______________________________________________
> OpenSCAD mailing list

> adrianv wrote
>> The hull function takes a list of points
>> and computes the face list of the convex hull of the point list, which
>> can
>> then be passed to polyhedron()
>
> Huh. So... this should work:
>
> module badCube(){
>   polyhedron(points=[[-1,-1,-1],[1,1,-1],[-1,1,-1],[1,-1,-1],
>                      [-1,-1, 1],[1,1, 1],[-1,1, 1],[1,-1, 1], [0,0,1.41]],
>              faces=[[0,1,2,3,4,5,6,7,8]]);
> }
>
> color("pink") translate([0,-3,0]) badCube();
> color("lightblue") hull(){
>   badCube();
> }
>
> <http://forum.openscad.org/file/t452/badCube.png> 
>
> I was going to request the 'hull' function be slightly modified to take as
> optional input a point cloud, but given that it seems to work with a
> polygon
> object with a 'faces' argument that only needs to define the used points
> [in
> any arbitrary order], creating a module that does that is fairly easy,
> e.g.
>
> module hullPointCloud(points) {
>   hull()
>    {
>      echo([ for(i=[0:(len(points)-1)]) i]);
>      polyhedron(points=points, faces=[[ for(i=[0:(len(points)-1)]) i]]);
>    }
> }
>
>
>
> --
> Sent from: http://forum.openscad.org/
>
> _______________________________________________
> OpenSCAD mailing list