ctchin wrote

Parkinbot wrote

I was playing with these two approaches and tried to make a generalized tree
branching with fillets between.

Its working well for direct code but I wanted to use a list and a for loop.
Alas the for loop does an implied union() and so the children() is not
working.
My brain has frozen on how to get around it.
E.g. I have this: (images in forum)

use <fillet.scad>
// from clothbot
// - https://github.com/clothbot/ClothBotCreations/tree/master/utilities

d0 = 10;
h1 = 30;
h2 = 20;
angle = -66;

// Parkinbot approach (filleted)
translate([30,20,0])
fillet(r=2,steps=2) {
rotate([angle,0,0])
cylinder(h1, d=d0);
cylinder(h2, d=d0);
// join
color("red")
sphere(d0/2);
}

// ctchin approach (filleted)
translate([30,-20,0])
fillet(r=2,steps=4) {
rotate([angle,0,0]) cylinder(h1, d=d0);
cylinder(h2, d=d0);
// join
color("red")
intersection() {
rotate([angle,0,0])
translate([0,0,-2d0])
cylinder(d=d0,h=2d0);
translate([0,0,-2d0])
cylinder(d=d0,h=2d0);
}
}

//-----------------------------
// Each branch is a cylinder with [diameter, startpt, anglexyz, length]
//    - startpt, anglexyz are triplets
// generalize later to use shape profile instead of circle (cylinder)
// (one day alllow random variation in deformation along length of the
branch object)
branches = [
[10, [0,0,0], [0,0,0], 40],
[10, [0,0,0], [45,0,0], 40],
[5,  [0,0,30], [66,-22,0], 33],
[5,  [0,0,25], [66,-22,250], 33],
];

module make_tree(branches=branches, blend=4, method="round") {
// add round/sharp in here
// how to get around implied union of for() ?
fillet(r=2,steps=4) {
for (b=branches) {
dia = b[0];
start = b[1];
angle = b[2];
dist = b[3];
translate(start)
rotate(angle)
cylinder(h=dist, d=dia);
}
}
}

make_tree();

To get a result like this:
http://forum.openscad.org/file/n14635/branches_5.png

How do I pass all the results of the for loop to fillet() as children ?

--
View this message in context: http://forum.openscad.org/Non-Linear-Transformations-tp14539p14635.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

Usually, an (implicit) union joins overlapping or matching (there is always a
numerical error!) objects. This happens also with polyhedrons. If you
calculate your own polyhedrons (your example doesn't look so) just provide
some very small overlapping.
BUT, when I try to implement your example along the "broadway", say, like
this:

I'm getting this:
http://forum.openscad.org/file/n14636/weird.png

I remember that the convexity parameter of linear_extrude() never worked
properly with me, so in the end to get more out of it I implemented a
fullblown library to bypass it on the basis of hull(). Just have a look at:
http://www.thingiverse.com/thing:648813

runsun wrote

http://forum.openscad.org/file/n14004/20150925_bent_pipe.png

--
View this message in context: http://forum.openscad.org/Non-Linear-Transformations-tp14539p14636.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

You are right. A better solution would be the following approach. I didn't
present it, as I forgot how to express a partial array (Was it ever possible
in OpenSCAD to express a partial array like branches[2:4] ) Overcoming this
obstacle with the tail function, this is your proper and fast solution
ver2.0:

(By the way: a nice approach .)

Neon22 wrote

--
View this message in context: http://forum.openscad.org/Non-Linear-Transformations-tp14539p14644.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

Yes I used this pattern
for (i=[1:len(branches)]) {
b = branches[i];

On 11/20/2015 2:03 AM, Parkinbot wrote:

parkinbot: Was it ever possible in OpenSCAD to express a partial array like
branches[2:4]?

Currently you must write
[for(i=[2:4])branches[i]]

Many people have asked for a more direct 'slice' operator, but we don't
have one yet. Multiple syntaxes have been proposed.

On 19 November 2015 at 08:03, Parkinbot rudolf@parkinbot.com wrote: