Jon,

I uploaded the newest version of Naca_sweep.scad to
http://www.thingiverse.com/thing:1208001. The sweep() in this version has
new parameters:

Using the default "planar_caps=false" triangulates the caps, as Ronaldo
already stated. Actually, I didn't have any problems to F6-render the code
with planar_caps set to true or false, using OpenSCAD 2017.01.20. Let me
anote that it might be a good idea to strengthen the blade at its root by
altering the last column ZZ in the matrix A. Try e.g.:

--
View this message in context: http://forum.openscad.org/error-preparing-boat-propeller-tp20570p20574.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

That did the trick!  Thank you!

Jon

On 2/25/2017 1:09 PM, Parkinbot wrote:

So, now on to scaling.

We want a 1/2" shaft with a 3/16" center bore and a dog slot.  We also
want propellers that are roughly 3" in diameter.

I tried to express everything in native sizes by changing the "size"
value to 1, but that threw everything off, even after I tried to adjust
the values in the matrix by dividing them all by 10.  At the moment, the
shaft is correctly scaled, but the propellers do not touch the shaft,
and they are way too large.

I don't mean to be ungrateful for this amazing technology, but would it
be possible to allow some of these things to be adjusted in a way that
was less obscure and simpler?  If the answer is "I don't have time for
this", I completely understand.

Thanks for listening!

Jon

use <naca4.scad>      // see http://www.thingiverse.com/thing:898554
use <shortcuts.scad>  // see http://www.thingiverse.com/thing:644830
use <naca_sweep.scad> // see http://www.thingiverse.com/thing:900137
use <splines.scad>

// some parameters
inches = 25.4;
N = 100;                        // refinement
R_axis = 10inches * 0.25;        // radius of axle
r_bore = 10inches * 3/16/2;    // radius of axle bore
size = 0.1;                        // sizing factor
slices = false;                    // set true to view interpolation slices
ty = 50;                        // common y-distancing of blades
dog = inches * 0.125;

// core data to be interpolated by nSpline()
//      transform data            airfoil nacaXYZZ
//--------------------------------------------------
A = [ // Tx    Ty  Tz  Rz  Ry  |  L    X    ZZ
[0,    ty,  60, -25,  0,  300,  .12,  .2],
[-50,  ty,  200, -50,  0,  600,  .2,  .12],
[-100, ty,  500, -75,  0,  450,  .1,  .06],
[00,  ty,  660, -82,  10,  10,  .01,  .1],
];

// Main object
module MyProp()
Rz(180)
S(size, size, size)
boat_prop(3);

difference() {
MyProp();
// drill out center hole
rotate([0, 90, 0])
cylinder(h = 100, r = 3.75, $fn = 100);
// dog slot
cube([dog, 2 * R_axis/10 + 1, dog], center = true);
}

module boat_prop(M = 2)
{
m = slices?1:M; // used for slices view
$fn = N;        // refinement
shift = -130;  // shift for axis parts
shl = 270;      // length of axis
B = nSpline(A, 2*N);  // interpolation

D()  // difference
{
U()  // union of blades and axle
{
for(i=[0:360/m:359])
{
C = gen_dat(slices?A:B,N);
R(i)                // = Rx(); x-rotates object
sweep(C, showslices=slices);
}
if(!slices)
axis(shift, shl);
}
Tx(shift+shl*.2)  // translate
Ry(90)            // rotate
Cy(r_bore,shl*.9); // cylinder
}

// generate all data - applies transformation and airfoil params
function gen_dat(B,N) =
[for(i=[0:len(B)-1])
let(R = B[i])
Tz_(R[2],  // z-translate 2D-shape
R_(R[4], 0, 180+R[3], // rotate 2D-shape
T_(R[0],R[1], 0, // xyz-translate 2D-shape
vec3D(airfoil_data(N=N, L=R[5], naca=[R[6],.5,R[7]], open=true)
))))
];
}

module axis(shift = 0, shl = 200)
{
T(shift)        // = Tx()
Ry(90)
Cy(R_axis,shl);
T(-shl/2+shift)
Ry(90)
Sp(R_axis);    // sphere
}

On 2/25/2017 1:09 PM, Parkinbot wrote:

Jon,

if you see what you get within less than 100 lines, everything is as simple
as can be ;-)
I admit, the design uses of some magic library stuff concerning the blades.
But the rest, hub and bore is just OpenSCAD stuff using scale, difference
and union. You should be able to work this out on your own, also the mm/inch
rescale.

The matrix describes a single blade using a series of four slices being then
interpolated by nSpline():

A = [ // Tx    Ty  Tz  Rz  Ry  |  L    X    ZZ
[0,    ty,  60, -25,  0,  300,  .12,  .2], // root slice is
translated by [0, ty, 60], pitched by 25°
[-50,  ty,  200, -50,  0,  600,  .2,  .12], //
[-100, ty,  500, -75,  0,  450,  .1,  .06], //
[00,  ty,  660, -82,  10,  10,  .01,  .1], // tip slice
];

The first line describes the root, the last line the tip profile. The second
and third are intermediate profiles to "guide" the cubic spline
interpolation. You can introduce more lines, but they have to "fit into the
scheme", otherwise the interpolation scheme will freak out. As for the
columns the meaning is
Tx = translate slice along x
Ty = translate slice along y,
Tz = translate slice along z, (distance from axis)
Rz = rotate slice around z, (pitch)
Rz = rotate slice around y
used here

and
L = NACA cord length
X = NACA camber
Y = NACA location of max camber = 0.5 (not interpolated in this example)
ZZ = NACA factor thickness/cord
used here:

The first five parameters describe how the 2D-shape is positioned in 3D, the
last three columns describe the 2D shape itself. It is a 4 digit Naca
profile see https://en.wikipedia.org/wiki/NACA_airfoil.
The 4 digit naming scheme I used is: "nacaXYZZ", with X, Y, ZZ denoting
profile parameters. The matrix adresses the cord length L and only the X and
the ZZ parameters. The generator function calls into my naca4.scad library
to construct the 2D shape. It has a "constructor" airfoil_data() with an
overloaded "naca" parameter. One overload accepts NACA-Numbers like 1230,
and the other a vector like [.1, .2, .3]. Both result in the same profile,
NACA1230.

Hope this helps.

--
View this message in context: http://forum.openscad.org/error-preparing-boat-propeller-tp20570p20614.html
Sent from the OpenSCAD mailing list archive at Nabble.com.

Very helpful, indeed!  Thank you!  I will keep this handy going forward!

I tried to make the props smaller by using a scale(0.5) on My Prop()
[see line 34].  The result is startling.  The hub vanishes.

Jon

use <naca4.scad>      // see http://www.thingiverse.com/thing:898554
use <shortcuts.scad>  // see http://www.thingiverse.com/thing:644830
use <naca_sweep.scad> // see http://www.thingiverse.com/thing:900137
use <splines.scad>

// some parameters
inches = 25.4;
N = 100;                        // refinement
R_axis = 10inches * 0.25;        // radius of axle
r_bore = 10inches * 3/16/2;    // radius of axle bore
size = 0.1;                        // sizing factor
slices = false;                    // set true to view interpolation slices
ty = 40;                        // common y-distancing of blades
dog = inches * 0.125;

// core data to be interpolated by nSpline()
//      transform data            airfoil nacaXYZZ
//--------------------------------------------------
A = [ // Tx    Ty  Tz  Rz  Ry  |  L    X    ZZ
[0,    ty,    0, -25,  0,  300,  .12,  .2],
[-50,  ty,  200, -50,  0,  600,  .2,  .12],
[-100, ty,  500, -75,  0,  450,  .1,  .06],
[00,  ty,  660, -82,  10,  10,  .01,  .1],
];

// Main object
module MyProp()
Rz(180)
S(size, size, size)
boat_prop(3);

difference() {
scale(0.5)
MyProp();
// drill out center hole
rotate([0, 90, 0])
cylinder(h = 100, r = 3.75, $fn = 100);
// dog slot
rotate([-40, 0, 0])
cube([dog, 2 * R_axis/10 + 1, dog], center = true);
}

module boat_prop(M = 2)
{
m = slices?1:M; // used for slices view
$fn = N;        // refinement
shift = -130;  // shift for axis parts
shl = 270;      // length of axis
B = nSpline(A, 2*N);  // interpolation

D()  // difference
{
U()  // union of blades and axle
{
for(i=[0:360/m:359])
{
C = gen_dat(slices?A:B,N);
R(i)                // = Rx(); x-rotates object
sweep(C, showslices=slices);
}
if(!slices)
axis(shift, shl);
}
// bore out the axle
Tx(shift+shl*.2)  // translate
Ry(90)            // rotate
Cy(r_bore,shl*.9); // cylinder
}

// generate all data - applies transformation and airfoil params
function gen_dat(B,N) =
[for(i=[0:len(B)-1])
let(R = B[i])
Tz_(R[2],  // z-translate 2D-shape
R_(R[4], 0, 180+R[3], // rotate 2D-shape
T_(R[0],R[1], 0, // xyz-translate 2D-shape
vec3D(airfoil_data(N=N, L=R[5], naca=[R[6],.5,R[7]], open=true)
))))
];
}

module axis(shift = 0, shl = 200)
{
T(shift)        // = Tx()
Ry(90)
Cy(R_axis,shl);
T(-shl/2+shift)
Ry(90)
Sp(R_axis);    // sphere
}

On 2/27/2017 3:50 PM, Parkinbot wrote: