I can't follow the code as I don't speak German and the error message I get
from running it doesn't make sense.

ECHO: "Pf = ", 268.803, " LF : ", 314.159, " * ", 3, " = ", 942.478, " * ",
3

WARNING: Bad range parameter in for statement: too many elements
(4294967295), in file RepRaptor, line 36

TRACE: called by 'for', in file RepRaptor, line 36.

TRACE: called by 'for', in file RepRaptor, line 35.

TRACE: called by 'for', in file RepRaptor, line 29.

TRACE: called by 'union', in file RepRaptor, line 24.

TRACE: called by 'difference', in file RepRaptor, line 21.

TRACE: called by 'translate', in file RepRaptor, line 19.

TRACE: called by 'SandFilter', in file RepRaptor, line 47.

The file is Untitled.scad, there is no file called RepRaptor.

I don't know where 4 billion comes from, seems to be 2^32-1.

The likely solution to it being too slow is to do it in 2D and linear
extrude the result.

On Sat, 9 May 2020 at 12:48, Bruno Boettcher bboett@gmail.com wrote:

Indeed... argh.... i could kick myself....

still its very slow....

ok, here's the corrected (and commented, does that help?) code:

//filter for a pump, outside diameter 4cm, partikel size as argument as an
offset to space them up
//overkill, factor to increase the needed surface, 1 same surface
module Filter(partikel,luft, overkill)
{
pumpFlaeche = PI18.518.5/4;//surface of the pumphole
lochFlAeche = PI * partikelpartikel/4;//surface of the individual hole
anzahl = ceil(PI18.5/(partikel+luft));//number of holes on the radius of
the filter
einKreisF = anzahl * lochFlAeche; //surface of one circle of holes
kreisAnzahl = ceil(pumpFlaeche/einKreisF)+2;//number of circles needed +2
to compensate for borders
//echo("Pf = ",pumpFlaeche, " LF : ", lochFlAeche," * ",anzahl," = ",
einKreisF," * ",kreisAnzahl);
difference()
{
cylinder(d=40, h=overkillkreisAnzahl(partikel+luft),$fn = 64);
union()
{
schritt = 1*(partikel+luft);
translate([0,0,-1])
cylinder(d=36, h=overkillkreisAnzahl(partikel+luft)-1,$fn = 64);
for(y = [1 : schritt:overkill*(kreisAnzahl)(partikel+luft)-1])
{
translate([0,0,y])
for(w = [0 : 360/anzahl:360])
rotate([0,0,w+((y/schritt)%290/anzahl)])

color("red")translate([0,30,0])rotate([90,0,0])cylinder(d=partikel,
h=60,$fn = 64);
for(r = [partikel : (partikel+luft):18])
for(w = [0 : 360/ceil(PIr/(partikel+luft)):360])
rotate([0,0,w])
translate([0,r,overkillkreisAnzahl*(partikel+luft)-2])
color("red")cylinder(d=partikel, h=10,$fn = 64);
}
}
}
}

Am Sa., 9. Mai 2020 um 14:32 Uhr schrieb Torsten Paul Torsten.Paul@gmx.de:

--
ciao
Bruno

---==========
http://nohkumado.eu/, http://bboett.free.frhttp://aikido.nohkumado.eu/,
http://bboett.free.fr
http://aikido.zorn.free.fr

1h later we are at 143/1000 for the compilation :(

Am Sa., 9. Mai 2020 um 15:06 Uhr schrieb Bruno Boettcher bboett@gmail.com:

--
ciao
Bruno

---==========
http://nohkumado.eu/, http://bboett.free.frhttp://aikido.nohkumado.eu/,
http://bboett.free.fr
http://aikido.zorn.free.fr

OK I can see my 2D suggestion makes no sense now.

I think it will go a lot faster if you stop the cylinders intersecting at
the middle. Any 3D operation with so many objects will be slow though.

On Sat, 9 May 2020 at 15:13, Bruno Boettcher bboett@gmail.com wrote:

On 09.05.20 16:40, nop head wrote:

Also it seems the code is generating the top holes multiple
times.

Maybe check if $fn = 64 is really needed for all those tiny
holes.

ciao,
Torsten.

Besides the high $fn values for holes, your code imposes extra avoidable
work to CGAL. For instance, the for loop that makes the top holes is inside
the for loop of the radial holes; the cylinders making the ring of radial
holes are duplicated and unnecessarily longs. You could use a high $fn
value for the outer filter cylinder but keep a smaller value for the hole
cylinders.

Correcting those flaws and redesigning a little the order of the
operations, I was able to render your filter in 2min 48sec while your
original code required 13min 26sec in my machine, both running with $fn=64
for the outer cylinder and $fn=24 for the others. If you need I can share
my code.

Em sáb., 9 de mai. de 2020 às 16:03, Torsten Paul Torsten.Paul@gmx.de
escreveu: