Why not update? It is already about a year old, and manifold changed a lot.
On 12/3/24 15:51, nop head via Discuss wrote:

I am using a version from January so, yes it is due to the massive cube in
the below module. Removing a zero fixes it.

module below(bh=1,ht=25){
intersection(){translate([0,0,-1000+ht])
cube([1000,1000,2000],true);
children();}
}

On Tue, 3 Dec 2024 at 07:42, Chun Kit LAM via Discuss <
discuss@lists.openscad.org> wrote:

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

Why not update? It is already about a year old, and manifold changed a
lot.

On 12/3/24 15:51, nop head via Discuss wrote:

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

Thanks to all of you for the replies. Glad I was not alone. (fwiw, I
found that fast csg gave errors, but normal csg was OK, if restarted,
but that would be too slow for where I intended to use this). This
stemmed from trying an earlier post '[OpenSCAD] making a round
container with the help of rotate (?)' by FPA. Part of the the
solution I posted used Minkowski, which was very slow in rounding the
top of the partitions and tray. Since I think we are only concerned
with 3d printing, by choosing steps of 0.1mm, then the bevel top gives
a good result.  (In any case, for fdm, chamfers are better than
rounding). This solution, that I posted, was intended to be a generic
solution to that sort of chamfering. It produces suitable results far
faster than Minkowski, but is obviously restricted to flat sections of
the solid.

I chose a large cube  for the intersection, since, as yet, I do not
have an fdm printer with a bed of that size, and i needed to allow for
the model being offset.

I'll update my version, and see how it goes.

Best wishes,

Ray

On 03/12/2024 08:54, Chun Kit LAM via Discuss wrote:

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

I've installed the build for last night. There is no problem with
Manifold, nor fast csg, since they, or much else of the earlier
features, are listed in the new features. Anyway, it works. i guess
manifold was a test situation.

There is a problem with this, in as such, now I can produce working
scripts, that will not run on earlier versions. Not a problem for me,
but obviously could be for others.

Best wishes,

Ray

On 03/12/2024 11:31, Raymond West via Discuss wrote:

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

I think i may revert to my earlier version. The latest, without manifold,
is far, far too slow.

Here is the console output for the February version using manifold.

Compiling design (CSG Tree generation)...
ECHO: 0.7, 56, 102.857
Rendering Polygon Mesh using Manifold...
Geometries in cache: 86
Geometry cache size in bytes: 3013408
CGAL Polyhedrons in cache: 0
CGAL cache size in bytes: 0
Total rendering time: 0:00:00.374
Top level object is a 3D object:
Facets:      31732
Rendering finished.

and for the latest version

Compiling design (CSG Tree generation)...
ECHO: 0.7, 56, 102.857
Rendering Polygon Mesh using CGAL...
Geometries in cache: 632
Geometry cache size in bytes: 4620624
CGAL Polyhedrons in cache: 101
CGAL cache size in bytes: 216265592
Total rendering time: 0:01:30.395
Top level object is a 3D object (Nef polyhedron):
Simple:    no
Vertices:    16240
Halfedges:  49032
Edges:      24516
Halffacets:  16562
Facets:      8281
Volumes:        4
WARNING: Object may not be a valid 2-manifold and may need repair!
Rendering finished.

The above results were for exactly the same script, and only a relatively
small part of the project. The fact that the cgi version threw a warning,
does not matter in this case, since the slicer did not bulk at the stl file
produced by the manifold version.

For anything useful to me,  the cgi rendering is now far too slow.  I
recollect someone mentioning it would be removed, a while back, but If it
gave a problem, then there was the option of re-rendering in the cgi
engine.

Are there any newer settings to speed it up a bit?

On 03/12/2024 16:18, Raymond West via Discuss wrote:

I've installed the build for last night. There is no problem with
Manifold, nor fast csg, since they, or much else of the earlier features,
are listed in the new features. Anyway, it works. i guess manifold was a
test situation.

There is a problem with this, in as such, now I can produce working
scripts, that will not run on earlier versions. Not a problem for me, but
obviously could be for others.

Best wishes,

Ray
On 03/12/2024 11:31, Raymond West via Discuss wrote:

Thanks to all of you for the replies. Glad I was not alone. (fwiw, I found
that fast csg gave errors, but normal csg was OK, if restarted, but that
would be too slow for where I intended to use this). This stemmed from
trying an earlier post '[OpenSCAD] making a round container with the help
of rotate (?)' by FPA. Part of the the solution I posted used Minkowski,
which was very slow in rounding the top of the partitions and tray. Since I
think we are only concerned with 3d printing, by choosing steps of 0.1mm,
then the bevel top gives a good result.  (In any case, for fdm, chamfers
are better than rounding). This solution, that I posted, was intended to be
a generic solution to that sort of chamfering. It produces suitable results
far faster than Minkowski, but is obviously restricted to flat sections of
the solid.

I chose a large cube  for the intersection, since, as yet, I do not have
an fdm printer with a bed of that size, and i needed to allow for the model
being offset.

I'll update my version, and see how it goes.

Best wishes,

Ray
On 03/12/2024 08:54, Chun Kit LAM via Discuss wrote:

Why not update? It is already about a year old, and manifold changed a lot.
On 12/3/24 15:51, nop head via Discuss wrote:

I am using a version from January so, yes it is due to the massive cube in
the below module. Removing a zero fixes it.

module below(bh=1,ht=25){
intersection(){translate([0,0,-1000+ht])
cube([1000,1000,2000],true);
children();}
}

On Tue, 3 Dec 2024 at 07:42, Chun Kit LAM via Discuss <
discuss@lists.openscad.org> wrote:

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email to discuss-leave@lists.openscad.org

I suggest you turn on Manifold in the newer version:
Preferences-->Advanced-->3D Rendering-->Backend

Dan

On Tue, Dec 3, 2024 at 6:53 PM Raymond West via Discuss
discuss@lists.openscad.org wrote:

 I think i may revert to my earlier version. The latest, without
 manifold, is far, far too slow.

 Here is the console output for the February version using manifold.

 Compiling design (CSG Tree generation)...
 ECHO: 0.7, 56, 102.857
 Rendering Polygon Mesh using Manifold...
 Geometries in cache: 86
 Geometry cache size in bytes: 3013408
 CGAL Polyhedrons in cache: 0
 CGAL cache size in bytes: 0
 Total rendering time: 0:00:00.374
 Top level object is a 3D object:
    Facets:      31732
 Rendering finished.


 and for the latest version

 Compiling design (CSG Tree generation)...
 ECHO: 0.7, 56, 102.857
 Rendering Polygon Mesh using CGAL...
 Geometries in cache: 632
 Geometry cache size in bytes: 4620624
 CGAL Polyhedrons in cache: 101
 CGAL cache size in bytes: 216265592
 Total rendering time: 0:01:30.395
 Top level object is a 3D object (Nef polyhedron):
    Simple:     no
    Vertices:    16240
    Halfedges:   49032
    Edges:       24516
    Halffacets:  16562
    Facets:       8281
    Volumes:         4
 WARNING: Object may not be a valid 2-manifold and may need repair!
 Rendering finished.

 The above results were for exactly the same script, and only a
 relatively small part of the project. The fact that the cgi
 version threw a warning, does not matter in this case, since the
 slicer did not bulk at the stl file produced by the manifold version.

 For anything useful to me,  the cgi rendering is now far too
 slow.  I recollect someone mentioning it would be removed, a while
 back, but If it gave a problem, then there was the option of
 re-rendering in the cgi engine.

 Are there any newer settings to speed it up a bit?



 On 03/12/2024 16:18, Raymond West via Discuss wrote:

 _______________________________________________
 OpenSCAD mailing list
 To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

I suggest you turn on Manifold in the newer version:
Preferences-->Advanced-->3D Rendering-->Backend

Dan

On Tue, Dec 3, 2024 at 6:53 PM Raymond West via Discuss
discuss@lists.openscad.org wrote:

 I think i may revert to my earlier version. The latest, without
 manifold, is far, far too slow.

 Here is the console output for the February version using manifold.

 Compiling design (CSG Tree generation)...
 ECHO: 0.7, 56, 102.857
 Rendering Polygon Mesh using Manifold...
 Geometries in cache: 86
 Geometry cache size in bytes: 3013408
 CGAL Polyhedrons in cache: 0
 CGAL cache size in bytes: 0
 Total rendering time: 0:00:00.374
 Top level object is a 3D object:
    Facets:      31732
 Rendering finished.


 and for the latest version

 Compiling design (CSG Tree generation)...
 ECHO: 0.7, 56, 102.857
 Rendering Polygon Mesh using CGAL...
 Geometries in cache: 632
 Geometry cache size in bytes: 4620624
 CGAL Polyhedrons in cache: 101
 CGAL cache size in bytes: 216265592
 Total rendering time: 0:01:30.395
 Top level object is a 3D object (Nef polyhedron):
    Simple:     no
    Vertices:    16240
    Halfedges:   49032
    Edges:       24516
    Halffacets:  16562
    Facets:       8281
    Volumes:         4
 WARNING: Object may not be a valid 2-manifold and may need repair!
 Rendering finished.

 The above results were for exactly the same script, and only a
 relatively small part of the project. The fact that the cgi
 version threw a warning, does not matter in this case, since the
 slicer did not bulk at the stl file produced by the manifold version.

 For anything useful to me,  the cgi rendering is now far too
 slow.  I recollect someone mentioning it would be removed, a while
 back, but If it gave a problem, then there was the option of
 re-rendering in the cgi engine.

 Are there any newer settings to speed it up a bit?



 On 03/12/2024 16:18, Raymond West via Discuss wrote:

 _______________________________________________
 OpenSCAD mailing list
 To unsubscribe send an email to discuss-leave@lists.openscad.org

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

I guess 'change the location to confuse the enemy.' 😳

Total rendering time: 0:00:00.152

thanks

On 03/12/2024 19:05, Dan Perry via Discuss wrote:

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org

Here is my latest complete bevel code. I'm happy with what it does. At
the moment, it only works with vertical, straight sided objects.
Useful for open-topped boxes, and they can easily be made stackable.
The 0.1 resolution is good enough for current fdm printers.  Note that
the bevels (really the are chamfers) are subtractive, as they are in
the machinist's  world.

module myshape(){
$fn=100;
  difference(){
    union(){cylinder(d=18,h=25);
            cube(25);}
    translate([8,8,-8]) cylinder(d=10,h=60);
  }
}

   ht=14;  // height of top of object
   bh=1.55;   // height of top bevel
   bb=2;   // height of bottom bevel
   bt=4;   // height of bottom of object

bevelobj(ht,bh,bb,bt)myshape();

module beveltop(bh=1,ht=25){  // bh = bevel height, ht = height to top
of bevel
   translate([0,0,ht-bh-0.1])
     for (j=[0:0.1:bh-0.1]){
      translate([0,0,j])
          linear_extrude(0.2)
          offset(delta=-1*j)
          projection(cut=true)
          translate([0,0,-ht])
          children();
    }
}

module bevelbot(bb=2,bt=0){  //bt= bottom of bevel, bb= bevel height
   translate([0,0,bt+bb])
     for (j=[0:0.1:bb]){
      translate([0,0,-j])
          linear_extrude(0.2)
          offset(delta=-1*j)
          projection(cut=true)
          translate([0,0,-bt])
         myshape();
    }
}

module middle(ht=10,bh=1,bb=2,bt=2){
     intersection(){
         translate([-5000,-5000,bt+bb])
         cube([10000,10000,ht-bh-(bt+bb)]);
         children();
     }
}

module bevelobj(ht,bh,bb,bt){
    middle(ht,bh,bb,bt)children();
    beveltop(bh,ht)children();
    bevelbot(bb,bt)children();

On 04/12/2024 11:07, Raymond West via Discuss wrote:

OpenSCAD mailing list
To unsubscribe send an email todiscuss-leave@lists.openscad.org