Anno domini 2026 Thu, 28 May 16:22:14 +0200
Karl Exler via Discuss scripsit:

Hi!

You could load the mesh in MeshLab and reapair/close it there. https://www.meshlab.net

Nik

--
Please do not email me anything that you are not comfortable also sharing with the NSA, CIA ...

If the model will not convert within OpenSCAD, it won't be something to be repaired externally.
Consider to post the link of the original file and also present the changes you've made. If the original file renders properly, the changes listing will be critical to determine the solution.
On Thursday, May 28, 2026 at 10:48:45 AM EDT, Nikolaus Klepp via Discuss discuss@lists.openscad.org wrote:

Anno domini 2026 Thu, 28 May 16:22:14 +0200
Karl Exler via Discuss scripsit:

Hi!

You could load the mesh in MeshLab and reapair/close it there. https://www.meshlab.net

Nik

--
Please do not email me anything that you are not comfortable also sharing with the NSA, CIA ...

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

On Thu, May 28, 2026 at 04:22:14PM +0200, Karl Exler via Discuss wrote:

OpenScad has, for ages, used the CGAL library for rendering the
geometry that you've defined.

About two years ago someone started a project to replace CGAL. It is
called "manifold". The openscad programmers, have for a while kept the
feature "included, but disabled by default" to allow for more
testing. I'm not sure if it would be "enabled by default" on the most
recent version or not. That said, if you have the latest stable release
it certainly has it available!

If you switch to "manifold" as opposed to CGAL you get a few benefits.

1. It is orders of magnitude faster in some cases.(*)
2. it accepts many of the STLs that CGAL doesn't.

You need to move to a reasonably recent openscad, and possibly
go into the preferences to enable manifold.
On my version it is
edit -> preferences -> features -> Manifold.

Even is the maintainers haven't enabled it by default, I've never seen
a problem caused by manifold, on the contrary I've had big struggles
with exactly the problem you've been having.

I previously was under the impression that everybody on the internet
was producing and sharing "bad", "broken" STL files. As "manifold"
does accept them, it seems that the problem with those STLs somehow
originate in CGAL.

Somehow I am under a strong impression that you're still using
CGAL:

Roger. 

(*) It doesn't really matter if your render takes 0.001 or 0.1
seconds. But once it renders in half a second as opposed to 10
minutes, you start to appreciate the speed. :-)

--
** R.E.Wolff@BitWizard.nl ** https://www.BitWizard.nl/ ** +31-15-2049110 **
** Verl. Spiegelmakerstraat 37 2645 LZ  Delfgauw, The Netherlands.
** KVK: 27239233    **
f equals m times a. When your f is steady, and your m is going down
your a** is going up.  -- Chris Hadfield about flying up the space shuttle.
**  'a' for accelleration.

On Thu, 2026-05-28 at 16:22 +0200, Karl Exler via Discuss wrote:

If you are running Windows 10 or 11, and if you have (or can download)
Bambu Studio, you can load up the STL, and when it tells you it's not
manifold, click on 'Repair'. It will try to fix it with a Microsoft
program.

I have found it to be pretty good.

Larry.

Larry, OpenSCAD will not create an STL, based on the error message. One cannot repair something that can't be created. The problem lies within the code modified by the OP.
On Thursday, May 28, 2026 at 11:47:24 AM EDT, larry via Discuss discuss@lists.openscad.org wrote:

On Thu, 2026-05-28 at 16:22 +0200, Karl Exler via Discuss wrote:

If you are running Windows 10 or 11, and if you have (or can download)
Bambu Studio, you can load up the STL, and when it tells you it's not
manifold, click on 'Repair'. It will try to fix it with a Microsoft
program.

I have found it to be pretty good.

Larry.

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

On Thu, May 28, 2026 at 09:46:27AM -0600, larry via Discuss wrote:

Guys, Stop telling this guy that it's the STL that's broken. It is
OpenScad / CGAL.

I found numerous STLs that you import, hit render: All is OK. Then you
intersect it with something or subtract anything... As soon as there
is overlap, you get exactly these errors. Then, switch to manifold,
and everything is fine.

It is CGAL or CGAL in combination with OpenScad. Yeah, the error message
is misleading because it points you in the wrong direction.

In my case I had a "wind turbine blade" STL and I needed to cut it into
smaller sections to print it (I printed three blades (eventually 4) at
2.54 m length per blade).

To demonstrate my point I tried reproducing what I did a long time
ago, but today I can't even import the STL. Not my project, so I'm not
going to spend more time on it.

Roger. 

--
** R.E.Wolff@BitWizard.nl ** https://www.BitWizard.nl/ ** +31-15-2049110 **
** Verl. Spiegelmakerstraat 37 2645 LZ  Delfgauw, The Netherlands.
** KVK: 27239233    **
f equals m times a. When your f is steady, and your m is going down
your a** is going up.  -- Chris Hadfield about flying up the space shuttle.
**  'a' for accelleration.

Hi,

There is often confusion between manifold, (the 3d calculation method),
and manifold, the object description

The following code shows some things to try.

//////////
polyhedron(points=[[0,0,0],[0,0,30],[0,30,0],[0,30,30],[30,0,0],[30,0,30],[30,30,0],[30,30,30]],
faces=[[1,0,4],[2,4,0],[1,3,0],[3,1,5],[3,2,0],[3,7,2],[5,4,6],[5,1,4],[6,4,2],[7,6,2],[7,3,5],[7,5,7]]);
//last should be 6, not 7 i.e [7,5,6] cube(10);

///////////

the polyhedron is a cube with an error, (the last tri has been altered.)

In order for booleans to take place, the shapes must be
manifold/watertight (each edge shared by two faces)

if you load the above code in openscad, and fn5, you will see that the
large cube has one triangle missing, that shape is not watertight. Run
f6, and the large cube disappears. If manifold, not cgal, is selected,
then the small cube disappears, too, but if cgal is selected, then the
small cube remains. Correct the last face in the code, and fn6 works for
both cases. You can select cgal or manifold from
'edit-preferences-advanced-backend', if your version of openscad is
fairly recent. If you have a downloaded stl/whatever, then add a small
cube, and fn6. If the shape disappears, you know it is not watertight.
If you difference the small cube from the faulty large cube, that is a
boolean, and that will only produce a result if both shapes are watertight.

cgal is more thorough than manifold in may ways. Did you note the the
fact that when the shapes were not differenced, it showed the smaller
watertight cube, whereas manifold render showed nothing.

fn 5 , fast render, give a very rough idea of any shapes, you cannot
export that as an stl/whatever. There is a multitude of non watertight
shapes, generated by blender and elsewhere on the various sites. They
can often be 3d printed, but unless they are repaired, that is generally
all you can do. If you want to give the link, and your code, we can
probably tell if it is your code, or the cone, or both that are faulty.

Best wishes,

Ray

On 28/05/2026 15:22, Karl Exler via Discuss wrote: