Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for
Show only
|
Search instead for
Did you mean:
This page has been translated for your convenience with an automatic translation service. This is not an official translation and may contain errors and inaccurate translations. Autodesk does not warrant, either expressly or implied, the accuracy, reliability or completeness of the information translated by the machine translation service and will not be liable for damages or losses caused by the trust placed in the translation service.Translate
What would be extremely usefull for people who 3D print a lot is to be able to convert STL files (on import) to real bodies that can be edited (push / pull / rotation / modification etc.).
I would love to live in a world with this feature!
I am fairly certain it is computationally impossible to convert an arbitrary mesh into an accurate solid, but they have said such things about lots of things that are now quite possible.
As a 2D metaphor I like to use is to think of it like Autodesk Vectorize It or LiveTrace/ImageTrace in Adobe Illustrator. You have a chaotic, high entropy JPEG that you want to convert into the elegant precision of vectors. Infinite resolution, smooth curves, great editablity, all this and more is yours with vectors! Unfortunately, even 10 years after the release of liveTrace in Illustrator it is still lightyears behind hand tracing, even for the simplest of shapes. Corners look weird, lines waver in strange and senseless ways— visual artifacts that can be seen all over billboards the world over.
The culture of Thingiverse etc. drives me a little crazy, where people upload STL mesh files of objects that were modeled as solids by the thousands. These files are intrinsically chaotic and essentially uneditable. They can be made only as they are using todays 3D printers. Tomorrows printers, with better resolution? Hello polygons. Todays CNC machines? Good luck. Few CAM packages can do the meshy-math to run toolpaths on a mesh. Tools like MeshMixer help a little, but compared to a SAT, STEP, or IGES solid file it is comparatively hopeless. I wish folks would upload solids along with their messy meshes.
Anyway, I have wanted this dozens of times.
Most existing tools for this do a horrible job. They simple turn each triangle of the mesh into a surface. This often crashes your computer, because there are sometimes millions of triangles.
It has worked for me in the past to convert 3D scan meshes into CNC machinable solids, but I have never tried running it on a mesh that is simple like the one in your image. I'd love to see a Fusion port of mesh enabler. I'd also like to see a version of it that has a real interface and some options. As it stands it is just one button, and often it will just crash if your mesh is intense.
Rhino has an equally poorly featured command called Mesh2Nurb, which I mention because they have a long post about the (intrinsic?) shortcomings of a mesh-surface conversion.
I have seen on the internet outrageously expensive software that guides a user in the manual recreation of solids from mesh data. The class of tool is called reverse engineering mesh to surface because it is almost never used by the original author of a part. They mostly act as systems to help users trace the parts manually and recreate features.
This post was intended to be much shorter... I just hate meshes so much. 🙂
@RobbGodshaw "I am fairly certain it is computationally impossible to convert an arbitrary mesh into an accurate solid, but they have said such things about lots of things that are now quite possible."
Hi Robb, i'm not so sure. You have vertexes, you have faces, you know what's inside from outside of the object with the normals. There are software that decimates faces to simplify STL files and that works ... "relatively" well ;).
I'm more concern by cracks. There are often a lot of them in stl files and that means having a solid object before doing anything else.
Meshmixer is able to give the exact volume of an stl file. So i suppose that it's able to "see" the outside geometry of the model, right ? otherwise how could it be accurate (unless it isn't 😉 ) ?
F360 is able to veneer nurbs around STL volumes (like scan of hand or arm...). I'm not a 3D ingeneer, but maybe the trick could be somewhere around this (veneering nurbs around stl volume)... Or maybe what i'm saying is a complete nonsense 😉
I agree with you : i hate meshes too! That's why i would be very happy if autodesk could solve this ! I know voxels are comming and it may solve a lot of things for 3D printers, but i think we're going to be stuck with STLs files for some more time, so having a solution for this would definitely be awesome...
The general problem is very difficult for meshes (without generating a huge mess), but recognizing things that are co-planar, or in a limited number of geometries is totally doable.
Voxels would be much harder for something like F360 to deal with, though.
I always presumed this would be impractical to even attempt to implement (beyond the current convert to brep function), but if it can be done reasonably well, even as approximation that then could be user tweaked, I'm ALL IN.
The first company to ever manage to successfully pull this off, and pull stl files in to be converted to true conventional CAD models, as in the original post, has a game changer right there. As far as 3D printing is concerned there would be that company and .... nobody. End of story.
So the last post I see on this topic (problem) was from last year. Has there been any progress on STL to IGES file conversion? Are all polygon CAD models really redrawn from scratch, in order to manufacture parts (not including 3D pronting). Any 3rd party file conversion offerings available. I would prefer to remain in the Fusion 360 universe. Please advise.
I was trying to import the Meshmixer Bunny into Fusion so that I could modify it and print it for a friends birthday present. What I ended up doing was taking sketch sections and lofting between them. If the development team could make a feature that would create many different sections and then loft between them couldn't that work for conversion to an editable body.
Along with this, we need a way to convert triangular mesh to quad mesh. Sometimes you need to convert the surface manually to make the geometry right, and Fusion 360 handles quad mesh much better than an imported STL.
As an avid member of the 3D printing community and maker movement, I could definitely see the many upsides that this could bring to the table. Further direct integrationcwith meshmixer would also be great!
This feature would be huge. As of right now, it is difficult to accomplish this without having expensive software. It can be done with inventor by converting stl to dwg. This link explains how(https://www.cadforum.cz/cadforum_en/convert-surface-mesh-model-to-3d-solid-body-tip8183). During the process, sometimes you have to close a popup inventor window with task manager for it to work.
There are programs that can do this already. Power Shape or Geomagic Design X for example but look up the price, don't see this technology coming to Fusion for free. Conversion of prismatic shapes like in the original post could be done quite easily. As in @odolyte picture you can do it manually now in Fusion so an option to automate on import should be doable. Geomagic\Alibre 3d had an option years ago that could import prismatic shapes with round holes and translate to solids on import. It only worked on simple stuff and was pretty disappointing.
![Autodesk Design & Make Report](https://forums.autodesk.com/t5/image/serverpage/image-id/1350519iCB55845EBFC43016/image-size/large?v=v2&px=999" "Autodesk Design & Make Report")