Subject:
|
Re: stl2dat conversion tool beta3
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Mon, 14 Feb 2000 19:55:40 GMT
|
Reply-To:
|
RUI.MARTINS@saynotospamLINK.PT
|
Viewed:
|
920 times
|
| |
|
|
On Mon, 14 Feb 2000, Marc Klein wrote:
> A new release of stl2dat is available here
> http://marc.klein.free.fr/stl2dat.html
Great, but I haven't tried it iet, I got sick this weekend, stomack sick,
haven't eaten all weekend, cheesh, Boy was I ungry last Night.
> It creates now correctly the type 2 lines.
> I removed the bug with "partial adjacent facets". (was a silly bug,
> triangles where merged too early)
> The part name, author name and angle limit can be given in the command
> line
OK.
> Some explanation of the conversion
> 1. merge adjacent coplanar triangles to quad facet
> 2. detect cylinder primitive
> 3. create type 2 and type 5 lines
> 4. merge adjacent coplanar triangles to triangle
Be carefull, with these merges, because every time you merge two polys, by
a facet, you remove 1 or 2 points (tris or quads). And you can only due
this if they are coplanar (which you mencioned already), but also, if
you can do the same with the new merged edge.
Example
1
/|\
2-3-4
merged into
1
/ \
2---4
Point 3 vanished, but you probably have two (or more) other polys wich
shared vertex 3, example:
1
/|\
2-3-4
| |\|
5-6-7
and if this polys aren't coplanar with the ones mencioned above, and among
them selves also, you can't remove the vertex 3, and if you don't you get
the BAD MODELLING feature I talked about in a previous mail, so in these
cases, you shouldn't merge anything.
By the way, tesselation routines, when done right produce just the
required amount of polys, which can't be further improved without changing
the shape.
The only exception is the fact that this particular format (.stl) only
allows for tris which can be merged into quads in only very specific
situations. so be carefull.
This makes me ask you to provide a disable flag/option to prevent these
optimizations in case they behave errouneously or in an insatisfactory
manner.
> Ideas for future implementation
> - Optional switch off of features of the program
> - Merge contiguous aligned type 2 lines
This one reveals the potencial for the same problem I mencioned above, so
be carefull.
> - Merge triangles and quad when possible
This one also.
> - Optimize creation of type 5 lines : no creation of unnecessary lines
> - Make a Dealaunay triangulation
Don't know this one !
> - Detect disc, ring, cone pimitives
> - Detect stud primitives
These are surrely a most, but it seems a bit too dificult to acomplish, if
the modeler, hasn't account for it, by using the primitives specific
measurements (4-4circle.dat comes to mind)
> stl2dat is an utilitary tool to convert .stl files
> issued from CAD software like Solidworks, Autocad, and many others to
> .dat files for ldraw compatible programs.
>
> I would appreciate any comments and suggestions.
> Marc Klein
> A Mindstorms fan.
About the comments & suggestions you got them right !
Rui Martins
|
|
Message is in Reply To:
 | | stl2dat conversion tool beta3
|
| A new release of stl2dat is available here (URL) creates now correctly the type 2 lines. I removed the bug with "partial adjacent facets". (was a silly bug, triangles where merged too early) The part name, author name and angle limit can be given in (...) (25 years ago, 14-Feb-00, to lugnet.cad.dev)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
