Conversion of virtual models into printable files

In the following weeks, a portion of about 65×35 cm was made in high resolution, to test the formats with the 3D printer. A sliced portion of 15×15 cm high resolution were late made to generate the g-code and make sure that the prints could be eventually produced.

After loading the model into the printer software, it was noted that:
.WRL file cannot be loaded.
.OBJ file can be loaded with the following results:
– Units of measurements are not correct exported to the file
X = 1.734
Y = 2.320
Z = 1.685
– The model contains inverted Normals.
– The model consists of 1 Shell, 18.797 Surfaces and 300.433 Triangles:
18.798 GAPS in the mesh which should be stitched or filled.
QUIBCX has suggested that all of the above mentioned issues could easily be corrected by GeoMagic.

It was advised that:

1) The issues should be corrected in GeoMagic and NOT in other tools. Applying this, we could prevent different algorithms will be used to solve the issues leading to unpredictable results.
2) Try to export the model in .ZPR format. .WRL is more complicated due to different versions in use.
3) The number of triangles should be reduced to prevent issues with larger models.

Accordingly, the mesh was optimized in Geomagic. A reamining issue was to get the precise number of the measurement due to zoom issues in Geomagic and the rough corner to put the trim point.

– the closest point that we could achieve for that moment was :  L : W : H = 20,029 cm ; 15 cm ; 1,125 cm
 – Units were kept in cm (for printing, it was decided to just multiply by 10). This because of unit calibration from the beginning process (from the scanner to Faro software to Geomagic).

– In this mesh processing,  Geomagic design 5.1.0.0 was used. In this case, Geomagic design has no option to export to ZPR format.
The available formats are .xtl, .obj, .3ds, .wrl (VRML 1.0 file and VRML 97 File), .ptc, .pts, .ply, .asc, .dxf, .icf, .xdl, .mdl)

– It was agreed that for the workshop in future, a reduced number of triangle will be provided. But for the test prints (15×20), the scanning team prefered to deliver a higher quality to show the depth of the crack.

Be Sociable, Share!

Comments are closed.

© 2011 TU Delft