tanya ismail

i've done a fair bit of 3d printing in the past and I didn't really have anything in mind for this project, so with my limited time this week I decided to just to test out printing delicate wire frame structures on the makerbot and zcorp. I used the majority of the time working on an Islamic pattern generator in Grasshopper and Rhino by combining arrays of polygons over grids, though there is still much work to do on overlays of cull patterns to achieve complex design. I took one of the radial patterns and projected it onto a spherical surface sitting them on elongated columns so I could test both the capability of machines in achieving both strength in long slender members, and definition in tightly woven patterns.

i printed the makerbot model upside down to minimize support material. The print resolution on normal wasn't great and it ended up 'hairy' from dragged remnants of pla, combined with the excess material from the supports which was too much of a hassle to remove from over openings and the center of the dome was members were no longer articulated instead all merging together into a disk... but the structure overall was stable and precise (despite concerns of printing an stl composed of multiple closed meshes - since for some reason neither rhino or grasshopper would allow me to union the volumes into a single closed mesh).

the zcorp models were already destroyed by the time I had reached the machine. Each time they were handled they would instantly disintegrate - not that this was surprising, however the definition of the members was much more refined especially near the center of the dome. It would likely require a few more tests at various scales to find an optimum resolution for both printers but I'm more interested in figuring out the pattern definition first before pursuing this further.

04 > 3d scanning
The handheld Sense 3D scanner was not complying with me at all when I tried to scan a random dinosaur toy. I  picked stegosaurus for his high depth in surface texture and its vivid colours, and despite placing it against a white background for consistent contrast and varying the location distance, it would consistently lose tracking often unable to continue model improvement, requiring a restart of the entire process and loss of progress. Frustrated i switched over to a different method, employing Autodesk's 123D catch phone application to compose 3d models from a series of static 2d images captured from every angle that it tracks for you, identifying areas requiring additional detail.

This process was able to produce significantly better results in composing a full model and though the quality was not great and alot of detail was lost, it had the advantage of being free, not requiring any additional software and conveniently accessed anywhere you would have a phone. It provides the files to download as .stl .mtl or .obj. So I went home and tested scanning my dog while he slept on the sofa. If you know huskies, you can image the struggle of keeping him still while I capture 20+ photos.