Week 3: Parametric Lampshade

This week we made 3D Printed Light Fixtures with geometry generated in Rhino / Grasshopper.

• Rhino 6
• Grasshopper
• Ultimaker Cura 4.13.1
• Silk PLA
• LED ring light

I really struggled this week without the step by step tutorial. I realized how heavily I rely on those tutorials to get through the assignments, and it exposed that there is still some basic modeling in Rhino that I don't know how to navigate. My initial idea was to use a small LED solar light that I have put dichroic film on in the hopes that it would cast light gradients on some complex geometry (as seen in the first two images). I began trying to translate Professor Jacobs' Rhino examples from a spherical form into a more rectangular form that would compliment the shape of my light. I made an STL of the rough outer dimensions of my existing light (didn't have digital calipers) to try to match the geometry in Rhino.

For some reason the "lamp3.3dm" file would not open for me, so i was using the "lamp2.3dm" file as a starting point. I was having issues seeing real time feedback with the sliders, so I had to bake the "mesh" output every time I wanted to preview a change in the geometry. I later learned from Mert that I didn't have the "draw shaded preview geometry" button clicked on in Grasshopper. I tweaked all of the slider parameters to try to see how they each effected the model. The "nu" variable was very useful, and when I would lower it down to 4 I would start to get square ziggurat geometry or 8 would produce octagonal forms that could mimic the rectangular form of light. I was never fully happy with any of these initial models, but I was also feeling the pressure to finish a model so I would have enough time to print. I went through all of the examples in the Class Drive folder to see if I could get results that I was pleased with. I went on some tangents with a 3D Lithophane generator, manipulating the STLs in Blender (where I am more comfortable), and exploring the parabolic mesh.

I was struggling with modeling my found light accurately in Rhino and combining it with my form through Boolean Difference, so I skipped that step to prioritize printing. If I have extra time, I will go back and attempt to model a test fit as a proof of concept.

The final image is the geometry I exported to attempt to slice and print.

Eventually, I made a form that seemed interesting enough to attempt to print. I began by using the "lamp_mesh_surface_example.gh" example. I was happy enough with the form, and while it was pretty minimal, I'm hoping the Silk PLA will look really good with this minimal form. I wanted to make the top hole as small as possible so you have to peek in to see the light source, but otherwise it is just an ambient glow.

I went through a lot of slicing issues and combinations in Cura. At first, I had exported the geometry from Rhino without properly performing the boolean union, so it would have failed if I attempted to print. I was eventually able to boolean union the interior and exterior poly surfaces, but for some reason it didn't cap the base of the two shells. I was able to get around this in Cura by sinking the model into the bed slightly before slicing. I attempted to slice in vase mode to save time, but each result looked like it would fail because of the steep overhangs.

I was eventually able to get gcode that seems like it will print. I'm printing with "tree" supports for the first time to support the top overhangs. The estimated print time is 17 hours 24 minutes.

The print took forever, but I think it came out pretty well. I was nervous leaving it printing overnight in my shared studio space, so I paused the print 11 hours in. You can see a distinct line at this layer. I was afraid the print was going to fail when I resumed printing, but after a few unpredictable layers it started printing reliably again. There must have been some slight expansion overnight as the temperatures dropped. In the future, I need to just trust the machine, and let it go overnight.

I was really happy with the surface finish of the Silk PLA. It really shows all the details, for good and bad. It highlights any imperfections, but it also shows the complexity of the infill pattern. There are areas where the print gets slightly lacy, but the patterning of these glitches are aesthetically interesting. I wasn't anticipating how prominent the infill patterning would be when lit from within, so I would plan that more consciously in the future. There was also a lot of stringing left from the support material on the interior, which makes the lamp more opaque at the top and ruins some of the glow of the shade. That said, I really loved the form of the "tree" supports, and think it is a pretty interesting form on its own.

Overall, I'm happy with this first test, and wish to continue working larger, but the long print times can be stressful.