MAT 238: Lamp Shade

For this assignment, we were tasked with creating a lamp shade that could replace an existing lamp shade in our house.

To start the process of designing my lamp shade, I first experimented with some curves that I could then loft to generate the outer surfaces of my lamp shade. I experimented with lofting between circles, squares, and octagons but finally settled on a set of curves that only involved octagons. These octagons were positioned and displaced so that smaller octagons corners were placed at the midpoints of the larges ones. This created a nice looking, rose-like base.

Once I created these curves, my next task was to offset them by a certain distance and create lofted surfaces between the curves. Once I was able to create a lofted surface, I them performed the mesh offset and mesh to ploysurface functions in Grasshopper. These functions allowed me to create a surface with a thin dimension as opposed to just a 2D surface.

After I was able to loft surfaces between different curves, I decided to experiment with randomizing the curve order. This allowed the lofted surfaces to change between smaller and larger octagons, and presented some interesting lamp shade ideas. However, after exploring this task and comparing them to an actual lamp shades, I decided against randomizing the curves them in my design. In my opinion, they seemed to look more like vases than lamps.

In the next step of my design process, I decided to explore the idea of creating holes in my lamp shade to make the design a bit more interesting. In order to do this, I took in points that were on the surfaces of my lofted curves and randomly sampled them. From these randomly sampled points, I created spheres that were placed randomly along the surfaces of my lampshade, and intersected them in some way. Using this technique, my final step was to take the BooleanDifference in Grasshopper between the original mesh lampshade mesh generated from the octagonal curves and the spheres. However, this final step proved to take an extremely long time, crashing my Rhino/Grasshopper software multiple times.

Before attempting to find an alternative path in generating holes, I decided to test my design using only a handful of spheres. To do this, I first manually created spheres in Rhino that touched the surface of my design and performed a boolean difference in Rhino. Unfortunately, the design didn't turn out as I had imagined so I decided to abandon the idea as a whole. Further, I concluded that the actual printing of the design would likely be very complicated and not turn out as expected

After all this experimentation, I finally decided to simply stick with my original design of just the largest to smallest octagonal curves, displaced by their midpoints. Once I settled on this design, my next step was to create the part where the lamp would be inserted.

To ensure my lamp shade design would fit into my lamp, I first measure the diameter (which was 2.1 inches). I used the revolve curve function to create a simple circular top to the lamp shade. As a test, I printed this circular shape to see if it would fit into my lamp, and sure enough, it didn't :). Thus, I made my curve a little smaller and once I reprinted it again, it fit perfectly into the inside of the lamp base.

Once I fitted the lamp base correctly, I decided to print my design in full. However, after 2 attempted prints, the lamp shade failed to get past the same point. I had thought it was an error because the design had moved, but it turned out to be a reproducible error that occurred when the printer had to print overhangs. Thus, I decided to include support structures to see if it would help the printer, and sure enough it was able to fully print. However, I noticed that at some heights, the printer seemed be pushing against the object, almost as if the gcode didn't raise the object high enough to make the layer correctly. Further, the support system seems to be tightly attached to the lamp shade, so I will have to figure out a way to remove it from the lamp shade design (perhaps someone could comment some tips! )The total print time with support system was 11 hours and 51 minutes.