W3: Lampshade

This week's assignment is designing a lampshade using Rhino and Grasshopper and manufacturing it using Ender 3 Pro. In my house, we have a ceiling light that we don't use regularly. I wanted to refresh its look in the hopes of using it more often. My initial idea was to print 4 variations of my design for each of the light bulbs, but due to time constraints and insufficient material, I will need to reconsider this goal.

Rhino, Grasshopper, 3D printer, and ~300gr of filament

I started designing the lampshade using its fitting component to the ceiling light. I measured the original cap with a digital caliper, recreated a section of it using curves in Rhino, and used the RevSrf component in Grasshopper to create the circular surface. I printed a test version (~40 minutes) and made sure that it fits perfectly.

The original lampshade has a brass ring around it. Since it's light enough and would go well together with my black PLA, I wanted to incorporate that into the design. I measured the ring with a caliper, modeled a curve and used the same method from the previous step. Then, I added a little holder for the ring, later to be connected with the cap, and the shade itself.

Then I started thinking about the main shade structure. The design goal was creating an interesting shape, and use pattern morphing on its surface to create holes (I have a black PLA, without the holes it would look too solid for a lampshade). For this, I tried the parametric mesh creation method using a modified version of Jennifer Jacob's sample code. This approach could create beautiful forms, yet I couldn't find a way to use the resulting mesh in SrfMorph component in Grasshopper. Is there a reliable way to convert meshes into surfaces?

I also tried creating a surface using parametric methods in another Python script, but I realized I need to know more about NURBS (knots and control vertices) before using it successfully for 3d-printing. I set it aside for future investigations.

In order to create the initial surface to be morphed on, I used Swp2 component in Grasshopper with two rail curves and one circular section curve. When setting two curves, I accidentally made one longer than the other, but I ended up liking this better than the symmetric version. I ended up creating a "folded" look for one side of the shade by rotating the ring holder by 7 degrees (around the origin). For the "holes", I experimented with different basic curves, replicated and extruded the curve to achieve a desirable and printable morph on the surface.

For manufacturing the lampshade, I tried various configurations and rotations of the object in Cura software. The object is 17x17x10.5cm and it is expected to take a long time. My hope was to print at the lowest quality (0.28 mm layer height) for time considerations. I soon realized the trade-off between thicker layer height and support overhang angles; recommended settings were set to 35 degrees for 0.28mm, and 45 for 0.2mm. Higher quality settings require less support but use more filament. The average time estimates with various settings in Cura were around 2 days and 3 hours. Once I placed the slanted bottom part flat on the bed I found one setup with less than 2 days print time (i.e. 1 day 19 hours 40 minutes).

Currently, I am printing the object. A small patch of support structures broke off, and the loose filament is filling that gap right now like a spider web. I hope it doesn't cause a problem with the overhang part and/or surrounding print area in the following layers.

Update #1: I couldn't leave the print without the support structures overnight. So I canceled the print 6 hours in, added a Raft underneath for better adhesion of small support pillars, and restarted. The new estimate with Raft is 1 day 22 hours. It's been 10 hours so far and the print doesn't show any problems.

Update #2: Finally done! Separating the final shape from the bed was a challenge; I had to use a spatula underneath the raft and pull from the top at the same time. I am glad that raft and the model itself didn't cause separation problems. Support structures were hard to pull off; I had to break them by applying force with a screwdriver.

Update #3: I also printed another symmetric version of my lampshade. It is installed now.