LED Vase

For my final project for MAT 238 I created a glowing vase with a plastic structure, acrylic pieces, and LED lights.

Material: White PLA filament, 1/8" acrylic sheet. Hardware: 16x16 neopixel grid, Arduino Uno, breadboard, wires. Software: Fusion 360, Rhino+Grasshopper, Prusa slicer. Fabrication: 3D printer, laser cutter.

I was inspired by maketvee's Neopixel Vase where they 3D printed a vase which slots into an LED ring and uses lines of clear PETG filament to guide the light around it. I wanted to do something that also guided light along a path over a structure but had more complicated patterns. Initially I thought about using clear filament and printing tree like structures, but after concerns that the layers would be cloudy and wouldn't carry light well, I switch to clear acrylic, leading me to a flat sided structure.

Using measurements of the LED grid I already owned, I tried some arrangements and decided that it would be better to have two pixels per acrylic piece, leaving five pieces on each side for a total of twenty. I sketched out how the base could be shaped so it would slot into the grid and wouldn't slide around.

At first I tried to create visually complex generative patterns using tutorials in Rhino Grasshopper. My intention was to extract lines from these complex patterns to use as outlines of the acrylic pieces. However I found this very challenging as I didn't have much control over the lines' qualities and struggled to make useable shapes out of them. In my peer review from Devon, they encouraged me to make new patterns using primitives and boolean operations as it would ultimately be easier, despite this approach being from scratch.

Following the peer review with Devon, I worked on creating patterns that I could use for the acrylic strips in Rhino Grasshopper. The patterns were made in python nodes, but I used grasshopper nodes to create guiding outlines for the size of each face and where the acrylic pieces need to fit into. I did some minor manual adjustments with what was created in Grasshopper to make sure it was ready to fabricate, such as removing tiny jagged pieces from the blocky pattern.

I did two main tests to make sure my vase would work. The first test was to see if the shape I had in mind would slot nicely into the neopixel grid. The second test was a tolerance test for the acrylic fitting into indents in the PLA. The slot test worked pretty well, though I realized later the fact that it was a flexible piece helped a lot, and the tolerance test indicated 0.25mm would leave a loose but acceptable fit. The latter test was delayed due to the fact that the print failed many, many times.

In Fusion 360, I first created the grid of pixels to allow to me visually check everything lined up. I then extruded the basic wall shape. Next up was indents for the base of each clear piece where the pixels would line up with the base of the acrylic- while each piece was a different shape, the slot where they touched the pixels was consistent. Lastly was taking the dxf files of the patterns from fusion and extruding them into the walls. Everything was parameterized to adjust later, though I did struggle with the way operations in sketches weren't saved in the timeline and didn't response to later parameter changes.

I also used Fusion to take the patterns and either use the move or offset tool (depending on the pattern) to create the 0.25mm tolerance and export a file for laser cutting.

My neopixel grid was controlled with an Arduino Uno and a simple breadboard setup. The code runs a rainbow pattern through only the pixels exposed through the acrylic.

The laser cut pieces were quick to fabricate and had no issues. Despite having had many problems 3D printing earlier pieces, the final (very long) print was successful on the first try. There was an axis shift towards the end of printing which resulted in a line across the print on two sides, but luckily it did not effect the fit of the pieces. Interestingly, while the press fit test indicated that a 0.25mm tolerance would result in a loose fit, the pieces in the final print had a very tight fit and were a little hard to get in. The artifacts from the supports on two sides may have contributed to this, as well as the fact that the test was printed with a different orientation than the final one.

I am pleased with the end effect. The fit on top of the neopixels isn't perfect and it was hard to get some of the acrylic pieces in, but the overall shape worked out. My favorite patterns are the wavy spikes (poking above the top of the plastic structure makes the silhouette interesting) and the hole-punch (the tiny laser cut rectangles diffuse a lot of light). I find the glowing pieces a very interesting challenge and seeing how these patterns worked out, I have a lot of ideas for future projects involving LEDs and clear structures.