Nyctous Acrylic Tunnel Book Fabrication

After selecting my design plans to create a nyctous acrylic tunnel book, I had about four weeks to complete the project. I planned and divided my time in those weeks as follows: 

Week 1 – Test Plexiglass Laser Cutting and Shapes
Week 2 – Design and Fabricate Shapes
Week 3 – Wire and Program Lighting
Week 4 – Make Stars and Finishing Touches

I would first complete a test cut of the acrylic panels to test their overall appearance, capability for detail, most effective arrangement, and other details of the process and look. Following the test cut, I designed and fabricated the final shapes for my acrylic layers with most complications and unknowns being smoothed out by the  test week. As the dynamic lights are a key part of my project, I would then dedicate a week to arranging LEDs, wiring the circuitry, programming the control, and refining the timings for an optimal rendering of the aesthetic. The final week at the end is left to make the background star panel and any other finishing touches necessary to bring the project together. This leeway was important to my schedule for any unforeseen circumstances that came up, as well as just providing time to make my result feel polished and finalized.

This test was intentionally rough and varied to gather information for a final cut while not spending overmuch time on details I may change. I put gathered a lot of information from this step, and the final acrylic panes would not have come out as good without first performing this step. My concept sketch for the initial cut layout is shown below.

Layout sketch for the test cut

The two components to complete a cut are a raster engraving of the scene element silhouettes and a vector cut to separate the panes and associated hardware from the sheet. With the raster engraving I composed three scenes, described as follows in their final arrangement from front to back. The first composes a lamppost over a road with blades of grass flanking either side. I primarily wanted to test the laser raster’s ability for detail in this pane, intentionally including skinny grass and lamppost features. The lamppost was also a feature I imagined putting an LED behind in my final cut, so included it here as a test for that. The second scene has two buildings with window cutouts on one side and a tree on the other, also ideas for final features being tested here. The window cutouts were especially an unknown for me, as I wanted the silhouettes to appear as buildings while not being completely unrealistic silhouettes. The third scene contains mountains and clouds, which I was most unsure of, but felt like a default background to use given my time in Colorado. In drawing the clouds I found myself including interior lines to emphasize their shape not possible with a strict silhouette, so I decided to try two different intensities of rastering on the clouds.

For the vector cuts, the sizing of the panes was a primary consideration, but also the connection to the hardware holding them together. I initially considered creating an enclosure to house the acrylic panes into. While sheet metal or concrete would likely fit this aesthetic, my default thought of making a wood box would likely not. Thus, I came to the idea of instead having the panes not enclosed but supported by minimal acrylic rails beneath the panes. The lower right section of this layout thus contains a large variety of rails of different thicknesses and different spacings between panes. To connect everything together, I planned for mating notches in the panes and rails, with a variety of tolerances to hopefully find a proper fit. If the panes are too strongly illuminated by the surrounding environment or the focus of an observer is not to look through the front, I can add a black paper around the panels.

I used CorelDraw to put this sketch into a form accepted by the laser cutters on campus, with cobbled together images for the raster.

Test cut layout in CorelDraw

And then I laser cut it, where numerous problems arose, most apparent being the warp of the entire acrylic sheet.

In progress test laser cut with visible warp

This is likely due to the larger regions of engraving and cutting on the right side and the protective paper still being on the acrylic. In some laser cutting processes it is more ideal to leave the protective paper on to prevent acrylic smoke from depositing onto the transparent sections of the acrylic, but I should have removed it here. The increased heat from these elements warmed the acrylic up so that it warped and led to several passes to cut fully through.

Another issue was the inconsistent engraving in some places, perhaps partially due to the warping, so I will perform the raster engraving in multiple lower power passes for the final cut. The differing cloud intensity looks equally bad to me, being relatively even but not a good mix of translucency.

Test cut arranged both closer together and father apart

For the arrangement of the panes I found I preferred them much more in the father apart spacing. I also decided not to change the size of the panes; I was expecting the 5.5 inch by 8.5 inch size to feel small on the test cut, but I found I didn’t mind the overall, felt it fit the amount of detail I was trying to fit into each pane, and would have to reconceptualize the content quantity of each scene at a larger size.

Taking all of these points into consideration, I ended up with the following design to cut my final panes.

The process for wiring the LEDs worked out without too many roadblocks. As planned, I used an Arduino microcontroller and breadboard that I already had on hand, I also used jumper wires with end connectors to not require soldering and enable more rapid iteration on my design. In order to get varying intensity and asynchronous timing across all the LEDs in my project, I ended up using every output pin available on the Arduino board.

Arduino and breadboard wiring layout

I wrote all of the LED timing code in Arduino myself, and an early test of the LEDs appeared far too synchronized, every LED turning on or off at the exact same time as every other. My final code solves this problem with a separate timing variables for each phase of every LED’s pattern. The entire program loops through every fifth of a second, and whenever a timing variable reaches a preset threshold, it switches to the next phase of that LED’s pattern. I set the threshold values at all different prime numbers so that the LED patterns very rarely sync up. I also used rgb LEDs within my project, so those phases shift the color every cycle by set increments within that phase. The abstract shapes in the back of the tunnel book are the most apparent use of this, though the streetlamp in the foreground also shifts slowly from white to slightly yellow light and back, Try examining the patterns of the LEDs more closely in this video:

The final week of this project did have to be condensed into the delayed time I had left, but I did manage to create a background of stars as shown below.

Illuminated rear pane of the tunnel book with holes in black paper acting as stars

For this final pane, I laser cut another acrylic panel with raster etching across the full surface. I then traced circles onto a rendered image of the night sky in CorelDraw and cut that pattern into a piece of black construction paper on the laser cutter.

Overall, the manufacturing process went largely as expected. The initial test cut for panes was very helpful for finding problems before the final design, and laser cutting proved very effective in creating a variety of elements from the raster etched panes, supporting rails, and cutout of stars. Some additional insights cropped up in the aesthetics of the piece, but the manufacturing turned out exactly as I originally planned.