project 4: lighting engines

Sarah Xi
22 min readNov 8, 2019

Light is highly integrated in our lives and we are of constant need of it. Without it, we cannot do even the simplest tasks, in other words, we as both humankind and as a society depend on light. Whether or not its too much, light effects not only our efficiency/productivity, but also our mood. In this project, we are to explore how we interact light and create a lighting engine that supports the given task.

11/5/19 research + explorations

For this project, we are to create ‘lighting engines’ for our picked out prompt. I picked out “design a lighting engine that supports collaborating with peers in a conference room”. As such, I started by looking deeper into what we discussed in class and looked and various existing lights located throughout campus.

The Research:
In class, we discussed the different types of lights. I went online to look at the difference between the types of light and types of lighting.

Types of Lights:

  • Incandescent: typical bulb, bulb has filament, can overheat and burn out.
  • Light Emitting Diode (LED): is becoming increasingly common, a semiconductor, energy efficient, won’t overheat, and consists of multiple diodes to produce light.
  • Halogen Lamp: improved version of incandescent bulbs, inert gas within the bulb increases brightness, has a longer life span.
  • Compact Fluorescent Lamp (CFL): meant to be a replacement for incandescent bulbs, has a longer lifespan, higher luminous efficiency.

Types of Lighting:

  • Ambient: Atmospheric lighting,
  • Cove: Museum lighting,
  • Work/Task: Spotlights, i.e. Orthodontist lights
  • Accent: Decorative lighting
  • Intermittent: Light that shines in pattern of time: i.e lighthouse.
  • Navigation: Source of illumination for a vessel (i.e. planes, boats) to show its position and status.
(above) line of incandescent lights near the University Center — picture 1 + 2
(left) incandescent bulb at Entropy; (right) LED light at a staircase near Entropy — picture 3 + 4
(left) hanging light in the UC; (right) exit sign with LED — picture 5 + 6


  • Each light has its functions, but you can use a specific light for a different function.
  • Pictures 1, 2, 3: many incandescent lights were strung up and used with both a decorative and ambient lighting function in mind.
  • Picture 3: Looks like cove lighting, but not quite, helps light up the entire wall space. LED strip.
  • Picture 5: Spotlight, many of these in a room helps light up the whole space.
  • Picture 6: Exit sign isn’t exactly a ‘light’ but it is lit by presumably an LED. It helps the sign get noticed in case of an emergency.
(all above) lights seen at the Pittsburgh Vintage Mixer Fair

During the weekends, I decided to go to the annual Pittsburgh Vintage Mixer Fair at Nova Place. There, I found many vendors that sold different knickknacks, clothing, and vintage items like cameras. In addition to that, I also found numerous sellers that displayed lamps/lights for sale.


  • Very few lights on display were plain white, most were (pretty saturated) colored.
  • Either the lampshade is patterned, three dimensional or the stand is decorated (i.e. Buddha sculpture at the bottom, plane cylindrical lampshade on top).
  • Some have two dimensional pattens: light shines through in interesting ways (i.e. gridded).
  • Three dimensional lights: shows relationship between shadow and light, light illuminates and reflects off varied surfaces.

After looking at the different types of light, I then brainstormed things I should consider for my prompt, which is to design a light that supports collaboration in a conference room.

Things to Consider:

  • How can I make the light illuminate a whole room?
  • Make sure that the lighting engine isn’t distracting, but also not invisible: has to complement the context.
  • Color temperature: cooler lighting, more formal, won’t make people sleepy.
  • What types of paper can support a big lighting engine?
  • What kind of adhesives that won’t show on paper when it’s lit?
  • On/off state of lighting engine: equally as enjoyable.


In brainstorming for this project, I also looked at a few artists and their artwork for inspiration and potential directions.

Mark Rothko:

  • Became famous for his colorful rectangle/square forms that was intended to depict deep emotions.
  • Look at the relationships between tones and color temperature.
  • What elements work well with each other?
(left) Light Cloud, Dark Cloud 1957 ; (right) Orange and Yellow 1956


James Turrell:

  • Likes to work with light, and the presence of space.
  • How can I use the way light reflects/refracts to portray something?
  • How does different levels of illumination affect how the audience views it?


Naum Gabo:

  • An artist known for his sculptures of glass, plastic, and metal.
  • Became a leading component of constructivism
  • Style of art is abstract, he aimed to combine sculptural element with the architectural element into one unit.
  • Flow of the form created by the nylon string portrays a sense of calmness, clean.
  • Organic shapes.
(left) Linear Construction No. 2 1971; (right) Spiral Theme 1941

Richard Sweeney:

  • Creates sculptures that are inspired by organic, natural structures.
  • Likes to take inspiration from growth patterns in nature.
  • Repetitive forms are constructed in an organic way.
(left) Untitled; (right) Seed II 2012

11/1–11/12/19 exploration of paper, adhesives, and sketch models

After looking into the types of light there are and is around us, I then explored a bit with the types of paper I already have.

(left) bristol paper; (right) palette paper
(left) tracing paper; (right) tracing paper crinkled
(above) printer paper

— Sturdy, good for structural support.
— Thick, light diffused a bit too much.
— Smooth surface and thickness allows for effective embossing and scoring.

Palette Paper:
— Waxy.
— Ultra smooth on one side (wax part), smooth but has a bit of grit from paper texture on the other side.
— Diffuses the light less than the Bristol paper due to its thinner nature.

Tracing Paper:
— Extremely thin, almost translucent-like.
— Diffuses the light, but is the most bright.
— Good for hiding the bulb.

After testing out the paper I already have, I then went to the art store and took a look at other potential papers I could use for this project.

(left) paper swatches; (middle) hot press watercolor paper; (right) silkscreen paper
(left) white and warm white stonehenge paper; (middle) 4 ply bristol paper; (right) 2 ply bristol paper

The types of paper I took interest in were mainly ones that had a cooler tone as warmer tones aid sleep and relaxation more than productivity.

(left) 4 types of adhesives on printer paper; (right) same 4 types of adhesives on tracing paper.

As for my adhesive experiments, I tested Elmer’s Glue, UHU Gluestick, double sided tape, and matte Scotch tape. I swatched all four on a thicker sheet of paper (printer paper) to see how obvious it showed on the light. I also swatched the adhesives on tracing paper as it was more translucent. What I found from that was that liquid glue like Elmer’s tend to be more obvious than both regular and double-sided tape.

I also looked at how distance could affect how much the light diffused. The papers I experimented with were the ones from the paper swatch: Bristol, Canson, Folia in white and pearl white, construction, tracing, and printer paper.

(left) bristol; (middle) canson; (right) tracing
(left) Folia white cardstock; (right) Folia pearl white cardstock
(left) construction paper; (right) printer paper

What I noticed after doing this, was that the thicker the paper, the easier it was to cover/diffuse the light emitted from the bulb. With this in mind, I got an idea for what kinds of paper I would opt for in the final.

11/11/19 sketch models:

Last thing before starting making iterations, I set out to create a few more sketch models to experiment and explore different possibilities in approaching the project.

Sketch 1:

(above) sketch model #1, with light on
(above) sketch #1, with light off

I already knew I was really interested in how light and shadow could cast interesting patterns through the use of layers. As such, for my first sketch model, I experimented with layering by adding different cuts of paper under the outer most cylindrical form. I also wanted to look into the concept of hiding and reappearing with the on and off stages of the sketch model. What I found from that is that it gives a good element of surprise, however, the off stage would lack interest (which is equally as important).

Sketch 2:

(above) sketch model #2
(left) OFF state; (right) back view ON state

I wanted to continue exploring layers, except in this sketch model, focusing more on the exterior and not hidden layers like the first sketch model. Here, I cut a piece of tabloid paper into four strips and cross hatched it over itself to cover the light. This gave me insight as to how the shadows created are similar/different to when the layers were hidden behind a piece of paper. The off state is definitely more interesting here than the first sketch, however, it loses a bit of the element of interest as it’s more clear where the shadows would appear.

Sketch 3:

(above) sketch model 3
(left) side view; (right) OFF state

For this sketch model, I decided to combine both elements of whats hidden and what’s presented. I cut a tabloid sheet in half and overlapped it with each other, while also having another strip inserted in between the two overlapped pieces. What I thought was interesting about this sketch was that in the off state, it is not clear as to how the light of the strip would shine through. However, in the on stage, the slightly curled inwards strip creates a light gradient due to the paper tapering off away from the exterior sheet. This then reminded me of how distance affects how clear the layer of a paper can be seen.

Sketch 4:

(above) ON state
(above) corners and inserts

For this sketch model, I was interested in exploring ways to cut and connect pieces of paper. As such, I cut slits and inserted paper strips through them to connect and hold the form. This ended up being an experimentation of the potential use of tabs in my final.

Sketch 5:

For this sketch, I curled and ‘stacked’ a sheet of paper in another. Craft was a bit harder here as it was tough to have glue stick and hold rounded edges and tape was hard to place neatly. I also wasn’t sure what to do with the corners (looked awkward and unintentional) so I ended up folding the corner.

Sketch 6:

Since the last few sketches mainly focused on layering, I was then inspired to explore pattern. For this sketch, I created varying sized (in height) triangular prisms and stacked them. It was interesting to see how light peeked through the open corners and illuminated the top and bottom of the paper. One thing I did notice though, were how the tape and paper tab on the corners could be easily seen when the light is on.

Overall Reflection:

  • Tape was actually very easy to use and hide.
  • Printer paper is not translucent, but is still quite flimsy.
  • Consider ways to utilize negative positive space with paper, how does that affect on/off visuals?


  • Sketch Model 1+3 were interesting.
  • Sketch Model 1 — ON stage is significantly more enjoyable to look at than its OFF state, looks incomplete.

11/13/19 drafting models:

In creating my first draft, I tried to work with cylindrical forms as I thought it’d be easier to make big and fit the conference room atmosphere: formal, simple.

Draft 1:

(left) sketch of first idea; (right) a tab idea to hold the form
(above) trying out the tab method to hold the cylindrical shape

However, as it’s better to work bigger for my prompt, I found that creating one large sized form was difficult. Not only would I have to consider the appropriate size, I’d also have to watch out for stability while maintaining visual appeal. Moreover, with a curved, rounded shape like a cylinder, it’s extremely hard to maintain good craftsmanship.

(left) side view; (right) failed attempt at lining the bottom with tracing paper


  • Cylindrical form is too hard to keep symmetrical, stable, and well-crafted.
  • Hard to cut out circles well (craftsmanship concern).
  • Bristol paper was too thick, hard to illuminate a conference room.

Draft 2:

This was a quick draft, but when I was looking at typical conference room lights online, I noticed that there were many that were long, bar shaped lights. I then attempted to create a lighting engine inspired by said shape.

(left) front view; (right) side view

Using a sheet of clear print vellum, I overlapped the sheet into a cylinder and folded it to create a top crease. Then, to help diffuse the light at the base of the socket (as the vellum was too close to it, too bright), I cut out strips of Bristol paper and pasted them at the top.


  • Bristol paper is too thick, doesn’t stick well against curved edges (didn’t conform to the shape).
  • The clear vellum used is equally as translucent as the tracing paper used in the beginning of the project, but slightly thicker/sturdier.
  • Light is diffused well in the front and back side, not so much on the sides due to the naked bulb.

Draft 3:

Moving on from the previous drafts, I then thought about ways I could create a large sized piece without having it droop or be too chunky. As such, I decided that instead of making fewer large pieces, many little parts creating a bigger whole would be better.

(left) idea sketch; (middle+right) bristol template for easy replication of the triangular prism net

For this draft, I created 6 triangular prisms using tracing paper and combining them. For faster and efficient replication of each prism, I first created the triangular prism net using Bristol paper and then using that to trace over the tracing paper. After creating all the prisms, I then arranged them in a way that there would be a cubic base in the center and four triangles wedged on the side. In addition to that, I also varied the heights of the center and the sides to aid enjoyability in both ON and OFF states.

(left) I used a ruler to help make precise creases when folding the prism nets; (right) the triangular prisms
(left) side view; (right) for the center base, I had pasted half half of tracing paper and printer paper to see how bright the light would shine through them.

In creating this draft, I came across numerous issues that I did not anticipate:

  • UHU glue stick marks can be easily seen through tracing paper, in both ON/OFF states.
  • Tracing paper is equally as fragile/unstable no matter how many pieces of it you group together. Adding more together does not make it stronger as a whole.

This comes to show how attention to craft not only applies to the size I’m working in, but more importantly the type of paper and how it behaves.

(above) what the light looks like hung up

In-Class Critique:

  • Different levels are intriguing.
  • However, it loses interest on the sides as its the only side that doesn’t show obvious varied heights.
  • What should a conference light do/have?
  • Lighting engine should not distract people in conference room.
  • At the same time, it should not be invisible, it should complement the task (in this case, collaboration in a conference).
  • Because of the numerous layers of tracing paper, the light is too blocked, which makes it hard to illuminate a conference room.
  • Scale up the lighting engine! Too small.
  • Consider ways to bounce light: baffling to reflect and illuminate stronger.

11/18/19 creating alternatives:

In class, Steve and Stacie discussed the difference between alternatives and variations. Alternatives being completely different ideas/approaches and variations being different versions/alterations of that same idea. Because I felt the need to explore more, I then brainstormed for another alternative.

Before starting on the fouth draft, I visited Home Depot to find light bulb alternatives, as a cooler toned bulb is more appropriate for productivity in a conference room. Especially with bigger spaces, proper illumination is another significant factor to consider.

(above) other types of light I encountered at Home Depot
(above) new bulb!

The bulb I ended up purchasing is a 75 watt daylight, 1100 lumens bulb. For comparison, the previous bulb was warm light, 400 lumens.

(above) design faculty conference room

I also ended up visiting the design faculty conference room for a point of reference. I’ve always envisioned a conference at a rectangular table for my lighting engine after all.

Draft 4:

(left) brainstorming different ways to approach the ‘pattern’ repetitive design; (middle) initial sketch of draft; (right) ancient Chinese coins

I was initially inspired by ancient Chinese coins, by how there is usually a square within the circular shape. On top of how I wanted to experiment with ‘pattern’s, or rather, using smaller parts to create a bigger whole, I then came up with the idea of having squares layer within itself. This not only makes the interior and the OFF state interesting, it also allows for more opportunity to add elements on the exterior to tackle the ‘enjoyability’ aspect of the project. As such, while I layered squares within itself, I also tried to decorate the exterior by using Yupo paper and clear print Vellum paper through layering.

(left) Yupo paper; (right) seeing how Yupo creases vs taping strips together
(left) layering clearprint Vellum to Yupo, layering swatch; (right) layers under light

For this idea, the lighting engine would be created in three main components. One, being the exterior shell made of 4 rectangular panels with Vellum layered on it. Two, being a smaller open square box shape within the exterior and three, being the same but even smaller and with tracing paper lined at the bottom to hide the bulb.

(above) process of making one of four panels for exterior structure
(above) front and back of panel

Originally, I was going to create the exterior shell by cutting out a long strip of paper and folding it for convenience and efficiency. However, because I needed to work in bigger sizes, I didn’t have paper long enough to do so. I then opted for the join four separate panels route, which ended up making it easier for me to create sharper and more precise corners.

(left) exterior shell; (right) second layer

After making the first layer, I realized that Yupo was way too flimsy to support itself—especially with this size. As such, I made my interior supports with Bristol, however, not till later realizing how its thickness would impede on how illuminated the engine would be.

To fix this, I then ended up cutting windows out to let out the light trapped by the Bristol in the center.

(left) cut panels; (right) all three layers done

In building and connecting every piece together, I was originally going to use the UHU glue sticks as they worked well when I did the layering swatch. However, I noticed that the paper liked to curl when it was much larger and longer, prompting me to switch to only tape.

(left) example of paper curling after attempting to use UHU glue sticks; (right) used double sided and scotch tape to build the whole draft
(left) side view; (right) under view
(above) group critique exercise from class


  • Craftsmanship, sides are warping.
  • Very industrial, architectural looking.
  • Work on OFF mode: looks like white rectangle
  • Does square shape support any kind of conference room?
  • Looking from underneath: kind of distracting, can still see the bulb and too many windows/holes.

What to do next time:

  • Careful of craft! Seal the sides well, avoid warping.
  • Play around with layers, how can it complement the shadows?
  • Vellum to cover the bulb.

11/20/19 creating variations:

Based off the critiques given last class, I’d gotten a couple of ideas of what to change and improve on.

To start, I looked at different ways I could layer in the exterior layer. This is mostly because last draft, I intended the light to only emphasize the layers on the exterior, but because of the windows inside, there were new shadows created that I didn’t anticipate. Moreover, because the layers and shadows didn’t seem to be intentionally paired together, I then sought out to find a way to have the layers complement the shadows casted from within.
I ultimately decided on a layering pattern that was symmetrical as I believed that it would match with the symmetrical nature of the shadow.

(left) different ways to layer; (middle+right) draft brainstorming/plans

In further exploring the layers, I also tested out different types of paper: mainly Bristol and printer paper. This was because during the last critique, some mentioned how the lighting engine looked like a white box in its OFF state and that the layers were hard to see. As such, I then tried four different alternatives to this issue. On each side, I layered a different combination between the two paper: Bristol on Bristol, printer paper on printer paper, Bristol on printer, printer on Bristol.

(left) Bristol on printer paper; (right) printer paper on Bristol
(left) printer on printer paper; (right) Bristol on Bristol

From this experimentation, I discovered a solution to the warpy edges issue. By layering Bristol on top of the thinner, flimsy Yupo, I can prevent warpage and add stability. In addition, it further adds stability by using thicker Scotch tape to join the panels together.

(left) sturdier Bristol; (middle) flimsy printer paper; (right) how the layers were taped on each panel
(above) joining panels together
(above) how I attached each layer to each other with Scotch tape.
(left + middle) placing all three parts and taping done; (right) adding clearprint Vellum to paper to hide the lightbulb

What worked and what didn’t:

  • Bristol layered on the exterior walls helped with stability
  • Overestimated the size of the interior, sides are still warpy.
  • Vellum paper effectively covered the bulb and is still bright.
  • White light was strong enough to illuminate through all the layers.

11/24–11/25/19 fine-tuning:

With the deadline approaching fast, I felt like I was in a pretty decent place with my design. As such, for my final lighting engine, I decided to go with the same design and focus on craft.

(above) creating the layers for exterior
(above) to prevent underestimating how long each panel is, I made each layer longer so that I could cut the excess off.
(left) cutting out windows from the Bristol; (right) taping clearprint Vellum

Even though I had only discovered Yupo paper recently, I ultimately decided to use it for my refinement iterations as it not only has the translucency of tracing paper, it also is significantly more sturdier.
In addition, something that came up during a discussion with Stacie last class was that I could try layering Bristol on both the front and the backside.

(left) Bristol layered on front and back; (right) What it looks like with light

Although the idea was interesting to experiment with, I found that the ON state was more visible than it were in the OFF state. Initially, I thought that it could work, as the layer on the back was very visible by itself (a single panel). However, as I mounted it onto the main structure, the layer was almost invisible.

(above) foam core fitted to the structure

For this iteration, I initially tried to fit the top with a board of foam core for stability and to hang the light. Although it did work, it felt too constrained and covered than what it was before where the foam core only took up the centermost section/square.

(left) foam core fitted to inner square; (right) foam core fitted to second layer/square

In doing this, you wouldn’t see the sides of the foam core peeking out if the foam extended to the second square versus the innermost.


(left) light off in context zoomed in; (right) light on on black background
(above) detail shots
(above) light on in context


Out of all the previous projects, I personally thought that this was the most enjoyable. This is most likely due to the fact that cardboard was significantly harder to cut and that I’ve never worked with ‘light’ before, so learning how to manipulate light towards a certain goal was intriguing to me. Overall, this project has not only made me realize that there are some things that you cannot visualize and have to make it to see it, but also to know how to decide what you want to spend your time on. This is mainly because this project was much more rushed than the ones in the past, and similar to the cardboard project, I learned to effectively spend time on ideas worth spending on.



Sarah Xi

Hey! I’m currently studying design @CMU with a focus on communications design + minor in HCI. You can find some of the projects I’ve worked on here.