First step to a functioning lamp: a light that turns on. I had my boss at Stemtap (the company I was formerly interning for and am now straight-up volunteering for) give me a crash-course in electronics.
He drew out circuits in technical terms, taught me the difference between parallel and series circuits, helped me setup a bread-board, and showed me how to test the current.
I took the technical drawings and the practical lesson and turned them into drawings my brain could understand, showing actual physical objects. When he confirmed that I was right, I got rid of the bread-board and built my own parallel circuit from scratch, using a latch-button switch to turn on three strips of LEDs. (I was pretty proud.)
Turning back to my mechanism, and how to run power up between the ply, I tested some more glue options. So far, all of my glues had been brittle and rigid when dry. I tested three new ones and silicone finally gave me the flexibility I needed for my design.
Having come so far with my flexi-lamp concept, I woke up at 4am one Sunday morning with the sudden urge to sketch a new design for my original concept. I had finally found a way to merge the elegance of my material test model with the function of my first cardboard prototype.
I really liked this idea and thought it would be achievable, so I decided to shelve it as a backup option for after I’d confirmed that my flexi-lamp wasn’t going to work.
I made a slab of home-made terrazzo to use in prototyping and avoid wasting my limited material. I gathered gravel, mixed it into cement and poured it into an A3 paper box with the same depth as my final material. I was really pleased with how it came out in the water-jet cutter, so at least I know that if a terrible mistake happens with my final piece, I can use a home-made batch.
I cut a block out of the full slab, turned it on its side and sliced it into a thinner strip. Finally, I did a third pass to cut a semi-circle shape out of it. When working with a pre-cast slab, this is the only way to alter the thickness of the piece, and it’s limited to 100mm width.
Here, I’ve made slidable weights for testing.
Before glueing or bending my central stem, I applied copper tape and tested it to make sure it worked.
I created a makeshift base using stacked MDF and inserted three springs around my button. In the future, I’ll use more springs, because this setup allowed the user to press the pad on the edge and make it skew unevenly.
I tried both birch and bamboo stems and started experimenting with them in formation. The birch was much stronger and held upright, though when strands were pulled out unevenly, they kept their kink permanently. The bamboo was much more flexible. It was difficult to get it to stand upright and when attempting to pull out individual strands, they were liable to break quite easily.
To give the bamboo some extra strength and stability, I tried making it 2-ply. I used silicone on the sections I wanted to stay flexible and aquadhere or epoxy on the sections I wanted to set in a rigid shape. This mixing of glues worked to an extent; the rigid and flexible sections were distinct. However, I couldn’t create bends nearly as tight as those capable with wet birch, so my goals for the final form would never be fulfilled with bamboo.
To twist the stem around a 90 degree angle, I wet a small section, set spacers evenly apart to ensure the strands would remain equidistant, and twisted the outer rings by hand. I then held the shape over a heater for ten minutes to set it. I had to try this a few times to learn how far I needed to bend the wood in order for it to find my desired form when it ‘bounced back’ into shape.
To create the lamp-shade, I split the central strand down the middle, soaked the ends in hot water, forced the strands around a mould and held them in place with rubber bands. The 2-ply shaft was difficult to work. It bucked at one point. I dried the whole head with a hairdryer for 15 minutes until I was confident it would keep its shape. I was happy that it had.
I was impressed at how tightly the curved kept their shape. I’d anticipated them travelling outward, but the final form is a relatively high-fidelity copy of the mould. This inspires me to use a more defined mould in the future.
The MDF spacers absorbed water, expanded and deteriorated. I had to cut them off.
Once I’d freed up the stems to move independently, I was able to get some articulation going.
I had originally thought of creating layers of rings; one with leds, one with a reflector, one with a lens, one with diffusion, and one with a back cover. Seeing the tumultuous form I’d created, I realised that an equally scattered lighting system would suit best. Filling the swirling structure with a swarm of lights showed me the character I’d been searching for.
I tested the articulation with myself and others. It worked, but the interaction took learning. This is something I would need to improve on.
The weights didn’t turn out to be useful at all, rather, it was the friction that the rings created against the stem in different positions that allowed it to hold its shape in various states.
Although the testing of each of my electrical components had worked prior to assembly, I learned that something had gone awry during the bending process. I was only able to get half of the circuit working at the top of the lamp. To solve this in the future, I’ll need to reorganise my bending and laminating processes and perhaps experiment with a higher gauge of copper.
My interim submission was a model that demonstrated articulation, without complete electronics.
After making this model, I was able to determine some changes I’d need to make in my design.
An obvious problem is the size of the base. I’ll need the base to be much larger and heavier to withstand the force of the bending stems. This encourages me to make the head larger to match, and as a consequence make the whole lamp shorter, to compensate for material distribution. Currently, the range of articulation is almost double that which is required in the brief. I’m hoping a shorter lamp will retain the same quality of character as this version and become more suited to a desk environment.
The interaction needs to be simplified. Fortunately, I discovered that only two friction rings need to be present for maximum articulation range. This will allow me to dedicate more effort to perfecting those two rings.
I’d like to create collars that open up and allow easy-sliding when squeezed, then lock tight when disengaged. After brainstorming a complex solution with a peer, I thought of a much simpler solution I’d like to try, involving two parts and a single spring.
This has been my favourite part of the process. I love coming up with wild ideas, building models and learning new skills to make them function. Next, I have to create a refined product that looks good, functions well, is pleasurable to use and can be manufactured repeatedly. I think I may have fluked the aesthetic quality of this model and it will be a test to see if I can repeat it, not just once, but many times.
I believe this product would look and perform much better as a full-sized floor lamp, so I’m hoping that if I can make it work well enough at small scale, I’ll be able to repurpose the design in my own time.