Mike's Spacetime

While the mechanics of the structure were still undetermined, we started to focus on the most important question – in fact our very first milestone: can we illuminate a panel of fabric, evenly, in bright and distinguishable colors, switchable by computer within a price and wattage range permissible by our budget ? We had set a hard limit on the date by which time had to have found a satisfactory solution or we could not go on.

Our first thoughts revolved around Halogen bulbs. Incancescent bulbs cannot be switched as much as we needed to switch them, and the possibility of using LEDs was quickly discarded due to being way out of our price range: We’d need either A LOT of them or special super-bright ones which were inherently expensive. We looked into LED panels or LED arrays but nothing we could find did what we needed for an acceptable price.

Halogen bulbs however are cheap. They can be switched relatively quickly. They have reasonably good life times. They’re bright. Theater gels are also cheap and come in all needed colors. Olli started researching and ordering various possible lights, from prefabricated, colored halogen bulbs to little, very bright and cheap stick light bulbs with custom made casings out of tin cans and gels attached to produce color.

Our first test at ATC (Barry & Heater’s old place) involved a piece of the aforementioned Stretch Tendo fabric, and a simple off-the-shelf 80W Halogen light. We attached the fabric in the door way and shone the light on it from about 8 ft away (at that point our imagined cube was still 24ft on a side). The results were mediocre. The biggest realization was that without massive amounts of diffusion between the lamp and the fabric, the light itself was always visible as a bright dot through the fabric reducing the apparent brightness of the fabric itself. Viewed from an angle (we ran up to the kitchen window one by one to get the appropriate angle, while someone blocked the direct line of sight to the lamp) the cloth appeared so much brighter. Diffusion however would loose a lot of the light intensity, shining it back into the cube.

Further we realised that light bleed was going to be a problem. Initially we planned to have little cones around each light, but due to the sheer width of the lamps (and we needed 6, not just 1) these cones needed to extend quite a long way into the cubie space before preventing significant light bleed into the next panels. In the corners this would quickly become prohibitive. Could we move the lights further back to make the cones narrower and thus sharper ? Not a lot, since the scaffolding would start to be in the way.
With 100W bulbs, the brightness was acceptable but not great (provided we could somehow shield the direct sight of the lamp or diffuse it).
That put us at 5600W without even considering computing or music, near 12000W if we went up to 200W. That would be a massive generator, one of those you tow behind a truck and a lot of fuel. A lot of fuel (Anyone have any figures on this ? ).

Switching the 110V, 100W bulbs was also a problem – that is finding a way for a central computer to control the on/off state of 54*6 = 324 lights. Mars had found something acceptable at whistler but at closer examination we found it to be way to expensive for what we needed.

Around that time (April 09) I had moved into a new place, “Danger House”, and  started to get to know my new roommate, Mark Anderson, who happened to be working on a project for a Fire station i think, that involved a pretty glass sculpture and super bright LEDs. He clearly was an electronic wizard (and generally a genius). One night we were at our friend Mez’s house party (i forget the occasion) and we started brainstorming this lighting and switching dilemma we were facing. Mars was feverishly searching for possible solutions on the net (I think we were looking for sufficiently cheap Switching Relay Arrays at the time), – Barry was sitting on the bed looking deeply troubled.
We were close to concluding that this cube project was in fact not achievable within the self imposed budget limitations (we were aiming for $10,000, with lets-be-honest-this-always-happens expected 100% overruns)
Having exhausted my general, non-specific ideas I had little more to contribute, so I went to fetch Mark from the party to see if he had any bright ideas. I fail to recall all the excellent suggestions Mark made that night but the most important single thing he did was to convince us to put super bright LEDs back on the table.

He had just recently worked with this new type of LED unit (Luxdrive), which were available for around 15-20$ per RGB unit.

Luxdrive RGB ENDOR

The solution was to consider all the costs together: While considerably more expensive per light unit, using LEDs reduced costs in switching and in power consumption, cabling, fuel etc. Thus making the overall costs comparable to the halogen solution.

However the final convincing argument however was still to come.

The Day of Truth

The final decision on the lighting approach was approaching and by this point we  had concluded to reduce the facet size to 5ft x 5ft and decided that the only way to prevent light bleeding was to subdivide the cube into little rooms, each with only a single outside facet. Originally  we had thought of black cloth (cheap), thinking mainly about preventing light bleed, but someone suggested using mylar instead, a plastic foil type material with a aluminum coating (Aka PET film  ).

It’s highly reflective (around 99% or so) and thus would increase our effective brightness, preventing diffusive loss into the cube.
Olli and Matt (?) had build a protocube out of 1×2″ wood, measuring 5ft on a side and brought it to David Lockhart’s backyard garage (”The Bronze”). Olli had gotten a roll of mylar and a variety of halogen lamps, Mark had a bunch of LEDs with him. We glued mylar sides into the cube creating a kinked cone like structure inside and covered the outside with the spandex material we had ordered back in Whistler.
We very quickly realized that the way to achieve even glow on the panel was to shine the light backwards onto the mylar which would then reflect the light back onto the panel. The lamp itself could then be occluded using itself or just some opaque disc. The results were stunningly better then the previous light tests and convinced us that the desired effect could, in fact, be achieved.

As for the light source, the brighter colors, such as red, yellow and white worked well with the halogens but the blues did not work at all. Turns out halogens have a weaker intensity in the blue part of the spectrum and filtering out the predominant portions with gels is far from trivial. In fact, by the time the gel is thick enough to leave only the blue portion, the intensity has dropped by 90% or so. In fact the “blue” was barely distinguishable from white.
LEDs on the other hand have a narrow band emission spectrum and should thus have the edge when it comes to saturated colors. Mark held one of his blue LEDs (which by the way have 7WATTs and have to be mounted on a big aluminum plate to dissipate enough heat so they don’t burn through) into the proto-cube and we were all blown away by the sight. The picture below (taken on Barry’s iPhone) doesn’t do it justice – we were looking at a flat, bright, saturated deeply beautiful blue panel. The colors were rippling in a beautiful water-like pattern on the surface because the mylar would ripple and twist with every movement of the air.
We were sold: LEDs would be the way forward – there was no question about that but it presented us with a challenge we had been trying to avoid all along: Having to build large numbers of custom electronics which had to be manufactured, tested and had to somehow withstand the conditions on the playa which are by no means electronics friendly.