Light Wall

NUSTEM’s interpretation of an Exploratorium classic

The Exploratorium in San Francisco have a magnificent Tinkering Studio. It’s a space where educators, artists and engineers collaborate to produce beautiful, thought-provoking, and just plain fun activities and installation pieces.

One of their classics is Light Play, a wall of frosted screens back-lit with weird and wonderful shadows interacting with the light, with colour filter, and with the screens. It’s a tremendous activity, particularly good with younger children and family groups. One aspect we particularly like is the way it fosters a sense of collaboration and contemplation. If you’re working on the arrangement of materials behind the screen, you can’t really see what effect you’re having, since that’s only properly evident from in front of the screen. So you have to work with others, articulate what looks good, and work out how to get that effect again.

The Exploratorium have written up their approach on their own site and at Instructables.

1 2 3 4 5 6

Light Play workshops from the Exploratorium. Images © Exploratorium, used with permission. Thanks, tinkering.exploratorium.edu!

Our version

We loved the idea of Light Play, and our colleagues at the Centre for Life were keen to produce a set for use here in the North-East of England. But while the Exploratorium’s setup is wonderfully dramatic, for our purposes there were some challenges around component availability, ease of construction, and cost.

So we simplified the components considerably. The results aren’t as gorgeous, but they’re still pretty good. The activity has prompted the response and interactions we wanted, so we’re happy.

Here are a few more details about what we – Think Physics and the Centre for Life – built:

Boxes

Making boxes for the light wall: our intrepid Think Club makers get stuck in.

There’s nothing clever about our boxes; they’re large cardboard boxes with big holes cut in each end. One end is covered (internally) with good-quality tracing paper to form a screen, the other is used for access.

One thing to bear in mind: your light wall will use a lot of boxes, and hence a lot of space.

Reaching inside a box during a school workshop.

Things we’d do differently next time

Use smaller boxes. The ones we picked were 465 x 465 x 600mm, and are just a bit unwieldy in use and particularly in storage. I’d go 100mm smaller in each dimension.
Cut the holes in the sides, not the end. Rectangular apertures are more interesting than square apertures.
Using the sides of the boxes also means you don’t have the tape holding the boxes together encroaching on the display face of the wall. Oops.
Absolutely use double-wall boxes. It’s tricky to cut, but the resulting boxes are tough enough to stand up to many uses – we’ve so far done two Maker Faires plus an art gallery installation and a bunch of school workshops. The boxes are looking scrappy, but they’ve at least one more busy weekend in them.
Curved corners to the cut-outs are a pain to make, but absolutely worth it in the long run. They look more pleasing than square corners.

Lights

The Exploratorium’s lights use single-LED high-output devices, which are lovely but quite costly in the UK. They also require heatsinks.

Instead, we adopted triplets of extra-bright conventional LEDs. Having three individual light sources, even closely-spaced like this, isn’t ideal – they cast rather awkward shadows rather than the crisp outlines a single source would give, and they’re only just bright enough. However, the effect looks better than we’d feared, and the trade-off was worthwhile in our case. We used sets of three high-output LEDs of various colours, ranging from about 8,000 mCd output up to something like 11,500 mCd. Bear in mind that the sorts of cheap LEDs you’ll find in every electronics kit are rated at about 300 mCd. They’re also fairly cheap from suppliers like Kitronik and Hobbytronics.

We mounted the LEDs up on prototyping board, in series with a single current-limiting resistor to suit the specific LEDs, and powered them off 9V cells (so, ~3V each).

1 2 3

Things we’d do differently next time

The gorillapod-style positionable necks were a pain to make. Literally. We bought cheap mini tripods from Amazon, un-popped all the segments, drilled them out, popped them back together, then threaded wires up the holes. There were blisters.
In use, we rarely see people position the lights other than by laying the entire unit on its side. So:
…next time, I’d just use a short length of dowel or plastic tube as an upright, and maybe mount the LEDs on a couple of segments of positionable leg.
I’d be tempted to make only white light sources, and offer a range of lighting gels for colouring.
Consider using small lamps like these from Ikea or these from Clas Ohlsen. On balance I’d probably prefer battery-operated lights, as they allow a finished box to be moved into a group wall display, but there’s a lot to be said for an off-the-shelf lighting solution too.

Fancy Lights

We made a small number of these ‘fancy’ lights. They use an Adafruit NeoPixel ‘Jewel’ array of RGB LEDs driven by a 5V Trinket microcontroller (like half an Arduino, and yes that’s a terrible description). Power comes from a single AAA cell via a step-up board which I still think must work via unicorns or something. I mean, how do you make 5V from 1.5V?! Mind-boggling.

When first booted, a small potentiometer controls LED colour. A second (button) input toggles the Trinket into a colour-cycling mode, with the potentiometer then controlling the rate of change of colour.

They can produce delightful effects… but the power drain is such that they last about half an hour, and hence aren’t terribly useful. Also: everybody wants one. Ultimately, we’ve withdrawn them from the workshop because they don’t add enough to make up for their limitations.

They’re still cute, though.

Turntables

To bring a little motion to the boxes, it’s worth having a few motorised components, like turntables. But they need to turn slowly, which is a challenge. Since I’ve never done the Exploratorium’s workshop personally I just assumed they’d made their own, but compiling this page I’ve realised they seem to be using jewellery display turntables, like these or this.

Since I didn’t know that, I ended up reverse-engineering something which didn’t exist in the first place. Hmm.

Well, I’m proud of our turntables. They turn slowly (about 1rpm) and despite looking pathetically fragile have proven quite robust in use. The old exhibit standby of ‘if it looks fragile, people treat it with care; if it looks solid, they’ll whack the heck out of it’ seems to apply. They do tend to fall apart in transit, but a minute’s ministrations with a glue gun solves that.

The core is a geared 3–6V motor, driven off a single 1.5V cell. There’s enough torque to work reasonably well, and undervolting it slows it down even before you get to the gearing. The drive pulley is a short length of plastic tubing with a cardboard end-cap. A slot in the cardboard cap fits tightly on the motor shaft, then hot glue poured into the tubing fixes it well enough. Gaffer tape wrapped around the plastic tube provides enough friction for the edge drive to the turntable itself.

Since we hand-cut the turntable discs they vary in diameter somewhat. The pivoted chopstick and elastic band tension spring keep the drive pulley and turntable in contact. The main turntable bearing is a short M10 bolt glued to the frame – ridiculous overkill, but it’s what we had lying around.

Things we’d do differently next time

For such a simple design, the turntables are surprisingly fiddly to make. There are a lots of components, many being slightly-hidden bits of wooden chopstick used as supports or guides. Each of those pieces needs to be cut, sanded, and glued in place.
I’d simplify further. A better bearing would allow for tighter tolerances, and you could then probably laser-cut ply to give a geared drive rather than the tensioned friction drive we used. You’d lose some of the charm of ours, but you could likely finish dozens in a day.
Tempting as it is to go to a stepper motor and variable-speed controller, the unit cost would go way up. Also, the power consumption of these things is impressively low; they’ll run all day on a single AA cell.
While the maximum payload is quite low, in some ways that’s a benefit as it forces participants to find creative solutions. As ever, we push against limitations and come up with inventive work-arounds. Completely ‘open’ play, without constraints, can be less creative.
One common approach from participants is to fix objects to the drive pulley directly, not just the turntable. They then have two rotation rates to explore. Another example of finding creative utility from apparently-poor design.
One advantage of this design’s rustic approach is that it can become a talking point itself. None of the components is unfamiliar, which helps it feel accessible; there’s a sense of ‘I could have built that’ about it. Yet it’s seen by a wide range of ages as ‘clever.’

Turntable base with platter removed to show bearing and support

Close-up of the drive system. Most of our units have an additional chopstick-segment guide stuck to the motor, which rides under the turntable platter and offers some extra support. I didn’t notice it had fallen off this one before taking the photo, because it worked perfectly well without it. Be aware that additional bodging may be required!

Turntables as far as the eye can see!

Effects, outcomes, and miscellaneous notes

This is a beautiful workshop and installation. The impact of the assembled wall and its ever-changing display is mesmerising, and it delights all ages. We’re not surprised that the new Tinkering Space in @Bristol uses a light wall as a prominent feature.

At Maker Faire UK we observe:

Broad appeal; families are drawn to the activity, but when there’s clear spare capacity pairs or groups of teens and adults will want to be involved.
Family groups working collaboratively.
Dwell times of typically 20 minutes, ±10 mins.
Lots of back-and-forth revisions and adjustments.
Discussions around aesthetics, properties of light, narratives, mood…
With about six work tables, we can work with ~400 people/day.
Many more people notice and watch the finished display, but we’ve not counted those interactions.
Two lights per box is about right. Many groups will use just one.

Things we perhaps still haven’t nailed are typically around the available choice of materials, and the nuances of facilitation:

Too many plastic dinosaurs = ‘Lost World’-themed light wall. A constrained choice of pre-made objects and an abundance of raw materials is probably better.
Few people try moving the light in front of objects (ie. casting reflected light onto the screen). Similarly, shaping the light by using masks close to the source is fairly rare.
Go back to the top of this page and look again at the rotating gallery of the Exploratorium’s workshop. Notice the amount of patterned acetate they’re using. It produces beautiful effects, but we’ve found it tricky to entice people to give it a chance.
The turntables are good, but they don’t prompt a particularly wide range of effects. I’d like to motorised widgets which could be fixed to the top of the box and turn things on strings, or could wave things back-and-forth, or … (etc.)
High-mounted lights + low turntables = all the movement is at the bottom of the box. There are many ways to fix this, but we need to make some of those solutions more obvious.