This week we were tasked with creating a press fit construction kit using the laser cutter and some sheets of cardboard.

As part of my research in the Center for Bits and Atoms, I'm working towards building a "digital material assembler." This assembler will be capable of assembling discrete pieces of material (much like these press fit pieces) into any arbitrary object. So, for me, this assignment was a perfect introdcution into the world of "digital material" and got me thinking about ways to automated the assembly of these kinds of objects

After reading George Popescu's paper on the GIK (Great Invention Kit) I decided I wanted to play around with GIK myself and to perhaps try integrating electrical circuits into the pieces themselves.

I first modeled the GIK piece in Solidworks. I know Solidworks is a little overkill for such a simple parametric piece but I wanted to be able to visualize a 3D assembly of the components before laser cutting. My first idea was to make a large cube and include circuitry to control 36 LED's on the outer wall. The idea being that it would be possible to multiplex the LED's in a grid and make them individually controllable with 15 I/O lines.

(left) GIK piece. (right) GIK cube. Blue and red are different types of conductive pieces.

To be able to make any arbitrary electrical path through the object, I figured I would need two types of conductive pieces:

• One piece would have all four terminal electrically connected. (pictured on the left below)
• The other piece would have two independent current paths (one vertical and one horizontal). (pictured right)

I used the vinyl cutter to cut out strips of copper sheet and stuck them onto the cardboard pieces like so:

I laser cut 208 pieces (only some of which I cladded in copper) and started playing with assembling them in different configurations. The assembly process was actually very different from what I was expecting. I was under the impression that the assembler would be able to assemble objects one piece at a time. I very quickly realized that assembling one piece at a time was harder than it sounded because of interferences and interconnections between parts. Instead, I found that it was much simpler to first assemble "voxel" sub-assemblies from multiple individual pieces and then assemble those voxels into an object.

Two types of "voxel" sub-assemblies.

The great thing about digital material (press fit kits) is that they're totally re-usbale. So, I made lots of things...

Along the way I discovered an interesting pattern that made me change directions... Certain configurations make a perfect 7-segment LED display:

So I set out to build a 7-segment LED display from digital material pieces that may one day look like this:

The seven-segment LED has eight connections: seven to control the individual LED's and one ground. I soldered the LED's onto the edge of some of the pieces. The two types of electrical components (described earlier) allowed me to create a ground signal that ran throughout the object without interferring with individual LED-segment signals.

A little programming and an hour or so of debugging and it was up and runnng...

Now, although it looks really nice, there are definitely some problems with the way I did the electrical connections. As it is, there's very little pressure enforcing electrical contact between pieces. This means that unless you do some serious fiddling and squeezing of the structure, getting more than ten electrical connections to work at once is a huge pain.

Earlier in the week I tried cutting copper on the vinyl cutter in a more interesting shape that would enable it to fold around the inside of the press-fit and make a good, solid connection. This was great in theory but in practice it just wasn't feasible to fold the copper without it folding on itself or becoming loose.

In terms of future work, I'd definitely like to explore better ways to create electrical connections in these types of digital material objects. While playing with the pieces I also had the idea of using something like a vacuum bag to act as a skin. This would allow for the pre-loading of the press-fit structure and make the whole object much stronger in tension. To be explored...

Download source files:

• solidworks gik
• solidworks gik assembly
• gik laser layout (dxf)
• copper vinyl (dxf)
• arduino code

Quick tips:

• Laser cutting cardboard settings: Epilog (60W), power = 25%, speed = 75%, freq = 700
• Mercurial order of operations: (courtesy of Kenny Cheung)
• hg pull
• hg update
• do work...
• hg add .
• commit -m "description Of Work Done"
• hg push
• if it works, great, otherwise the destination repo changed after your pull, so continue with merging
• hg pull
• hg update
• hg merge
• hg add .
• hg commit -m "merge"
• hg push