In July, we gained more confidence in our ability to physically manifest a scaled-up version of our smaller prototypes. We figured out that by tilting the machine so that the sticks dropped at about a 45 degree angle, we could build angled walls.

. 

Likewise, structural tests performed with our partner lab, led by Jun Sato, gave us more confidence in our ability to build larger structures with the system.

Interestingly, during structural crushing and bending tests, it is very difficult to establish when the sample truly "fails" as it displays an ability to twist and bend in order to find more strength. You could say it tries to locally optimize itself in order to prevent total failure. 

While these tests were being conducted, we started to refine our design concept and form-finding exercise.

At its core, the geometry of the pavilion design is based on the movement of a harmonograph. We quickly built one to see how it functions in reality, and then worked more on our design script to approximate how we could make form out of this phenomenon. The final processing script creates a 3D shape by adding the element of time to the cyclical 2D-motion of the pendulums. 

Over the course of the summer, we were able to build-up a script that tested the location of each local point and calculated how much of an angle it sat in relation to the the same point below. We then colour-coded these angle calculations back to the points. In other words, red dots would mean that we had to pour vertically at 45 degree angles, while green dots would mean we only had to pour at a 25 degree angle.

We weren't sure what the limits were in terms of what angles we could accomplish, both in terms of aggregation as well as structurally, but the analysis did become useful later on when we had to decide on a final form to target.