petebooth[iaac]

Moving right along form the last post, I have now combined the data read script as a function into a new script that actually does something remotely interesting! The script reads the data, stores it in an arra and then uses this data to propagate a surface with the specific number of random points. It continues through the dataset, propagating each value of points on a new layer, so eventually it will create a file that has a base surface, with x number of layers, each with a unique number of points in a different color.

It isn’t very intelligent yet, as it tends to put a lot of values near the edges of the surface, but with a bit of proximity detection it should be able to scatter the points more evenly. Hopefully this script will evolve into a packing script of some description. That is the ultimate aim, but i’ve already achieved the most important part of (semi)live data transfer to the rhino environment.

download the script here

After a lot of trial and error, I have finally completed the rhino script function that reads the txt file from processing. As the examples given were all concerning excel files, a lot of further research has been undertaken in order to get to even this basic stage.

The script opens the txt file selected by the user, reads the data until the end of the line (one value), determines if the string is <= 3 characters, and if successful converts the value to an integer and stores it in an array. Anything larger than a 3 digit value is discarded as processing has a tendency to output weird numbers, particularly in the first few entries of each file (creates either a ‘null’ or a very long number).

The next step is to use this data to actually achieve something!

download the script here

Well it’s not completely true, the bulk of it is finished but if I get time there will be some adjustment later on. For now it is done.

download the scripts here

I continued to pursue the live link from the dsLib, but after a couple of emails from the man himself, it is apparent that there are a few technical problems that I am unable to overcome. So taking that into account, I am now going for a semi-live link.

The code for the arduino circuit has been updated to allow for a better delay time between sensor reads. This script is written in arduino/C++, but as it is a fairly simple and well documented operation is was quite easy to get it working properly. Unfortunately the arduino language doesn’t support export to a text based file natively (to my knowledge), so I have used processing as an in between interface for this task.

The processing script (also based on C++) establishes a connection with the serial port, reads the data flowing through it and when it has received 40 sets of data, dumps it to a csv text file that can be read by rhino. This script will hopefully evolve a little more to create multiple csv files, allowing for rhino to bed fed data in a stream like manner. At the moment it will overwrite the file each time.

The final script I have been working on is the beginnings of the rhino side of things. Presently I have created an endless loop that prints the rhino build information continuously until the snap control is turned on. It create a graceful way for the loop to end. This is going to be built upon to allow rhino to read the csv files and do something with it. At this stage I am aiming for rhino to create a series of objects and ‘animate’ them based on the input data.
Hopefully all goes well.

Now thatI have the arduino board and script working properly, the time has come to tackle the rhino script to open connection with the serial port. As previously posted, vbscript and rhinoscript has no builtin functionality to access this information. The additional vbscript functionality is made possible by Stylianos Dritsas, who has completed some work in this area previously.

It is a case of downloading the additional library, and installing them on the computer, although this proved to be a little harder than first though, needing things to be added to the system registory and a script loaded in rhino. After a lot of fiddeling around, it seems to be setup correctly.

With the limited help files on the support page for the libraries, I have written a script that in theory should allow rhino to read values from the serial port, and print them on screen (once per script at this stage). With all of the scripting errors now elimated it is returning a value, but it is always the number 1. I think there could be a problem with arduino sending the data correctly to the live port or, more likely, my script isn’t addressing the port correctly.

download the script here

The first part of my physical sensing link to rhino is the creation of the circuit and the beginnings of the arduino script. For the initial testing a simple circuit has been created using the arduino diecimila in conjunction with a bead board. At this stage the circuit has a light sensor and a LED.

The arduino script that powers the circuit is also relatively simple at this stage. It takes light readings from the light sensor, and turns the LED on in low light conditions and off when the opposite occurs. The script contains a threshold value that can be set between 0 and 1024 (the range of the light sensor) to allow for fine tuning in different light conditions. The magic starts to happen when the script outputs the sensor data to the serial port, allowing other application to access the data.

you can download the script here

At this stage the circuit and script are working correctly, and the next stage will be to create an open communications link between it and rhino. The additional vbscript commands that will allow this to happen are provided by Stylianos Dritsas, who has completed some work in this area previously. Hopefully I can get this up and running soon, as it is fairly pivotal for my project.

The initial research for our Deleuze article is complete and received good comments. It should form a solid foundation for the final paper.

A link to our presentation on the above article can be found here

After one of three forms was selected, we needs to test different variants of sub-division in order to develop a construction method. The division of Juergan’s model below is based upon spiraled triangulation, similar to a lot of Foster’s more recent work.

Somewhere along the line I managed to change the division pattern, which at first i was pretty disappointed with (especially since it wasn’t until I has laser-cut the physical model that i realised). Although it looks kind of odd, in retrospect it was actually nice to look at two different triangulation patterns, both of which developed natural during the topsolid modeling process.

Looking at domes from our our countries, there are two from Australia that stick in my mind.  Both are not domes in their true sense, but are segments of domes.

The first is the Atrium dome at IXL Art Hotel in Hobart (Morris-Nunn + Associates, 2004). The curved glass roof is formed from a very flat dome, and is cut square to fit over the courtyard.  The primary structure consists of relatively small laminated timber gridded framework, complemented by a steel wire tensioning system.  This enables the roof to provide maximum light penetration, while keeping the visual weight to a minimum.

The second is one that everyone should be familiar with, the Sydney Opera House by Yorn Utzon, 1973.  The primary shell structure consists of a series of vaulted shells, that take their curvature from a single sphere.  Different sized portions and angles of placement create the variation within the overall building.  Consisting of a pre-cast concrete panels, the form was easily divisible, allowing or easier and most cost effective construction fabrication.

Interestingly, division of the surface is not triangulated.  The tiling systems was developed in sheets, allowing for them to be fabricated at ground level and placed in large pieces, rather than individually placed at height.

For the first stage of studio this semester we were required to develop a dome model within topsolid. mine is base on three pure ellipses and has a symmetrical cutting plane to slice openings in the sides.

You can view a video of the live topsolid model here on youtube.