This high-performance / low-tech design develops an ongoing interest (begun in 1998) in using passive off-the-shelf components to activate individual architectural apertures or entire enclosures for environmental control. In this case, taking readily available greenhouse window/vent auto-openers which use pistons (filled with a temperature sensitive wax) that expand and contract to create mechanical movement and using them to actuate dynamic facade/surface ventilation.
Various origami folding techniques were explored for methods of opening and closing the apertures. This particular system uses a hex or honeycomb geometry, but a number of geometries could potentially work.
*the assembly shown here is an approximately half scale mock-up with manually adjusted operation.
Step 1: Overview
This will be a brief overview of the steps involved in building / prototyping this model. I won’t go into the specifics of every aspect - it will be more of a synopsis of the workflow, fabrication processes and organization of assembly.
Various small diameter stainless rounds
Aluminum and brass tubing
Various small fasteners
Pantone drafting film
Objet polyjet 3D printer
Allen wrenches, small tools, superglue, etc.
I actually debated whether I would make this during my residency at Pier 9 since it was an old project (living on my computer since 2011). I decided to go ahead with it for a few reasons:
1. I had already been working on parametric node/joinery design research and so using a parametric system to generate almost all the node connectors was fairly easy. I simply adapted some of the parametric definitions to accommodate the geometry, node type, and strut sizes I was working with.
2. Pier 9 has so many resources, so I figured while I could do this at another time with other machines and tools, I could make it fairly quickly and easily with the tools I had such immediate access to - mostly the Objet high resolution printers.
Step 2: Design
While the general concept design was already done in a large part, I still needed to model the ‘cells’ I planned on prototyping very exactly so I could build right from the model. Once a general sizing and configuration was settled, I was able to use my parametric definitions to generate most of the nodes. Some of the nodes were specialized (such as the integrally printed hinge pieces) and so I modeled those individually. The ‘petals’ were simplified from an earlier scheme where I had them interlocking/nesting to make this first prototype more feasible. The patterning on the petals is a hypothetical frit pattern which could be adjusted as needed to provide a gradient of shading if desired. This was done parametrically and I created 3 patterns to test - one for each ‘cell’. Since I wasn’t using actual automatic vent opening cylinders (at this point I just wanted to test the design and mechanics) I needed to design a stand in - so I went with a simple tube with the central ‘piston’ rod sliding through and a thumb screw to stop it at any point along its travel. Once I had these aspects all designed in 3D it was time to fabricate the individual components!
Step 3: Struts
Probably the most straightforward set of this project - my parametric definition was designed to output exact lengths for all the strut members. I used a cold saw to cut a few small diameter rods to exact length - because they’re so small you have to cut very slowly and carefully. I then deburr on the belt or disc sander.
Step 4: Petals
To give the cells a little variation between the three I made each corresponding set with a different height ‘peak’ - so essentially I had three sets of six identical petals. Since each cell had two neighboring cells (in a larger field, each cell might be surrounded on all six sides) I needed to coordinate how the edge hinging was justified. Pretty simple, but something that could also be overlooked. Cutting the 3/32” acylic on the Metabeam laser was quick and easy. I was just cutting the holes and profiles.
To get the ‘frit’ pattern a little more labor was involved. I first etched a placement profile in some card which let me know where to set the petal. I then placed a piece of Pantone film on the petal (roughly cut with an x-acto by hand) and used the laser to engrave the frit pattern. After that was done, I just removed the remaining film by pulling it back and the frit dots were left on the petals. If needed I burnished them on better. As I mentioned, I had three frit patterns and had to repeat this six times for each one.
Step 5: Nodes
Once all the parametric nodes were generated and the remaining ones manually modeled, they were almost ready for export. I exported each unique type in .STL format to be printed (some were printed in multiple). For some of the nodes, especially the integrally hinged ones, I had already done individual prints/tests for fit, strength and movement. This saves you from printing an entire batch of unusable parts.
The printing was easy, it’s the cleanup of 3D printed parts (especially of LOTS of small parts) that takes time and patience! After spending some meticulous hours at the cleaning station, the nodes were finally finished and ready for assembly.
Step 6: Assembly
Being organised during fabrication or assembly of just about anything is paramount to a successful (or at the least frustration free) project. I like to lay out all the hardware, pieces and parts, tools, etc. After laying out all the tools and components I began by putting together sub assemblies. Order-of-operations might be the most important consideration when assembling. I find it helpful to outline each step before beginning and to assess and cleanup after finishing each step in preparation for the next. Since this project had to be assembled to test, I didn’t want to glue each piece as I went - I wanted to ‘dry-fit’ the project first before committing to full attachment! This meant I had to glue at the joints/connections after the project was already assembled - a bit more tedious, but i wasn’t comfortable gluing as I went.
Step 7: Conclusions
Overall the prototype worked well - I think it communicated the design and intent fairly well. Over time some of the joints/nodes have begun to sag or have cracked. While the Objet Vero Clear material makes some great looking prints the material is a bit weak and brittle. If I decide to make a second working prototype (with the actual automatic vent opening piston) then I’d probably change the design to be made using more robust materials and methods.