Extruded Exercises

Ultimaker Cura

Ender 3

Design Process
An essay on 3D-printing  for architecture and Operative Design’s catalogue of spatial verbs.


After reading Operative Design by Anthony di Mari and Nora Yoo, I started a personal exploration into simple architectural design applications for 3D printing. The content of the book seemed like a perfect match for a series of practical mass modeling exercises; Operative Design provides a library of “spatial verbs” that are used to generate architectural forms. It also provides partis of existing architectural works using their “spatial verbs” for parti generation. By 3D printing my own library of forms based on this work, I had three goals in mind: 1) to generate my own library of physical and digital masses which I could use as design tools, 2) to practice 3D mass modeling in Rhino, and 3) to understand how I can use 3D printing in my own design practice.

There is no disputing that the strengths of 3D printing are in being able to produce and reproduce complex forms that have been digitally modeled or scanned from life. Its potential is exploited in custom replacement body parts and in fantastical haute couture garments. It seems excessive to 3D print a rectilinear model for architectural design purposes rather than just make it with glue and card. However, from my own experience with digital fabrication technologies, I know that simple forms can be deceptively difficult to fabricate and so a foundational knowledge of 3D printing will aid in successfully increasing print complexity. Despite the lack of “organic” or curvaceous forms, some of the volumes in Operative Design are sufficiently complex that recreating them by hand would be challenging.

Technology and Materials

While there are many types of 3D printing now available, fused deposition modeling (FDM) is the most prevalent and accessible form of personal 3D printing; the most common printing materials are polylactic acid (PLA) and acrylonitrile butadiene styrene (ABS) in a filament form. Composite materials that include a percentage of other materials like wood or stone are available. The printer I used was the highly regarded but relatively inexpensive and fully open-souce Ender-3, and the modeling was completed in Rhino.

I chose filaments based on the relationship between the material’s use in architecture and the spatial verb categories used in Operative Design:

[ Category ︎ Architectural action ︎ Material ]

Subtract ︎ Carving ︎ Marble
e.g. taper, inscribe, grade

Add ︎ Framing ︎ Wood
e.g. offset, extrude, expand

Displace ︎ Forming ︎ Concrete
e.g. twist, bend, lift

These relationships ended up being more poetic than anything but they did allow me to test a variety of 3D printing filaments to see how their marketing compared to reality. I was also drawn to materials that seemed more sensual than the plain PLA. I used a wood PLA composite, a concrete PLA composite, and a marbled semi-translucent PLA. I also used a white PLA for the aggregations and parti models.


Moving from one spatial verb category to the next (add, displace, subtract, aggregate, parti) I modeled each verb in Rhino, beginning first with a base model upon which to perform the transformation. Operative Design used subdivisions of a cube for its selection of base models and I did the same.
My 1/4, 1/2, and 1/1 base models.

I printed each model, taking care to deliver the highest quality I could given the combination of filament and form. This was a difficult process and after a lot of trial and error I developed an increased intuition of what is required to print certain forms in different materials. Fortunately there is a massive online 3D printing community, so debugging specific print issues wasn’t always painful.

Without experience it’s easy to take for granted that the object modeled will be the object that is printed. So when I say “the highest quality” in this case it means reasonably free from surface defects. However, it’s nearly impossible to avoid the distinctive layered appearance left by FDM printing. The layer height can be set small enough to diminish the appearance of the layers but the trade-off is a long printing time. Understanding the settings for, and placement of, any printed supports is important to ensure that they won’t leave surface marks when removed.

There are so many different surface textures created with this type of 3D printing. Examples of factors which affect surface texture are the set height of the printed layers, the orientation of the model on the print bed, the type of surface being printed on, the angle of the model’s walls, whether supports are needed, and whether a setting called “ironing” that smooths out the top layer of a model by running the hot printing nozzle over top of the last layer is used. The important take-away is that the object printed will hardly have a uniform or “neutral” appearance without forethought on how to achieve this.

The basic forms were mostly simple to print as compared with the aggregates and the parti models. Some creativity was required: to orient the model correctly for printing, to understand what sort of removable support was needed, and to find the correct settings for each type of filament. The parti models allowed for more creativity in both editorial and production capacities. The architectural partis in Operative Design are good illustrations of their spatial verb approach but as they are rectilinear buildings they do not exploit the capabilities of the 3D printer. Nevertheless, I printed a few of them and chose to focus on other aspects of architectural modeling that might benefit from 3D printing: the creation of topography and of parts that require high tolerances in order to fit together. This part of the exercise was not about understanding what 3D printing can do; that has been well-proven by the thousands of existing 3D printed architectural models. It was about gaining the experience for myself.

Carve + Offset:  Poli House (Pezo von Ellrichshausen)

Embed + Branch: Villa 1 (Powerhouse Company)

Embed + Overlap: Casa para un carpintero (RCR Arquitectes)


On the Operative Design Forms

I have yet to design a project using the forms themselves as a design aid. It’s important to remember that the forms illustrate actions and do not illustrate spaces (which is consistent with the intent of Operative Design ). Using them as an idea-generating aid in the design process is less a matter of stacking them into totems than it is meditating on the operation each embodies.

While it’s easy to see how how the basic operative categories of  “add” and “subtract” can be used as starting points while using the form library as a catalogue (which, by the way, is not something Operative Design purports to provide), that is less the case with the category “displace”. The category of “displace” is for verbs that signify a type of movement, e.g. split, skew, bend, but also, somewhat confusingly, for a type of coupling, e.g. intersect, overlap, interlock. The category further breaks down when you compare the “displace” verb “lodge” with the “add” verb “nest”; what exactly is the difference between nesting and lodging from a spatial standpoint? Ultimately it doesn't matter, because what Operative Design is trying to do is be a point of departure for how space design might be thought of as a set of actions one performs.

The ways I chose to interpret the verbs in Operative Design were heavily influenced by the ways the book interpreted them, and this resulted in entirely rectilinear forms. There is nothing in their process to suggest that spaces formed using this kind of active language must be rectilinear.

On 3D Printing and Architectural Design

Sketch models vs Presentation models

One of the first advantages I noticed when creating the parti models was that 3D printed topography is great and the ability to integrate it easily with section models is a benefit. The high tolerances achievable between model parts allows for articulation and integration. When printing curvaceous or undulating site models, the extruded filament can create a high resolution version of the manual cut-and-stack site model technique (which is time/resource consuming and expensive even with laser/CNC cutting). 3D printing should be considered anywhere cut-and-stack would otherwise be employed, depending on the model scale.

PLA has the additional useful feature of recyclability; sketch models can easily be chipped up and turned back into PLA filament (a variety of machines are available for this purpose). In the grand scheme of sustainable behaviour however, it’s difficult to believe that focusing on the recyclability or biodegradeability of a design professional’s models is an important leverage point compared with the impact the design of the built work will have.

The parti models I printed are too rigid to be manipulated and used as sketch models, but my intention wasn’t to create sketch models. However, my first impression was still that 3D printed models have low flexibility when compared with other kinds of sketch models. Thinking a little more on it, it’s easy to see the utility for exploring designs that are configurations of modular components. Additionally, the ability to accurately reproduce existing buildings in 3D printing for adaptive reuse/restoration/rehabilitation projects is an advantage. For example, 3D printing takes any guesswork out of creating an accurate physical model of existing structures or features if they have been laser-scanned.

Materials vs Filaments

The range of available materials for 3D printing is misleading because it doesn’t matter if a filament is wood or stone, the resulting printed material has gone through an identical process. This gives traditional architectural model materials a distinct advantage as materials like cardboard and plexiglass maintain their distinctive characteristics regardless of how they are cut or glued. Yes, there are some differences between filaments, e.g. the wood filament can be sanded and the concrete filament can appear quite matte in some circumstances, but I really don’t feel like the differences are enough to counteract the uniformity imposed on them by the printing process.

Materiality vs Form

Small changes in materiality are important in architectural models because what is presented must be carefully edited for clarity at smaller scales. Unintended information can be conveyed through byproducts of the 3D printing process if unconsidered, e.g. a glossy surface created by being face down on the print bed might convey a change in material if all the other surfaces are matte. This seems like a trivial observation because all materials require some amount of forethought when using them for models but it deserves mentioning as some of the byproducts of the 3D printing process are unavoidable.

Layerability and unpretentiousness of material are often praised characteristics of handmade architectural models. It’s hard to imagine in fully 3D printed models the same ineffable cohesion of design concept and model execution as can be found in brilliant sketch models (as in the compelling Zeitz MOCAA cardboard tube carved silo model from Heatherwick Studio). But I hypothesize this has less to do with space illustration as with persuasiveness. Sometimes I think that the penchant some of us may have for the sketchier model comes from appreciating a certain cleverness in execution rather than the articulation of architectural space. But I recognize that it’s incorrect to assume these techniques are in opposition. The utility of either is determined by different factors, like stage in design process, content of design process, and personal need of the designer to design with their own hands versus their ability to express a design digitally.

My roughly printed concrete and wood filament versions of the original cardboard tube MOCAA model.

There are many different purposes for architectural models. We need clever and sensual models to illict emotion and describe materiality, we need graphic and abstract models to explain key concepts, and we need whatever form our model takes that helps us navigate our design process. Sometimes we are best served by Xactos and cork, and sometimes we need a nozzle carefully extruding filament because our own hands can’t do our forms justice.

Form is Not Enough

Due to the scale involved, both complex forms and materiality are expressed when 3D printing is used as the fabrication technology for architecture itself. I think the expressive capabilities of 3D printing are more fully exploited in this way. Architectural purposes aside, I would like to learn more about 3D printers with regard to their expressive potential, and not just their potential for mimicry and democratized fabrication. I believe there is real potential in the technology for art and design, and I see it when the materials used are sensual in their own right or when the method of printing exploits the machine's formal restrictions. This can be seen when creators experiment with modifications like extra large nozzles or add moisture content to their filament, solidifying steam bubbles into their prints. Objects like this are able to demonstrate the surprising beauty and authenticity of their fabrication.