Shells of Light, Blobs of Shadow

Translation of digital parameters to textural qualities in the physical


Material Systems
Fall 2021
Harvard GSD
Instructors: Nathan King, Zach Seibold

Team: Leonard Palmer, Sarah Hopper, Annabelle Li
The team’s investigation began with experimenting with additive manufacturing processes and experimenting with fabrication parameters such as speed and over-extrusion. We discovered that spherical geometries, or what we will refer to as Bubbles, could be created in two different ways, utilizing two different toolpaths. We decided to call these two toolpaths BLOB and SHELL. BLOB follows the spiral path of a base cylindrical surface and slows to over-extrude material in the places where the bubbles are located. SHELL follows the boundary of the combined bubble and cylinder geometry and maintains a constant standard speed. It creates a hollow shell of the outer surface of the bubbles instead of a fully solid geometry. Through these two defined toolpath methods, we are able to create bubbles which look the same in their digital form but differ in characteristics in their physical form.
How can fabrication parameters associated with the additive manufacturing of ceramics be used to create bubble geometries on a surface that read similarly in the digital, but push past its constraints to manifest into two differing languages of novel textural and porous qualities in the physical? To what extent can these qualities be predicted and prescribed to a designed object?

BLOB creates a textural and leafing surface with gradual transitions from base cylinder to Bubble and thick, fully opaque walls. SHELL creates a smooth and layered surface with more legible boundaries between cylinder and sphere and thinner, constant wall thickness. When the bubble overhangs the circumference of the cylinder, SHELL also creates holes, droops, and gaps among the droops which begin to add levels of porosity in the SHELL bubbles that contrast the opaqueness of the BLOB bubbles. These effects push the design of the object beyond what can be defined in the digital 3D model and creates a design process that has components in the digital creation of the model and the understanding and interpretation of the details and surface characteristics that will be added by the robot through the additive manufacturing process. In a sense, the craft of the machine and the characteristics of the clay body create something even more beautiful and tactile than a simple 3D model can emulate.
The BLOB toolpath was scripted so that whenever the contour lines are inside the Bubbles, the speed would decrease to 9mm/s, as opposed to the standard speed of 25mm/s for everywhere else. The 9mm/s slow speed was determined after a series of initial tests to balance between the legibility of the BLOB and avoiding over-leafing.
The SHELL toolpath was scripted so that the extruder followed the contour lines of the joined surface of the base cylindrical surface and Bubbles. The toolpath speed was scripted to run at 25mm/s.
There are two variable parameters in this research: sizes and cantilever spans of the spheres on the cylindrical base. With the results of previous print tests, we speculateed that both parameters will result in various degrees of drooping and hole opening effects on spheres. To test the effects of sphere size, we tested 8 incrementally sized spheres ranging from 10mm diameter to 80mm diameter to see how size difference correlates to drooping distance of spheres. The constants of the test include print speed, cantilever span, cylindrical base and clay body consistency.
To test the effects of sphere size, we tested 8 incrementally sized spheres ranging from 10mm diameter to 80mm diameter to see how size difference correlates to drooping distance of spheres. The constants of the test include print speed, cantilever span, cylindrical base and clay body consistency.
To test the effects of the cantilever, we tested 8 spheres at various lengths of cantilever spans. The cantilever span is calculated as the distance from the center of a sphere to the perpendicular point on the surface of the cylindrical base. Each sphere moves away from the cylindrical base surface by 2.5mm starting at 2.5mm span length. The constants of the test include 90% print speed, 50mm diameter sphere size, base geometry and clay consistency.
To test the effects of the cantilever, we tested 8 spheres at various lengths of cantilever spans. The cantilever span is calculated as the distance from the center of a sphere to the perpendicular point on the surface of the cylindrical base. Each sphere moves away from the cylindrical base surface by 2.5mm starting at 2.5mm span length. The constants of the test include 90% print speed, 50mm diameter sphere size, base geometry and clay consistency.
From top row to bottom row: Rhinoceros digital model, 3D scan of physical print, overlay of digital model and 3d scan of physical prints
In order to demonstrate this new set of design tools, we created a standing floor lamp as a proof of concept. The physical interaction offered by furniture enhances the distinctive characteristics of our BLOB and SHELL bubbles. Using bubble size and overhang distance as key parameters, we modeled the lamp’s bubble formations digitally, anticipating their physical behaviors such as droop and structural coupling.

The lamp comprises six vertically stacked cylinders, each progressively transitioning from heavier BLOB-rich compositions at the base (2/3 BLOB, 1/3 SHELL) to lighter, porous SHELL-dominated forms at the top. This gradual transition leverages BLOB’s thickness and weight for stability at the bottom and SHELL’s lightness near eye-level for visual delicacy. The resulting structure features two spiraling ribbons of contrasting densities and textures. This final lamp is a testament to the beautiful textural and light qualities created with these combined toolpaths and to the research that built our semi-predictable set of bubbles, holes, droops, and leafing. 

Pieces exhibited at Harvard’s Ceramic Studio in Allston, MA.
©2025 Leonard Palmer  All Rights Reserved. No visual or written material on this site may be copied, reproduced, distributed, or published in any medium, in whole or in part, without prior written consent by Leonard Palmer