Design Machines: Don’t trust Robots….. and How to Make (almost) Any Machine

James Coleman
MIT/SUTD Research and Teaching Fellow


Today it is common for industrial CNC (computer numerically controlled) machines (robots, routers, 3d printers, etc) to be transplanted into architectural studios/schools with the hope of expanding design potential. While this addition does extend the studios capabilities, it also makes them beholden to the preconditions, fundamental restrictions, and motives inherent in the machine’s industrial roots. This can quickly lead to projects being abstracted and reconsidered based around machine functionality, lowering the transfer of design intent during the production process.

The direct interaction between architects and manufacturing technologies is useful while designing because it centralizes design, engineering, programming, and fabrication roles. The conflation of these roles has the potential to diversify (and generate) production methods by streamlining the design to fabrication process. Unfortunately, despite the possibilities of high precision control, CNC tools are used predominantly in the same way they were in the 1950s.

Little has changed in CNC machines since the 1950’s and their creation remains rooted in the manufacturing sector. Industrial machines are easy to recognize, robust monsters crafted around longevity not flexibility. These machines are arduous to program, tedious to use, dangerous to operate and expensive to acquire.

Design intent alone is not sufficient to affect production trends and capabilities, a multidisciplinary skillset is required. To create custom automation tools, expertise in a wide array of fields is required. This breadth of skills necessary (mechanical engineering, material science, automation, controls, electronics, design computation, etc), is the primary factor limiting participation in the development of new fabrication technologies.

I am interested in how personal digital fabrication, absent of industrial biases, might expand the capabilities of non-experts and sponsor design experimentation. My research advocates that if the barrier to entry to the creation of custom computer controlled machines is lowered, personal fabrication and automation is achievable. With expanded participation in the machine creation process, an ecosystem of design driven machines would diversify the way we make things and circumvent the pitfalls of predefined workflows.To reach these goals, the process of machine creation and programming needs to be simplified, cheaper, and more flexible.

The use of modular mechanical components and an extensible control system, enables the rapid prototyping of rapid prototyping machines. The Modular Machines that Make Project and the ecosystem of reconfigurable machines born from it, demonstrates the reality of bespoke fabrication tooling and holds vast implications for the future of manufacturing. Don’t trust robots, make robots.