Embroidery with a Shopbot by Gabrielle Benabdallah

“This alerts us… to the possibility of encounters at the interfaces conceived very differently than as the meeting of a human and a machine, each figured as a self-standing entity possessed of preestablished capabilities. Rather, effective encounter at the computer interface are those moments of moving complicity between persons and things achieved through particular, dynamic materialities and extended socialites.”

— Lucy Suchman, Human-Machine Reconfigurations: Plans and Situated Actions, p. 245.

Developing a feminist framework for interactive CNC machining

Computer Numerical Control (CNC) machining has a history rooted in managerial control, military research and factory work optimization (Noble 1984). Despite the recent and growing access to digital fabrication tools and a commercial shift to user-centered interfaces destined to small scale professionals and hobbyists, certain assumptions remain embedded in CNC machines and their software.

I have spent the last six months learning how to use a three-axis CNC router from the company Shobpot. In the course of these six months, I have become familiar with some basics of subtractive manufacturing and CNC milling: speeds and feeds, roughing and finishing tool paths, raster passes and other area machining strategies, zeroing axes, changing and ordering end mills, specifying and sometimes converting dimensions. I also became familiar with engineering concepts such as machine allowance, tolerance and model boundaries, among others.

I also learned that the Shopbot, as well as the Computer-Aided Manufacturing (CAM) software I used to create my tool paths (Vectric's VCarve) were made for a different kind of user than me.

That user, unlike me, had milling and machining experience. That user most likely knew exactly what they wanted in terms of form factor, understood dimensions and scale, knew about end mills and had probably an idea, even vague, of how the machine was going to move. That user meant business, execution of plans, while I was seeking something less definitive, less tangible: exploration and expression.

There are several reasons why CNC milling remains a tricky fabrication method, even among digital fabrication enthusiasts. One reason is that there are higher stakes in operating a machine that spins a bit of metal very fast. Another is that it is more intuitive to think of how objects are made in terms of adding material (as in 3D printing) rather than remove it. Another is a more fundamental design assumption at the core of CNC machines and CAM softwares: that the user seeks to realize a design that has been developed on the drawing table, not with the machines. The machine only executes the design, the plan.

Yet, Lucy Suchman reminds us in Plans and Situated Actions that plans are inevitably vague: they sketch a general direction but can never account for all the contingencies of real life. In digital fabrication, every step from the importation of a geometry description into CAM to the final artifact is a careful execution of the Computer-aided design (CAD) model. While Suchman's observations derived from her anthropological work with early AI systems, they apply with equal relevance to the design paradigm of digital fabrication.

Embroidery with a CNC router

Inspired by a recent project presented at the ACM conference on Designing Interactive Systems (DIS), I propose to turn the Shopbot into a giant embroidery machine.

The Shopbot PRSalpha. X & Y-axis motors highlighted.

By doing so, I seek not only to divert its initial modality of use and develop a new tool for embroidery but to start outlining a feminist interaction framework for CNC machining that is not rooted in execution but in exploration; not in plans, but in situated actions. The idea would be to generate drilling tool paths and use a magic needle as end effector.

Embroidery machines usually don’t have as much X and Y travel as the Shopbot, which would make it an exciting embroidery machine that can punch a big surface. The Shopbot would initially be limited to one type of embroidery called punch needle embroidery, given the lack of bobbin case/compartiment under the bed.

Standard magic needles seem to come in diameters of 1.3, 1.6 and 2.2mm which, technically, the Shopbot could handle. If we have a collet that accommodates 1/16 end mills (which I think we do), then it should work for standard sized magic/punch needles used for embroidery.

Adapting the Shopbot for embroidery

Of course, using a machine designed for a specific use (in this case, substractive manufacturing) for another (embroidery) requires potentially many adaptations. Here are a few considerations before we dive in:

Z axis speed

Embroidery machines punch very fast, and the Z movements of the Shopbot are not as rapid. Here is a video of a Tajima embroidery machine. Below is a video of the Shopbot on a drilling tool path.

Threading & thread supply

Another consideration is the threading and thread supply. I’ll need to find a way to rig the spool to the Shopbot to ensure continuous thread supply. Since the YZ car and the gantry move, they’re not optimal placement for the bobbin, although the metal brackets that hold the gantry might work fine with well placed guides and proper tension. It’ll require some experimentation.

A sketch of how I could potentially rig the spool.

End effector

In theory, that part could be as easy as adding a magic needle in the right sized collet and never turn on the spindle. But… it might prove more complicated than that. In the plotter embroidery paper, the authors went through many iterations of the punch needle handle, as shown below.

The bed and fabric tension

Yet another consideration is the bed itself. Can we remove the sacrificial layer and, if we can, should we create a giant frame to install on the bed? It would have to be a little elevated to ensure that the needle doesn’t punch into the bed itself (see sketch above). Tension was an issue in the plotter embroidery project so most likely we’ll have to come up with strategies to ensure adequate tension despite the giant size of the bed.

Materials

Another consideration has to do with the materials, specifically thread and fabric. Depending on the fabric and the movements of the Shopbot, it might be that some threads work better than other (in terms of thickness, strength, etc.)

Concluding remarks and timeline

I intend to develop the hardware and software for this project in the course of the spring and summer of 2022, starting in May.

There are many design paradigms I would like to explore through this project, for instance the interaction modality for digital fabrication, which is traditionally (and almost exclusively) GUI or code. What is fabrication was instead driven by more evocative or embodied input such as voice or breath? These ideas might seem wacky, but there was a time when making electronics with soft materials such as thread and fabric seemed equally bizarre. It is the strange ideas we must pursue, as they lead us to new ways of being and making things —which might amount to the same thing.

A photo of a tool path simulation on the Shopbot. Could it embroider instead of cut? Could the software be designed for exploration rather than execution?