Machines and Craft

(Fair warning: This post discusses some of the technical process of developing our 3D advent calendar— while this is a light overview, it’s still a bit technical for this blog.)

One of our goals for the advent calendar was to user-test 3D printing. Less than the technology per se, we wanted to find out what we, as designers, might be able to use it for. What happens when non-experts come up with a load of different ideas? Can we use 3D printing as a design tool?

Yes and no.

How it Works – A Quick Primer
People who work with 3D printers talk about developing a ‘toolchain’ – a complete set of software and hardware for the multi-step progression from idea to object. To make a 3D printed object, you need three things: a 3D file, software to translate that file into instructions, and a machine that will implement those instructions to create an object.

When we invited everyone at Moving Brands to create a model, we tried to make sure that people without experience could dive in. We provided tutorials and support for 3DTin, TinkerCad, and Google Sketchup. Moving Branders used all three, as well as Blendr, Cinema4D,  Excel, downloaded files, last-minute requests for technical support, and external commissions.

While most  found the process to be time-consuming, painstaking and not always intuitive, they were able to apply a variety of skills to create a variety of solutions.


Once you have a model, the next step is to translate from a shape on a screen to instructions detailing exactly what the printer needs to do. Here’s where it gets complicated.

The best program for generating these instructions (known as G-code) is Skeinforge. It does the difficult maths of taking an object, slicing it into layers, and describing the exact movements the printer will need to make. The problem is that Skeinforge doesn’t really have a user interface. Its minimal front end offers a list of 30 categories with evocative names like “carve”, “splodge”, and “oozebane”.

Good interface design could make the program more usable, and a number of programs offer front ends built on top of Skeinforge. SFact changes the units and adds a bit more description, so it’s easier to understand, while ReplicatorG provides an all-in-one interface for viewing and modifying models, connecting to the printer, and generating instructions using standard machine profiles.

The problem, though, is that there is no standard machine. Different printers are all a little bit different, especially among the ever-evolving RepRaps. Even standard kits can produce different results if they are put together slightly differently.

So — even though a few settings control most of what you need to sort out — buried somewhere amongst Skeinforge’s ~300 settings could be the difference between a final object that looks good and one that looks like goo. And for now, the best way to find that number is to delve into Skeinforge itself.

Once you have your instructions, you need software that connects to the 3D printer (we used Pronterface, which was minimal but worked well), and then you finally build the object. The printer does this by extruding a thin line of molten plastic while tracing the shape of one layer; it then moves up a little bit and traces out the next layer. Each layer sticks on top of the previous layer, ultimately creating a solid object.

The most interesting aspect of this part of the process was that it moved very quickly away from software and interface issues, and into physical considerations.

We couldn’t just press ‘go’, walk away, and come back to find something that looked like what we’d imagined. We had to start caring about the physical properties of molten plastic and the structural integrity of how layers are placed. We had to become conversant with how the machine was put together and even how it sounded (and smelled), so we could catch things as soon as they began to go awry.

3D printing starts with a computer, but ultimately the process of making something physical requires understanding and paying attention to the messy aspects of the physical world.

On to Chocolate
Once we had a handle on printing with plastic and switched over to chocolate, even though the software tools were basically the same, we found ourselves starting nearly from scratch.

To state the obvious, melted chocolate is a different material from melted plastic. While melted plastic hardens quickly at room temperature, chocolate holds onto its temperature for longer. If the chocolate is too hot, you get a melted mess. Too cold, and it falls apart or  clogs up as you are trying to extrude it. In order to make something with chocolate, we had to become chocolatiers. (Although in the end, we were too afraid to eat any of the results.)

The Upshot
Before getting started, I had imagined 3D printing to be a way to take some of the characteristics of computing, such as scalability, repeatability and quick iteration, and apply it to small-scale manufacture.

The complexity and continuous evolution of the hardware and software are still huge limiting factors for usability and broad appeal, but this will undoubtedly improve over time.

Perhaps more importantly, I was surprised to see how much craft is still necessary – understanding materials and physical processes, attention to detail, developing a ‘feel’ rather than a programmatic understanding of what’s going on. In the end, this was both a limiting factor and a strength – 3D printing isn’t promising to replace the human art of creating, but to provide another avenue for complex expression.