Nadya Peek, researcher at MIT’s Center for Bits and Atoms in Boston, is working on developing new digital fabrication machines for the Machines that Make program. Interview.
São Paulo, from our correspondent
At the Center for Bits and Atoms at the Massachusetts Institute of Technology, directed by Neil Gershenfeld, a program aims to develop open source, low-cost computer numerical control (CNC) machines. Nadya Peek, PhD candidate and research assistant at Machines that Make, explains why MIT wants to rise above the infrastructural restrictions of the old industrial model and is looking for solutions in China, especially with FAB12 in Shenzhen this summer.
What is the Machines that Make (MTM) program?
It’s a directory where students of Neil Gershenfeld’s How to make something that makes (almost) anything course document the plans, materials and files of the machines they’ve made. Many students have contributed to it over the years in the spirit of sharing. Historically, digital fabrication originated from industries where CNC machines were developed for very specific applications. We’re interested in personal digital fabrication: How can we make it easier for people to make complex things, with the precision of digital fabrication, and with small tools that are made in small quantities? The CNC machines that are currently on the market are simply smaller and cheaper than industrial ones. In order to operate, modify and improve them, you need a lot of skills, you need to be able to deal with bugs and often poorly designed interfaces, which means not everybody can use them.
Is it about democratizing digital fabrication?
How can we transform the industrial infrastructure of CNC machine development to meet demands on a personal scale? We can talk about democratizing, but I don’t think it means that much. Democratizing this kind of project means making it simpler, but not necessarily accessible to everyone. The problem is not that the machines are poorly made, but that the components and their infrastructure are developed by historical players according to the demands of the industry rather than those of personal digital fabrication. MTM’s objective is to change this logic.
What are you working on concretely?
With Ilan Moyer and James Coleman, we developed a 5-axis machine based on the distributed control model. In order to build a machine like a 3D printer, you need at least three motors, one extruder and something to measure temperature. So we need to develop a system that controls all that. We’re going to design the electronic card, the axes, program the firmware and develop the interface that will control the electronic card. What happens when we decide to add a fourth motor? We have to redesign the whole control system. To make this process easier, I worked on a networked control system. Every time we want to add something, such as a motor, all we need to do is plug a control card into the network instead of having to redesign all the electronic cards.
Machines that Make, Slashbot metal prototype:
Nevertheless, we weren’t very satisfied with the structure, which worked, but not optimally. We implemented the networked controllers, but we needed a system that would be just as fast to set up for the mechanical part. So we turned to modular machines. Each axis is a modular part, we can pile up as many axes as we want. The first plastic version wasn’t stiff enough. By the sixth version, we decided on metal sheets. They worked great. However, not everyone has access to metal sheet cutters, whereas our objective was to allow other people to replicate these modular machines.
“The whole point of a modular machine is that you can rapidly prototype rapid prototyping machines. If nobody has access to this machine, there’s no point…”
How did you resolve the issue?
We opted to make it out of cardboard, so that as many people as possible can replicate it. Using cardboard is controversial, it’s not very sturdy, if you spill water on it, it dissolves, but cardboard still has the advantage of being a versatile material. You can glue it, attach aluminum sheets to make it stiffer, anyone can modify it with a pair of scissors or a cutter, and it’s very cheap.
We wanted to make the construction process extremely easy, so that someone who had never built a machine before could do it in half a day. When you work on a project, the machine is just a tool for rapid prototyping. Spending a year to build this tool can get very discouraging.
“We believe in an agile design method that implies quick, successive tests.”
We make one project work within a day, we see what the problems are with this first version, we adjust and we redo the design until we’re satisfied. With cardboard, we can use a laser cutter. It cuts fast, with high precision and is accessible in a fablab. With our system, it’s very easy to build and modify each axis.
I’ve given a number of workshops for this modular machine, especially for the How to make (almost) anything course and the Fab Academy. Many people have built this machine, sometimes for very amusing projects! We’ve seen machines to make Japanese zen gardens, to make coffee, cocktails, and even to cut cucumbers. You can do very serious and much less serious projects, that’s what’s exciting!
Step-by-step documentation to build the Cardboard CNC:
Neil Gershenfeld often talks about a roadmap for fablabs to develop machines that can build other machines. Is this project part of that vision?
I don’t think we’ll replace all the fablab machines with cardboard machines. But I’m working on making it as easy as possible to create machines for rapid prototyping. Making these machines a reality is part of the work objective of tomorrow’s fablabs.
One of the problems is documentation. You almost have to spend as much time documenting a project as doing it. I feel that the modular machines project is well documented, but I still get e-mails from very upset people who can’t duplicate it, who say that it’s not explained well enough, etc. We tend to think that people who will replicate an open source project have the same skills, but that’s not how it works. Through these projects, some people also want to acquire new skills.
In the academic world, the problem is that open source documentation of projects is not at all appreciated… I should spend my time writing research papers in order to be published. But good documentation is not a trivial matter, it’s even essential for “physical” projects. The code allows us all to have similar tools on our screen, to share them and reproduce them more simply. I hope that we’ll find a way to collaborate massively on these projects, which isn’t yet the case.
The FAB12 global meeting of fablabs will be held this year in Shenzhen, China. Will one of the objectives of this meeting be to develop these fablabs 2.0?
These days I spend a lot of time in China, where I buy lots of motors, electronic boards and other components. Instead of buying machines in China, in Japan, or in any other country where they’re produced, people could be buying components and making their own machines in fablabs, for fablabs. That’s one of the objectives of FAB12, to get the plans for the machines of fablabs 2.0 ready for the conference. Neil Gershenfeld always says that we’ll be ready for the 2016 event, but I’m the person working on it and I don’t want to promise too much…
The Chinese government, and especially the representatives of the city of Shenzhen, see the maker movement as a good thing. Makers come and not only generate economic activity around the manufacturing of pieces in small quantities, but also things that didn’t exist before. Digital marketplaces such as Taobao offer very easy access to components, and it has become easy for makers in the West to buy 10 motors in Shenzhen and have them shipped.
I was introduced to the company AQS in Shenzhen by Bunnie Huang [MIT alumnus famous for hacking Xbox, who now lives in Singapore]. When I was working on these modular machines at the Fab Academy, I wanted everyone to have the same motors, so that there wouldn’t be a problem in case one breaks. I ordered a motor from them, specifying a number of details, a very specific auger, bolt, etc. They let me order a smaller quantity of 10 pieces at $30 each. In the U.S., just the auger itself costs $30. They accepted to produce 10 pieces with the assumption that I would order more, and I ended up reordering 800. But by industry standards, 800 pieces is nothing. Unlike five years ago, today you can find factories that will produce small quantities. The small-scale issue is central to creating personal machines. And currently, China is the only place where you can access this type of production.
More info on Modular Machines that Make: Cardboard CNC