Hacking Manufacturing: research on the factory floor

Media Lab researchers spent a month hacking ideas and factory processes in southern China.

By Jifei Ou, Jie Qi, and Artem Dementyev

Since 2013, members of the MIT Media Lab community have set out from their research groups every summer to observe work in ten or so factories in China. This year, we went with a different approach: not just to visit but also to collaborate on experimenting and co-operative design with two facilities in or near the city of Shenzhen.

During the Hacking Manufacturing summer course this August focused on the theme of “Soft Robotic Materials,” we three instructors and seven Lab students worked with a digital knitting factory called K-Tech and a flexible printed circuit board (FPCB) manufacturer called King

Shenzhen has been the home of the Hacking Manufacturing course since its first year. As a manufacturing hub, it’s one of the world’s highest concentrations of factories across many industries. This ecosystem gave us access to diverse knowledge, tools, and raw materials. Also, leaving our comfort zones and living in a new country prompted many new perspectives for us on a personal level.

Media Lab instructors and students with workers at K-Tech. left-right: Zhenyu Yu (K-Tech), Don Derek Haddad, Amos Golan, Jifei Ou, Daniel Oran, Guillermo Bernal, Andy Su (K-Tech), Ben Su (K-Tech), Miguel Perez, Ani Liu, Laya Anasu, and Angela Chen (K-Tech). Credit: Bai Chuan

The overall goal of our Hacking Manufacturing month was to explore how to do academic research on the factory floor. Specifically, we wanted to see new outcomes from using manufacturing machines to prototype, rather than traditional prototyping tools. Usually, when people go to a factory, they want to produce something that has been well-planned and has clear economic value. Instead, we wanted to bring the spirit of a lab to the factory: What if we could collaborate with its workers, directly prototyping and trying out ideas on the very machines that are used for production? What if curiosity were our only pressure point? Benefits would flow — not only from new academic research for publication but also from bringing prototyping closer to production in the ideation phase. By definition, the prototypes would be production-ready.

Collaboration through conversation

Today, many people talk about interdisciplinary collaboration. For that to happen, there must be a shared medium and shared context for the conversation to start. For us, the manufacturing processes create both the medium and the context. We all come from different backgrounds — bio-engineering, mechanical engineering, architecture, industrial design, fine art, neuroscience, to name a few. Yet we all share a love of making stuff, with our hands, with machines, and now with factories.

Circuit assembly line at the King Credie factory. Credit: MIT Media Lab/Artem Dementyev

We started our Shenzhen trip with different priorities: some students wanted to learn more about manufacturing and how to scale their previous research into a product; some wanted to experiment with ideas that are not possible to produce in a lab due to facility constraints; some were excited to use the experience as inspiration for new directions in their research; and some were curious to explore the social and emotional sides of manufacturing. As well, there were those who just wanted to experience a different culture for a month. In the end, we all learned a great deal beyond our expectations; we created many new collaborative connections and made many new friends. We also produced a lot of work that we could develop into a research paper and more.

Meeting our collaborators

Often the factory is depicted as a place run by machines and automation. Humans are invisible in this picture. But that’s not what we saw in the Shenzhen-area factories: Every piece of a product had its maker, its builder — someone who spent time and effort to fine-tune the machine and to work with the machine to achieve the products people use every day. The most exciting part of our trip was that we got to work with those makers. They have accumulated knowledge, expertise, and skills that aren’t common in a lab — wisdom that only comes from working for years with the machinery and materials in a factory. We believed that by collaborating with them, we could create something amazing in a short period.

Media Lab instructors and students with collaborators at the King Credie factory: (l-r) Artem Dementyev and Jie Qi, with King Credie workers Simen Zeng, Bin Yao, Zhiqiang Jia, and Zoro Zeng. Credit: Bai Chuan

So often, cool creations and research come from unexpected discoveries that happen while we play with processes rather than when we simply progress toward a planned goal. We found that to be the case in the Shenzhen factories: Having access to play with industrial-level machines gave us new perspectives on how things are made, and opened new operation spaces for innovation. In the Media Lab, we usually prototype with a laser cutter, a 3D printer, or other personal-scale digital and manual fabrication tools. While we’re able to make some tangible pieces with those tools, they’re not the kind that people typically use in the manufacturing process. The differences in tools create big gaps between outputs of lab-based research and those of factory-scale production.

Zhiqiang Jia (King Credie) and Guillermo Bernal (Media Lab) inspect the flexible printed circuit board just etched in the row of etching stations in the background. Credit: MIT Media Lab/Artem Dementyev

Then, there’s the added benefit of being able to research manufacturing without economic pressures. In this course, we could customize and try crazy ideas with these industrial-level machines in actual production settings, which are usually inaccessible to researchers, without worrying whether the results would immediately turn a profit. This gave us much more freedom to take leaps in design which might fail, rather than smaller, incremental steps toward progress.

Spending a whole month in a factory gave use more than expertise in the manufacturing processes; it also it gave us time to really know the people who work with the machines and find out how the factory operates. We were able to build community and trust with these people. That is, beyond being a source of production capability, the personnel in the factories became our collaborators, so that even now — after we’re no longer physically at the factories — we can continue working and creating together. If we have questions or curiosities to share, our communication channels are open so that we can still pursue these explorations with our factory collaborators.

We selected the Shenzhen-area factories based on the following criteria: first, they’re open to working with researchers; second, they include some degree of manual work that allows for fast modification of the processes because, if factories are fully automated, there’s less room for researchers to change and remix without breaking the machines. While we course instructors were split between the two factories, the students were free to go to either one depending on the nature of their projects.

Rong Fu Zhong (King Credie) and Jie Qi (Media Lab) experimenting with the heat press at the flexible electronics factory. Credit: MIT Media Lab/Artem Dementyev

King Credie

The King Credie factory, in the Bao’an district of Shenzhen, focuses on the final steps of flexible printed circuit board (FPCB) manufacturing: silkscreen, coverlay lamination, and reinforcements. FPCB manufacturing interested us because flexible electronics are popular in human-computer interactions (HCI) research and are featured in multiple HCI projects. This manufacturing process is similar to the usual rigid PCB manufacturing process. The main difference is that, instead of rigid underlying layers of fiberglass (FR-4, for example), King Credie uses a thin polyimide plastic film. Furthermore, the facility does automated circuit board testing and assembly using automated pick-and-place machines. The exposure, etching, and laser cutting is done in a separate location, and we moved between the two. We found some manufacturing steps to be particularly easy and interesting to modify. For example, since coverlay is laminated into the circuit board using a heat press, we discovered that we could sandwich various materials between the coverlay and the circuit, thereby creating flexible printed circuit boards with materials, such as leaves and paper, embedded into them. This discovery became a foundation for our various projects on the trip.

This flowchart shows different steps involved in manufacturing a simple two-layer flexible particle circuit board. Credit: MIT Media Lab/Artem Dementyev

K-Tech

K-Tech is a digital knitting factory in the city of Dongguan, which neighbors Shenzhen. The facility produces knitted fabric from a variety of yarns, using a special piece of computer code. This factory houses about a thousand flatbed computer-controlled knitting machines, mainly for the upper parts of athletic shoes. The machines can knit various types of stitches as well as automatically switch between different types yarns during the knitting process, and the patterns are programmed with a graphical software. The fabrics can be knit into numerous 3D forms, such as pockets, channels, holes, and ribbons, and they can also be shaped into three-dimensional shapes in the post-processing phase.

Jifei Ou (Media Lab) and Jordi Montaner (K-Tech) examine the knitted fabric produced by the machine in the background. Credit: RedZ Studio

We were particularly interested in how K-Tech’s knitting process uses various yarn types: for example, conductive yarn, shape-memory alloy wire, andoptical fiber that can add novel functionality to the fabrics. This variety gave us a wide design space to try out new functional textiles in production-level quality and quantities. We started by learning how to program our own knitting patterns, and then we hacked the process by writing our own computer software to control the knitting machines independently and to test materials that are usually considered too difficult to be machine-handled. Along the way, this created many hacking opportunities to make the machines work with these experimental materials.

Not only were we embedded in two factories, but we were also immersed in the manufacturing ecosystem. Thanks to our borrowed “factory” status, we could order raw materials online in production quantities to be delivered within days to the facility. It was also easy for us to purchase components, sometimes in vast quantities, at the famous Huaqiangbei electronics markets in Shenzhen, among the largest in the world. These public markets span multiple city blocks and sell numerous components, from resistors to complex pick-and-place machines.

It was also cool that the course gave us chances to visit other facilities. For instance, several students in our course were interested in using shape-memory alloy in their projects, so out of curiosity, we went to a nitinol factory to see how these metals were formed. We then learned that even this place, which shapes nitinol into wires, gets its raw materials from metal refineries further north in China. On the other end of the supply chain, we learned about how other factories in the region recycle waste materials — like raw copper board cuttings from the metal and sacrificial particle boards — into gift boxes.

What surprised us

We went to Shenzhen with certain preconceived notions. We didn’t expect to see such cross-collaboration among factories, for instance, because that doesn’t usually happen when factories work in different domains. Typically they look at the manufacturing ecosystem vertically through a single product, rather than horizontally across product domains. After our trip, we managed to bring together processes and materials from the K-Tech knitting factory and King Credie’s flexible PCB factory, which usually would not experience such exchanges. And, we demonstrated that the blend of these two produces new innovations. In essence, we researchers were like bees, helping to cross-pollinate by doing research that combined the manufacturing processes of two factories. That said, it was also surprisingly difficult to bring the two factory groups together, perhaps partially due to our limited time in Shenzhen, to create collaboration without our direct involvement. This is something we hope to facilitate going forward.

Laya Anasu (Media Lab, in the background) watches Angela Chen (K-Tech) demonstrate how to operate the knitting machine. Credit: RedZ studio

We were also surprised to realize how much more we learned by spending sustained time in the factories, and how that has changed our perspectives on the objects around us. Some of us had previously worked with the flex PCB factory, mainly when we ordered materials for projects, such as Circuit Stickers/Chibitronics by Jifei Ou and Jie Qi, as well as SensorTape by Artem Dementyev and Cindy Hsin-Liu Kao. We’d also visited some of these factories before, but more as “tourists,” so we didn’t grasp the processes as thoroughly as we do now after this trip. In fact, our recent factory experiences changed the way we designed. As Simen Zeng, general manager of King Credie commented toward the end of our trip, “I can tell by looking at the designs that now you understand how these boards are produced. There is a much more efficient use of the materials.” Similarly, many of the students in our group had little background in textiles until they spent time in the knitting factory. They started out not knowing the difference between a knit fabric and a woven textile. However, after the trip, now they can “read” how a textile is produced, simply by looking at it.

A pleasant surprise was the turnaround speed we saw: What would usually take days to turn around in a lab, took an hour when we were working inside the factory because we could quickly explain what we needed and come up with solutions to any problems that arose. For assembly, which typically takes weeks, we could pick up reels of electronic components from the Huaqiangbei markets in the morning, then catch a train to the factories, and then hand-deliver our parts to the factory in the afternoon, and have the board assemblies done that night. Granted, this was an especially rushed order due to our limited time at the factory. But it was empowering to see that such an undertaking was even possible.

Ji Qi with a circuit speaker. Credit: Bai Chuan

As we’ve noted a lot in this post, the factory personnel were incredibly helpful: Some even worked overtime with us, to finish our knit samples or even hand-assemble some of the flex boards we designed. Plus, after several weeks, they were making design modifications to our digital files (which added assembly steps and cost more to produce, but they generously waived the costs — that also surprised us) to improve our designs.

We weren’t surprised that language and culture differences could present challenges, more so than technical constraints. Most of our group members don’t speak or understand Mandarin so it was difficult to order materials, especially without access to a local bank account. As well, the language and culture differences often led to misunderstandings and miscommunications, even when everyone was clearly working with the best intentions. Ultimately we found that spending sustained time together really helped us all to understand our different working and communication styles and to build overall trust in our collaboration. It’s great to see that one month’s effort could influence long-term research directions.

What’s next?

To show what we’ve created during the course in Shenzhen, we’re planning a show for the Design Society Museum there in December/January. The exhibition will showcase the physical examples of our Hacking Manufacturing projects along with a video documentary about the trip. As well, some of us from the Hacking Manufacturing course have already written a research paper, which is currently under review, and others are thinking about taking what they’ve achieved to their theses directions. We plan to offer a class in the 2018 Spring semester to share everything we learned with more students from MIT and other universities.

Artem Dementyev (Media Lab) with a circuit robot. Credit: Bai Chuan

At the heart of Hacking Manufacturing is this question: How can we demolish the barriers between a lab and a factory — for a future where design, development, and deployment are tightly coupled? We hope that our model of “researcher-in-residence” can encourage other researchers, designers, and engineers to go to a factory and focus on playing and innovating with processes. We also hope the projects stemming from our trip could give factory owners greater incentive to be more open to play and to look beyond the bottom line. As well, we hope that the Hacking Manufacturing course might become a new model of industry-academic collaboration, where both sides share the perspective that the process, not just the product, can drive innovation.

Then, the next question would be: How can we bring a factory to the lab? After seeing what we did in just one month in Shenzhen, the K-Tech factory owner decided to donate two machines to the Media Lab so that we could continue our Shenzhen projects back in our research groups. This is a great start for us in thinking about how to create a new type of facility on campus — one that could truly blend prototype and production.

Ultimately, we hope to be able to create a network of factories, in which students, researchers, artists, scientists, designers, and engineers could freely create and innovate by devising new processes. A factory should not only be a place where products are being accurately manufactured but it should also a playground where creativity can be fully expressed.

Watch this video wrap-up of the 2017 Hacking Manufacturing course

Media Lab Hacking Manufacturing — 2017 from Tangible Media Group on Vimeo.

This post was written by the course instructors: Jifei Ou, a PhD student in the Tangible Media group; Jie Qi, a postdoctoral associate in the Lifelong Kindergarten group who recently earned her PhD in the Responsive Environments group; and Artem Dementyev, a PhD student in Responsive Environments. The Lab students in the summer 2017 course were Laya Anasu, Guillermo Bernal, Amos Golan, Don Derek Haddad, Ani Liu (a recent graduate), Daniel Oran, and Miguel Perez (a recent graduate).

The Hacking Manufacturing course instructors, students, and factory collaborators share a selfie moment in Shenzhen. Credit: Bai Chuan

More info: Visit this site to find out more about the Hacking Manufacturing course project. You can also read blog posts from previous years and see last year’s video wrapup.

Acknowledgments: Many thanks to the Shenzhen factories K-Tech and King Credie, for sharing their time and expertise, as well as to the Design Society Shenzhen and Hong Kong Design Trust for funding our video documentary and supporting our research efforts. We are also grateful to the MIT Media Lab, especially Joi Ito, Andrew “bunnie” Huang, and Gavin Zhao, our advisors on the Hacking Manufacturing program. Thanks to the students’ Lab research groups for funding support: Tangible Media, Lifelong Kindergarten, Responsive Environments, Design Fiction, Synthetic Neurobiology, Fluid Interfaces, City Science, and Playful Systems. Our gratitude also to Chris Wawrousek, Michelle Dunbar, Seeed Studio, Zhao Ma, Dan Chen, and Sarah Schade.


This post was originally published on the Media Lab website.