| April 19, 2023
Interview: Plastic scanner
Jerry de Vos developed the plastic scanner during his master thesis, it will be a device that can identify the most common types of plastic, in order to make plastic recycling simpler, more accessible, and more viable!
by the Open make team, and Jerry de Vos. Copyright to the authors, distributed under a CC-BY 4.0 licence.
Interviewee: Jerry de Vos (TU Delft)
Interviewers: Robert Mies (TU Berlin) & Moritz Maxeiner (FU Berlin)
Transcription and editing: Diana Paola Americano Guerrero, Fabio Reh, Robert Mies, Moritz Maxeiner & Julien Colomb
Screenshot of the interview.
The plastic scanner will be a device that can identify the most common types of plastic, in order to make plastic recycling simpler, more accessible, and more viable! All in an open source way.
Jerry de Vos developed the first prototype during his master thesis, and its development continues in collaboration with the TU Delft.
Photos of the plastic scanner, screenshot of video from Jerry de Vos.
- Main website: https://plasticscanner.com
- Project start: 2020
- Core development team size: 2
The plastic scanner is an infrared spectrometer. If you do a scan of a plastic object, it gives you a graph, which will change depending on the type of plastic. By comparing this spectrum to known values, you can hopefully tell what type of plastic it is.
At the moment, it’s a printed circuit board, an Arduino Uno that is connected to the computer and a sensor area. We have schematics of all the different components and how they’re connected together. Further, we have firmware and software.
Prototype stage. There is always the uncertainty of if and how well it will work. We’re at a point where we will not change much on the PCB itself anymore. We are still designing the housing (currently using a spacer, or divider between the sensor and LEDs). If we change it, properties could change. There’s a lot of dependencies on hardware, firmware settings, PCB and software, which are all dependent on each other.
I think at that point, people will hopefully hack it, or adjust it to their needs. There’s one guy in Shanghai who wants to replicate it and we agreed that we would cover all of the material costs and support him as much as we can.
The initial seed for starting the plastic scanner project was me helping out at precious plastic, another open hardware project. I saw the challenges in sorting different types of plastic. I proposed the project to my professor and he said, “That sounds good to research this and to make the graduation project around this.”
First, I did research on what are the different sorting methods and how are these used around the world in small or large scale, talking to people from the precious plastic community and some other independent recyclers. All of them have one form of infrared spectroscopy, I wanted to see if it would be possible to make a DIY version or a simpler version that people either make themselves or have it made at a FabLab or makerspace.
Do you think the plastic scanner would be an interesting project we could focus on? It would be good to focus on one project you’ve been involved in closely.
We would focus on the plastics scanner if you’re flexible because it’s very spontaneous.
Let’s start with the project area and afterwards we broach the issue of product hardware output and participants. How was the project initiated?
While looking into the working principle of this infrared spectroscopy, I found the reremeter project (https://github.com/arminstr/reremeter). I started to contact the author to build upon that work.
I wanted a fully standalone device without the requirements of internet access.
One development was keeping the plastic scanner fully open source, try to incorporate more feedback from the community and based on that build a product like the best possible open source plastic scanner. Another option would be seeing how this product could be in the hands of people that need it as quickly as possible. I chose the first path.
We chose GitHub because it’s well known and relatively easy to use. Since a few weeks we have a Discord server where people can join for questions if there’s something that they don’t understand.
I first work on this project inside my master thesis, then got the James Dyson award came with prize money (35k€) that I wanted to use to develop the plastic scanner.
How much money was that?
I think it can be very interesting to make either a small series we can sell or maybe like a Kickstarter where people can just order it. If you buy 100 or 500 of LEDs, it’s around a quarter of the price of single units. Further we think about something for calibrating.
I think at that point, people will hopefully hack it or adjust it to their needs which is something that we wouldn’t be able to do if we make hundreds of them.
Have you major issues you came across during the project and how did you resolve them?
Could you classify these different modules in terms of mechanical, electrical and software you developed?
Since you have a great documentation, are you aware of anyone having successfully tried to replicated your project?
The project for me started out as my Master’s thesis and with that came my thesis report. There is a paper that Armin wrote about his first prototype and why it works (I think it was in the process of being published somewhere).
Have you published the findings in other ways than the documentation on GitHub, like a journal or something?
Why did you choose GitHub? Was it easy to use it or were there any barriers?
Do you think it is possible to reproduce and modify the plastic scanner from what you’ve published, even if you don’t have much feedback?
Where are things you didn’t publish? Where do you see difficulties?
At what maturity are the different parts of your hardware? Are the firmware, software and hardware a prototype, demonstrator or market ready?
It’s super interesting to see all of the different people that reach out like companies who want to improve.
What was was successful about this project and what wasn’t?
The project isn’t successful in being able to delegate tasks or parts of the projects to specific people. There’s a lot of communication involved, and a lot of waiting for each other.
We always need to find a balance between high tech, nice solution and low tech easy to implement solutions.
Since I work at the TU Delft, I don’t pay myself for the project. Jure gets paid by the hours that he actually spent on the different tasks.
One guy read it in the newspaper and wanted to help. He has been his whole life into spectroscopy and gave feedback on our design choices. In one afternoon, people can tell as much as you could learn in years finding out or learning about.
What’s the background and the occupations of the people your working with?
Are you still looking for these persons?
How did you find the people you work with? They have very valuable content competences.
we have a Monday morning meeting where we go through the week. We make a plan and based on that we assigned tasks. We have a place where we put all the tasks, assign a person to it, assign in which area it is, the estimated hours and the actual hours.
How did you coordinate the work between the members of the project?
People like to get a purpose and a specific project that they can work on.
The benefit for citizen scientists would be to be mentioned in videos or include the project on one’s own resume. For me, it’s being associated with this project and doing the things that I like: inventing stuff. That’s very beneficial.
As a conclusion, could you maybe say how the different members in your project have benefited from working on it?