Interview: OpenFlexure microscope

| January 1, 2023

Interview: OpenFlexure microsocpe

Julian Stirling has been a core developer of the OpenFlexure project, which makes high precision mechanical positioning available to anyone with a 3D printer - for use in microscopes, micromanipulators, and more.

by the Open make team and Julian Stirling. Copyright to the authors, distributed under a CC-BY 4.0 licence.

Sections:

Interviewee: Julian Stirling

Interviewers: Robert Mies (TU Berlin) & Moritz Maxeiner (FU Berlin)

Transcription and editing: Diana Paola Americano Guerrero, Robert Mies, Moritz Maxeiner & Julien Colomb

Banner image: derived from an icon from the Gathering for Open Hardware Community (CC-BY) and photo from Julian Stirling.

Screenshot of the interview.

Screenshot of the interview.

The OpenFlexure microscope

The OpenFlexure project makes high precision mechanical positioning available to anyone with a 3D printer - for use in microscopes, micromanipulators, and more.

Its core development has been mainly organised by Julian Stirling in the UK, in cooperation with a group in Tanzania.

The OpenFlexure microscope in a nutshell

Image of a built microscope, by Samuel McDermott, CC-BY-SA 4.0

Hardware products

The OpenFlexure Microscope is a customisable, open-source optical microscope, using either low-cost webcam optics or lab quality, RMS threaded microscope objectives. It uses an inverted geometry and has a high quality mechanical stage which can be motorised using low-cost geared stepper motors.

Hardware maturity

The microscope is about 80 to 90 percent of the way to product ready. In terms of performance, its product ready. We’re now doing clinical evaluations to check whether it is good enough to diagnose malaria.

Rebuilds

The device has been used and built on every continent, including Antarctica, and in at least 40 countries. 250 people participated in the forum, but it is difficult to know how many built the microscope. We assume that there are thousands of microscopes out there, but we don’t know for sure.

Lab Maker, a company in Germany, has just started selling the parts kit, not the printed parts, just the fasteners.

The project

Project process

We’re very interested in medical microscopy, where to diagnose malaria, you need to look at 100 fields of view with a x100 objective and count the parasites. So it needs to be automated and reproducible.

Rather than looking at how professional metal motorized microscope are made, and building a cheap imitation, the whole thing was designed for being 3D printed from the ground up.

We are now moving on to the stage of improving the software and making it more robust as we get towards production.

The aim has always been around local manufacturing and to be maintainable : it’s not just about building it in Tanzania, it’s about being able to make the parts and service on site.

How did it all start with the OpenFlexure Microscope project?

Show answer

Project start

The project started as a hobby project for Richard. We got a research project which is how I became employed about two years later. Also on that research project, there was some sort of wider team around some other low-cost instruments, but the main OpenFlexure team was me and Richard and a group in Tanzania.

Could you describe the overall process to us?

Show answer

Funding

We’ve kept funding going or I wouldn’t be here.

There have been a series of different grants which are maybe not such big grants. Our first couple of big grants were from something called the Global Challenges Research Fund.

It’s generally been government funding. We’ve looked at a few smaller charitable grants, but it’s very hard to make charitable grants work inside the university, because at universities it costs a lot to employ someone as there are so many extra costs.

Did you keep the project going through lots of smaller grants? You said it was done three years ago? Is the project finished, then?

Show answer

Hardware importance

My main focus is the microscope. But, for instance, a lot of my time is spent is developing a program for writing our documentation. We had a malaria project, which was using (and modifying) the microscope.

Are those like government grants, or are they like from independent entities like GOSH or where do these grants generally come from?

Show answer

Major issues

I think one of the big problems is that we just don’t have the adequate hardware design tooling. For instance, we’ve always done our CAD in OpenSCAD. It does mean that we get some members of the community which can’t contribute because OpenSCAD is hard to learn. We’re basically using software engineering processes to manage the hardware design.

Then, there’s a lot of problems with how we do documentation. I think documentation is just hard. But, there are some tools to help manage that: being able to openly share a design, have people track how it changes over time, have people contribute. Trying to keep the entire process open as you move into production, it means that you start inventing new processes and writing new tools.

What major issues overall have you come across then? And how did you resolve them?

Show answer

The hardware

Hardware components

The microscope is mechanical and electrical. In terms of it’s a motorized microscope, it’s digital. The mechanical parts of it, there’s quite a few. But most of the mechanics is in one single part, the main body. Then there’s peripherals like, where we’ve got some gears and somewhere to mount the illumination and somewhere to mount the objective and a little base.

For software we run a client server architecture where on the microscope is a server that deals with all of the interaction with all the different pieces of hardware. That has an API.

We’re hoping, as we move towards microscope version seven, there will be something that can be sold as a product by Lab Maker and definitely others.

What products have you developed now, in terms of like, different components? Mechanical, electrical, and software? Could you briefly outline that?

Show answer

And has this system been built or produced by others independently?

Show answer

You see all these mods and apps. This is great. This is without any supply of kits, is this completely independent?

Show answer

Are you in contact with Labmaker? Are you assisting them in some capacity?

Show answer

What is the state of the art in terms of maturity, technology-readiness? Is it a demonstrator already, or market ready products? And since Lab Maker is selling it, where would you situate it, roughly?

Show answer

That’s a demonstrator for sure.

Show answer

Local production

One of the biggest problems, if you look at medical devices in Africa, is that about 70% of them are non-functioning, because people can’t get spare parts and company-certified technicians to fix them. So, the aim has always been around local manufacturing and to be maintainable.

We’re never going to plan to be ISO 1345 certified in the university. You have to worry about if you’re setting up a new type of colonialism where at any day we can pull the plug and say: “It’s still ours because it’s our certification.”

We’re very keen for BTech in Tanzania not to be reliant on us. We need to make sure to provide a good enough dossier of how everything works, so that whoever is going to be the legal manufacturer can write a good technical file and convince regulators that they understand it and can maintain it for the future.

We are looking for a good way to transfer control in a sustainable way, such that they can eventually maintain it when our funding runs out and we disappear.

That means you are next going for certification in the some midterm future?

Show answer

Research outputs

Academic outputs

For impact, the way the university sees the transfer of knowledge into a company as impact, and they’re sort of pump-priming for long-term impact.

We’ve got papers on the microscope, papers on the Delta stage version of the microscope, papers on the block stage version of the similar design that’s x, y, z translator for aligning optical fibers. We’ve started publishing on individual parts of it, such as the robustness of the autofocus algorithm, publishing on the software, on how we do the camera calibration.

Let’s go on the actual research output. Obviously, this was as part of several grants and it’s a research project or multiple projects. What was the output of this hardware development so far, directly or indirectly, academically and practically?

Show answer

Publication strategy

Because we do everything in (public) git repositories, I don’t really have anything of the microscope that isn’t shared. A lot of what I do is probably not the best thing for getting me lots of papers and a permanent position. I would get more papers if I would just moved onto the next thing rather than documenting the current thing (so everyone can reproduce it perfectly). And it is very hard to write a paper on a small update.

We’re getting more and more what I call sort of meta papers. We had one, where we took all of our thinking about how we’ve carefully considered medical devices being open, and how you track all of that knowledge.

What information have you shared mainly, on the documentation? Like, Bill of Materials, you mentioned, CAD files, or in this case, notice that files but other code, and so on?

Show answer

Work Coordination

We do our project management using the Gitlab issue tracker. If Richard or I want to change something, we make a merge request. And we have a discussion in the open about it. All of that project management is open.

Most people won’t look at the git repository, they’ll look at what is generated from the repository automatically. We are a bunch of nerds who are basically programmers doing a hardware project. We are using automated pipeline (GitLab Continuous integration) for the production of the documentation website.

I can tell you, by having set up CI infrastructure in our lab, it’s a lot of work to get it right. This is this is something quite advanced in terms of using Git. Have you seen other groups do this where you got motivated?

Show answer

That’s very interesting. Then you have basically this output documentation which people can generate from there and they can use it to to rebuild the hardware. Did I get that right?

Show answer

And was it easy? Or what were the barriers to do this so far?

Show answer

Successes and failures

Now we’re at the point where there are very few projects of our size with our community. The core team sort of grows and shrinks based on grants, but community members are still very active. So, it is very hard to estimate how big the project is.

I think there’s a lot of people who now have microscopes that wouldn’t. There’s a lot of people who have things that they could customize.

I think it’s very hard to get that level of different interactions from most academic projects if you’re not open. If you wanted to have as many collaborators as we have, you would have only time to fill in the NDAs by now.

Participants

Core team and community

It’s a reasonably big project, it’s not huge, like Linux kernel development, but it’s big. In total there must be 50 to 100 people who sent changes to at least one of the repositories.

About three or four people worked on it full time. More when we had more time of our partners in Tanzania who were sort of quite actively involved. It is going to be around 10, but maybe not all of them show up in the Git repository.

Our group is mostly researchers and we have other researchers at other universities that use it and will comment. But then we’ve had people who are doctors, medical doctors who’ve got interested in it from a medical perspective and fed back a lot of information.

The question arises, of course, how did you end up working on that project? And how many people are working on that project?

Show answer

Absolutely, I wouldn’t compare the two anyway. Can you say a little bit about the occupations, at least from the people you know or see? Are there doctoral researchers, postdocs, technicians, students, professors?

Show answer

How did you find suitable project members with the competences to work on the project?

Show answer

Personal gain

In terms of personal growth there will definitely be some, there’s lots of skills you pick up doing anything. In particular, working with a much broader set of people, in a large collaboration with humanitarian goals, has definitely broadened my horizons.

If I tell people who am I, I’m still basically a physicist and engineer. I’m quite unique that I have all of those technical skills which I had beforehand but then all of this experience working in different places. I think that’s what I’ve got out of the project. I think it’s very hard to get that level of different interactions from most academic projects if you’re not open.

How have you and all these people who have contributed the project members benefited from their work in the project? Can you say something about that?

Show answer

That’s a nice analogy. And other people, how have they benefitted?

Show answer

At the end we have a closing question that’s just within the scope of the question before. Did you have any students who did their bachelor or master or something similar thesis within the scope of this project?

Show answer