Print your own laboratory-grade microscope for US$18 | Hacker News Reader

Print your own laboratory-grade microscope for US$18

phys.org

100 pointssamizdis25 comments

Print? Forget 3d printing, if you want a cheep medical microscope you can 2d-print and the fold one out of paper.

https://en.wikipedia.org/wiki/Foldscope

Foldscopes illustrate an interesting reality when it comes medical testing: more often than not a general-purpose medical microscope is total overkill. For medical testing, seeing if some pathogen is present or not, you do not need massive optics creating large fields of view. Answers can be found using extreemly small optics of, literally, disposable microscopes. The problems of foldscopes largely surround issue of contamination and testing methodology, not the limitations of the optics.

Does FOV help improve the accuracy?

That depends on the test. If you want to count the number of something, a larger area (the sample size) is better. But if you are looking for the simple presence of something, say a parasite, then seeing only one of them is all you need. It's like checking a dog for ticks. To say "no ticks" you have to check the entire dog. But if you check one spot and see one tick you can safely say that the dog has ticks.

This isn't really comparable on any dimension. The foldscope uses a tiny ball lens that introduces a ton of distortion and even its ability to detect disease is pretty questionable (that said I really think it's got a bunch of clever ideas).

Note that this scope really isn't as cheap as described; the printer and the labor involved massively outweight the dollar cost of the components.

It's good for rough pathology of certain infectious diseases and fluid and tissue conditions. Also, you're throwing around a strawman about fixed costs that are amortized in volume. Plus, robotic manufacturing and packing can produce these in massive quantities that drive the net unit costs down.

I've played with similar microscope prototypes IRL and they were quite impressive for something seemingly primitive at first, they were quite usable and adjustable.

I saw laser-cut prototypes of something similar at Stanford around the med school at some sort of "science fair" ~6 years ago. The cost is mostly the magnifying bead. What's also great, besides being cheap, is that it doesn't need power, packs down quite small, and can work in the field.

Hey! I'm an author on the OpenFlexure paper. The Foldscope is awesome. We know the project well, but we fundamentally have very different aims. The foldscope is a fantastic simple microscope at rock-bottom cost.

Our design is focused on providing automated, accurate positioning, hence the titular "flexure" design. For pathology applications described in the paper, being able to scan huge samples at high magnification is really important, but relies on motorized positioning and autofocus to work well. That's where our design is really unique.

I've built several scopes by hand and looked at various projects; this one is far and away the best combination of design, price, and features that I've seen. It's not 'perfect' (if you really care about that, go to a top lab and build your own scope from scratch), but given the limits of what's available, it's pretty damn impressive.

> I've built several scopes by hand and looked at various projects

It would be cool to see comparison between such DIY microscopes.

JFTR, IBM rolled out own version of "low-cost" (~ $300) DIY motorized microscope based on Lego + RPi.[0]

[0] https://github.com/IBM/MicroscoPy

The problem is that so few people do this, and needs are so specialized, that head-to-head comparisons are really hard.

I never understand why people build things like this out of legos. They really not a good substrate for microscope parts. I know it's fun and they're easy to get, but the other parts of the scope are so demanding that you might as well go the full way and design/print 3D plastic (solid body) or use aluminum.

> I never understand why people build things like this out of legos.

Think, Lego chosen because it is really good for prototyping & and widely used for education.

And according such projects like MicroscoPy (or any other Lego-based project) main target is not create "production ready thing", but instead learn how things could be invented using basic building blocks.

Hey, co-author of the paper here. Thanks so much for the feedback! That comment is a real poster-piece for me!

Very cool but misleading title - you need to bring your own optics.

That's seems to be the case most 3d printed contraptions. Extrude a framework, and the populate it with a bunch of traditionally manufactured parts, but then call it a revolution.

It's a disappointing pattern, but somewhat understandable given the state of the technology.

Came here to write this as well. I mean, I was hoping this wasn't one of those cases, because then it would be a true revolution. But alas, it's another "print a different frame for the professionally manufactured thing".

Agreed -- feels like a real waste of time after I Google'd the costs for the optics ($100 minimum if I did it right) and other parts (Raspberry Pi, etc), then Google'd what a pre-built way-nicer-than-I'd-need microscope goes for ($157 shipped at https://www.amscope.com/40x-2500x-advanced-home-school-compo...)...

Definitely seems to fall in the "3d printing is cool" bucket more than "look how much money I can save" bucket.

The big difference here is the stage is automatic, and it supports multiple types of illumination.

I've built high-end 3D microscopes using aluminum extrusion and 3d printed parts before (with commercially produced objectives, but everything else pretty much sourced from places like Adafruit and OpenBuilds). I got pretty good results and contributed the hardware drivers to MicroManager (an open source tool to drive many different scopes with the same UI).

In retrospect, I would have instead found a way to get a Nikon Ti scope body and accessorize that. It's hard to replicate the high quality and full ecosystem of the Ti for anything less than the list price of a Ti.

Seems to imply you also need to add electronics and motor(s).

Isn't this just a holder that keeps the objective lens at the desired distance from the sample/camera?

And if you can buy the optical parts for cheap on Alibaba, then why not buy the holder there as well?

One of the benefits of the design being open-source, according to the linked site, is that it can be adapted locally to meet end users requirements, in this case labs in Tanzania and Kenya.

That would be more difficult with pre-manufactured parts.

The project name is the answer to your question:

https://en.m.wikipedia.org/wiki/Flexure

From https://openflexure.org/projects/microscope/

> Optomechanics is a crucial part of any microscope; when working at high magnification, it is absolutely crucial to keep the sample steady and to be able to bring it into focus precisely. Accurate motion control is extremely difficult using printed mechanical parts, as good linear motion typically requires tight tolerances and a smooth surface finish.

> This design for a 3D printed microscope stage uses plastic flexures, meaning its motion is free from friction and vibration. It achieves steps well below 100 nanometers when driven with miniature stepper motors, and is stable to within a few microns over several days.

Hey, paper author here. You can build a simple microscope like that. However, for many applications, including the Malaria diagnosis work we discuss in the paper, you need to be able to do high-magnification scans of huge sample regions. This needs automated movement and autofocus to work nicely. The unique part of our design is the low cost, ~50 nanometer precision automated sample positioning. As far as I'm aware, no other low-cost open-source design provides that kind of mechanical positioning system.

https://openflexure.org/projects/microscope/

Somewhat related... it makes me wonder whether 'old' zoom camera lens - which you can readily find for a fraction of their original price at thrift stores - could they be disassembled and their high quality glass lens used in something like this?

The current trend not to include the name of or a link to the article subject is annoying and should stop.

Reply on news.ycombinator.com