OpenFlexure Microscope, an open-source optical microscope | Hacker News Reader

OpenFlexure Microscope, an open-source optical microscope

openflexure.org

113 pointsfybs39 comments

Is it just the my lack of coffee or did they manage to build a website for a microscope without one single example picture taken with that device?

the paper has some pictures https://www.biorxiv.org/content/10.1101/861856v1.full.pdf

That's a fair criticism of the website and one that we had been meaning to address for a while now. We did start collecting pictures over on this Gitlab wiki but most of the pictures out there that are not listed there yet.

https://gitlab.com/openflexure/gallery-wiki#galleries

Isn't that because the focus of the project is on the mechanics of the microscope and not the optics? You could show nice photos even with a non-flexure based mechanism. The advantage is the usability.

The Foldscope Project has been an interesting one-- a low-cost microscope made primarily from paper, offering 140x magnification: https://www.foldscope.com/

"Foldscope Instruments Inc’s mission is to break down the price barrier between people & the curiosity and excitement of scientific exploration!"

A piece of paper and glass for $10? I own $5 pocket microscope from aliexpress with good tube optics (as for $5), bright LED and easy one-hand operation, and it is excellent https://www.aliexpress.com/item/4000574841522.html

It is $10 because the money from individual kits is being used to subsidize the foldscopes for classroom kits.

See "Why the difference in price between Deluxe Individual Kits and Classroom Kits?" in the FAQ: https://www.foldscope.com/faq

I found this 2019 PyCon UK talk a good overview of the microscope.

https://pyvideo.org/pycon-uk-2019/robotic-microscopy-for-eve...

Isn't that a digital microscope? I thought an optical microscope would be one that just uses lenses and eyepieces to magnify and show the images to one or more eyes.

Maybe I've missed something. It looks cool - but I'm not sure I understand what the purpose of OpenFlexure is from reading the website, what does it do that an old fashioned optical/mechanical microscope doesn't (or one with a digital eyepiece)? Is this meant to find things automatically?

The main value of this is a low-cost precision positioning system that can be 3d-printed. A lot of the challenge of microscopy is getting the sample into the right place. And yes, this design can be motorised and controlled automatically as well.

"Optical" when talking about microscopes means it's using visible light rather than, say, electrons.

"Flexture" refers to the type of mechanism that moves the sample stage relative to the objective lens. It means that instead of a high precision rack-and-pinion gear system that moves the stage, it uses "flexure joints" which don't necessarily need to be precision machined (they can be 3D printed).

In a flexture mechanism you apply force to a flexible bar and, because of geometry, that displacement gets translated and reduced into a much smaller displacement somewhere else. It only works with very small displacements (thus fine for microscopy). If you want to position the sample large distances, you just need to move the sample.

It's a neat idea made inexpensive by 3D printing, but the major expense here will be the optics (the objective lens). To get decent pictures of cells, like in a textbook, you're talking about $1K minimum, and it can go much higher, into the 10's, depending on application, performance and other optics (the light source and it's lenses).

Is it possible to use software to post process those images, apply a deconvolution filter and correct the optics?

Might that enable the use of cheaper lenses?

Unfortunately you can't recover information that was never captured, which is the main issue with cheaper lenses. They tend to have more abberations [1], which prevents capturing high quality information.

[1]: https://en.wikipedia.org/wiki/Optical_aberration

What if you shift the lens or sensor and capture multiple images, then process the result, would that be able to overcome chromatic abberation?

Well if you use narrowband colour filters (or light up the sample using narrowband LEDs) and refocus every time you switch, then you could eliminate chromatic aberration completely.

You would have to slightly rescale the image for each colour but that is not impossible.

Sounds like a tradeoff between processing time and image clarity / cost, same as computational photography in smartphones. That looks like a win to me, hopefully we will see innovations in this area so cheap microscopes with great resolving power becomes available to the masses.

Yes.

One problem in amateur astronomy is that people love spending money on expensive gear (like hifi audio nuts) and they think software is hard. So they sneer at technically complex solutions even if they end up cheaper in long run.

I assume microscopy has the same problem. Why make it cheap when you charge thousands of dollars for selling a top optical quality microscope?

I could be wrong. Haven’t tried it yet. But I haven’t given up on the idea yet.

Do all optical aberrations go away if you use a curved image plane?

And if you correct aberrations in software then you can use the cheapest lenses (a single element objective and eyepiece) without having to correct for these aberrations in glass.

Thanks, that makes more sense!

You'd be surprised at what you can accomplish with a couple hundred dollar plan achromat objective. (Though OpenFlexure's design complicates the use of immersion oil...)

Why are these kinds of lenses so expensive?

No, a very good Chinese made RMS objective can cost less than $25 at lower magnification. Plan Achromatic objectives cost a little more, but will give flatter, clearer images than cheaper ones.

It seems a tube lens is used anyway, so the advantages of infinity system objectives are not as obvious. I’d stick to DIN160 or so called Olympus compatible objectives.

Earlier iterations of this were used successfully as part of a water quality monitoring in Tanzania:

https://www.waterscope.org/about-us/

Richard Bowman (previously Cambridge uni, now Bath) was the main driver of these microscopes and is a great, highly intelligent person. Many other people have contributed to the project in small and large ways over the years.

Some other links:

https://3dprint.com/165457/openflexure-3d-printed-microscope...

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7249832/

This is a super cool project. Though I think the website could be a lot more upfront about what is needed to make this and what this can accomplish. A quick list of pats, estimated cost, specs and a sample photo would be really helpful in assessing whether this is a better deal that a $150 used microscope. As it is, it takes three clicks to find the list of parts required and no cost estimates are offered. https://build.openflexure.org/openflexure-microscope/v6.1.5/...

On first glance it appears to be vaporware but there is at least one published image apparently created with the device:

https://openflexure.discourse.group/t/sharing-an-almost-whol...

you are right, I found the site watching this:

https://www.youtube.com/watch?v=9TYlQ4urcg8

- this is the microscope from the link. OpenFlexure Delta Stage 3D Printed Microscope: Hunting for Tardigrades

https://www.youtube.com/watch?v=eATcGuTz6To

- And I really liked this DIY

https://www.instructables.com/Low-cost-Fluorescence-and-Brig...

I've built variants of these years ago for a school. The positioning system works extremely well, especially given the fact that it is 3d-printed; the optics were rather limited, both by lens quality and by the use of the old rasperry pi camera sensor.

I've looked into building a "high resolution optics" version. Most parts [1] are easily sourceable, but the tube lens [2] has proven to be difficult.

It seems half inch diameter, 50mm focal length achromatic doublets are not that common; Thorlabs does seem to be the only source, and are (at least at the moment) not in stock.

[1] https://build.openflexure.org/openflexure-microscope/v6.1.5/...

[2] https://build.openflexure.org/openflexure-microscope/v6.1.5/...

Compatible 50mm achromatic doublets are not that common? That part needs a redesign...

Since you've done the research, I hope you don't mind me asking (I have little to none formal knowledge in optics).

I am trying to build a homemade projector, I was looking for what I think is a 'broadband' (visible light) 'non-polarizing' 50:50 beamsplitter but all the beamsplitters I found are quite expensive. Is there a 'lower quality' hobbyist-grade product range of beamsplitters that you could point me to? (Brands or distributors is also fine)

I'm certainly no expert, my knowledge is limited to what I picked up after a few nights of reading about the microscope.

Does the beamsplitter have to be non-polarizing? My understanding is that standard beamsplitters do the splitting by polarizing the incoming light.

Now that I think of it, you're probably using a standard display as your image source? That does indeed produce polarized light. In that case I would try a (common in photography) circular polarizing filter. Rotate the filter to match the display polarization, after that the light will be circularly polarized. I think you could use that in combination with a "standard" polarizing beamsplitter. Depending on the desired dimensions, there are a few available on aliexpress.

Thanks for the pointers, I had assumed non-polarized was a requirement but indeed didn't consider that display sources are themselves polarized.

Thanks for the firsthand report. It occurs to me that optics have not gone through any of the disruptions that electronics have. It’s still difficult to find good lenses at a student budget price.

I’m genuinly curious: what do you mean by vaporware?

It usually means some software or hardware which is advertised but not available to buy. Here with a few clicks you can find literal printable files, detailed manual on assembly and the software and a scientific paper describing the project. And that is just a quick scan. What makes it vaporware then?

Clearly you can’t buy it, but you can download it right now and start printing it.

I used the terms “appears to” deliberately. There appeared to be no shots of actual output (see other notes on this page), finished builds or builds in progress, and nothing but renders of the product. A little disconcerting. Normally you’d see both on the homepage of a cool project like this one, along with smiling students showing theirs off, etc.

I built one a few months ago for our middle school science lab. It's relatively mature-- a decent mechanical system for the optics, with an OK (simple) optical design, and great build instructions.

The places where it's lacking: the electrical parts for motor drivers are clunky and annoying to build. The software could be a little better. And the transmissive microscopy setup is a bit dodgy.

Some more images here: https://gitlab.com/openflexure/gallery-wiki#galleries

And someone used the Openflexure delta stage to make a scanning laser microscope recently (and made an entertaining video about it): https://old.reddit.com/r/microscopy/comments/mvmznq/i_made_a...

and what magnification can you expect?

Reply on news.ycombinator.com