UROŠ updates: DIY microscopy revisited

Cool! I’m following the activities from distance. Keep posting!

Currently I do not have much time at hand though. :stuck_out_tongue:

Out of interest, what microscopes are there? Also optical? And in the case of the fluorescence ones, do you use some fluorophores?

How can I join the call on Friday?

If people want to join our calls, follow the calendar in the wiki.

All the online interactions are on our BigBlueButton. We regularly meet on Mondays, 15:00 CEST:

1 Like

Did anybody once made an overview and comparison of different webcams, pi-cams, checking their performance, prize, availability for low-cost microscopy?

Let’s start one here as wiki:

Cheap webcams (3$)

PS3 Eye cam (mbe outdated)

Logitech C270 (30-35$)

CMOS chip Sony IMX335 (25-40$)

Raspi cam V1 / V2?

1 Like

A few years ago we (with @ffederici) did a sort of review of cameras an optics to look at yeast for our Brewermicro project


Hi all
on friday we (with @pcremades and @naikymen) had a great workshop about DIY soil microscopy

Here you can find some more videos and pictures we took
We also collaborated with @julianchollet and @dusjagr on a small zine/guide to soil microscopy (english and spanish) that maybe is useful to some else.

We will be talking a little bit more about our UROS remote residence in Mendoza during the next GOSH community call


Having fun in the wood workshop, i just made a new microscopy stage design, with a vintage look and a flat broad sample surface withoug anything obstructing to place samples of whatever size.
Also changed the classic “screw and rubberband” method for focusing the stage, to threaded M4 inserts that can be glued into the wood and a turnable screw to adjust the level of the sample stage by slightly tilting.

As mentioned above. We have explored some new cameras, to make more quality images to be used in labs (unlike the cheap ones that are super good enough for workshops).
Experiences with the IMX335 are sofar amazing. Good colours, framerate and 5Mpixel seems to be the way to go.

What I want to develop now is a 3d printed case to fit the camera pcb with a dedicated lens holder, that would allow to slide / switch between 2 different lenses. Thus changing the magnification while looking at the same spot.

Anybody wants to join to design such a thing?

This filter switcher modules are also intersting…


I’ll continue the discussion with @adrianMolecule here, from the idea posted on the thread for the Experiment funding.

Although i really appreciate all the work and innovative designs of the openflexure project, we have to keep in mind that at the very core of the openflexure is digially controlled and precise automation of a microscopy stage, it’s not about building a low-cost microscope. And in the real world only very few applications need automation…

On the optics side and the ccd camera chip, the same applies to every design of a DIY microscope.

For keeping the costs down, we can even skip the raspi, thus switching to an ESP32, and use already owned smart phones as display for the microscopy image streamed via wifi. Recorded images can be stored in higher resolution on an SD card and later retreived through a website.

I personally never use stage automation, and i am happy to move around my samples by hand to find stuff to look at. also z-focus can be easily done by a simple mechanism, as with every standard microscope, just turning a wheel.

although i would be interested in using some inventions from the flexure for purely automated z-focus. leaving the x-y out of the game.

1 Like

That is a nice clarification.

So you don’t consider OF as a really low cost and you want to do a Lower Cost one replacing PI with ESP and part of the opticals with a common digital camera.

Indeed, OF was about 200$ as base model before PI prices spiked up due to pandemic and auto manufacturing swallowing all production. I consider the 200$ in a high range yet still reasonable low cost for doing research. Also OF is an inverted microscope. I have building an OF on On my back-burner list.

One thing I was not fully convinced by OF was having to build the entire optical part and not using a common cell phone. It increases the price tag quite a lot.

I also agree that in many cases automation is unnecessary yet in some cases I felt the need of an interpreted programming language with a decent IDE which ESP is lacking.

If we go manual focus and use a phone we can concentrate on the mechanical and optical parts, do the casting to a monitor directly from the phone and save the money for ESP.

With an App we could also do simple image processing like contrast control and edge enhancement or more complex like measuring colonies etc if desired. Testing will require some funds for growing/collecting specimens etc.

It’s important to find what would be the common use cases for such a microscope in advance. The “if you build it they will come is not always” true.


Depends on your goal. As this states above it is DIY microscopy. The OpenFlexure goal was malaria diagnosis.

The issue with phones is twofold in the OpenFlexure case. One you need either need a spare phone or you need to take the phone out every time you get a call/leave the room (something that makes no sense for an automated microscope). The second issue was that if you are working towards something highly reliable for medical use you need to be very sure of the performance of you optical system, and have control of the processing. Each smart phone has different sensors, different lenses, different processing. The was not optimal for our application.

It is possible to build and openflexure using the board of a webcam. This way you need a laptop next to it. And the webcam lens. This drops the prices way way down.

There are many takes on this and it depends on personal optics.

Personally, when I set to do something like that I look at what is the lowest price I can achieve research grade results so make it competitive with commercial products. I don’t do strictly education products and I don’t do soon-to -be-CS-for-profit products.

Of course, there is always a trade-off. In my case if I need, I sacrifice convenience like having to use a personal cell and sometimes time. There are many medical microscopy project using a cell. For instance CellScope originating in Berkley.

Also, considering the life span of cell phones it’s easy to recycle some older ones to become permanent microscopes. Two birds with one stone. An old generation decommissioned cell it’s easier to find than a web cam as web cams seem to have gone out of fashion.

I love OF as a project and I consider it to be a torch bearer for OS projects. It’s amazing that among the overwhelming number of OS projects paid by the public and then close sourced and hijacked for profit (CellScope from Berkley, Foldoscope etc), Open Flexure remains open and still thriving.

I also think OF has grown well beyond malaria and that’s a great thing. And automation, while many times unnecessary is sometimes the way to go so kudos for OF.


Agree totally.
To clarify, a low-cost microscope with the ESP32 and cam, allows us to have a defined camera module of high-quality, with reliable manufacturers and datasheet. The phone (or laptop) is only used as a screen to view the live images provided as a website and interfaced with a browser.

and… being low-cost and DIY, most probably NOT towards human diagnostics, but generally as a microscope to watch whatever you want.

In that case it would be similar to OF. Is there any other saving except fro swapping PI with ESP? And, I saw several live video streaming projects with PI like OctoPI. Is there any know project that can do the same with ESP32?

There are many new ESP32 integrated with camera modules, and finished software to stream it to its’ own hosted webserver to be access by ip adress. they usually already come shipped with that installed.

but for more specific uses as a microscope, this can be tailored and improved easily for someone who can do software.

to clarify again different aspects of the “system” of a microscope, i always enjoy to look at this image.


  • (the wine glas) we need something that holds the sample
  • (the candle) we need a light source and some way of focusing the light onto the sample
  • (the box) we need to position the sample (thats what openflexure is good at and can be automated)
  • (the microscope optics) thats a bunch of lenses in a black tube.
  • (the srew on the optics) we need to be able to focus precisely
  • (the pillow) we need a comfortable way to “look” at the sample
  • filters are missing here… or maybe the are part of that lens that focus on the sample?

Additionally in the 21st century, we will digitally be able to save the image files and share them online.


The matchbox scope of Vittorio et al is interesting in this case. They used an ESP-CAM, similar to how webcam’s were used by Hacteria.


The illumination problem is also solved very cleverly.

yes. it totally was the inspiration to follow their path. @VSaggiomo is also on this forum

1 Like

Well, if I read it right they use 2 processors - one on the camera module that seems to be at 2 MP and a “SEEED Studio Xiao Sense (UC2-fueld) board” which I’m not familiar with but it seems to run Python. I’m afraid If you cost both you get to the price of a PI so no significant savings.

I think open flexure uses an 8 MPixel cam so much better than a 2 MP.

I know ESP32 has 2 cores so it would be interesting to see if both the camera and the code that’s running on “SEEED Studio Xiao Sense (UC2-fueld) board” can be run on one ESP32. Unfortunately, while I did ESP32 programming I don’t have a good practical understanding of the multi core programming in ESP32.

Does anybody on this forum have a decent experience with that ESP32 feature? If so I would be happy to hear from you.


It’s running on the XIAO exp32 sense. yes, it’s only 14$ including a camera.
But what i said above, you can use a variety of different camera’s up to your wishes.
Another thing is, that microscopy is limited by the optics. not by the resolution of the camera. an 8M camera just gives you more “information” about how blurry it all is down there…

and again… we can use the same cams and optics exactly like the openflexure, just without the flexure, but make a much more simple stage design, that doesnt need 3 days of printing, motors, etc…

i think we first should just rebuild this amazing version with z-focus…

1 Like

Random question @dusjagr where’s that cool microscope image you shared from???

it’s from one of the earliest woman scientist / microscopist, the honorary mrs. Mary Ward,…

we dedicated a diy microscopy stage to here too:

the original book can be viewed here:


Just catching up here, cool! Looks like an inspiring woman scientist, and I see that the image is from 1869, so almost certainly in the public domain! :+1:

Thanks for the info.

1 Like