I make my own Microscope.

creator: x14km2d | build: 2021-11-06 | update: 2022-01-05


I like to be inspired by other projects and I found a lot about USB microscopes at Instructables. I could use one for my own projects and by the way my roommate had an old one lying around. After a short conversation, she sold it to me for a small price. There are many people who think that hardware hackers always need to create something new. This is only partly true, because hackers also learn by recreating projects. I like to learn e.g. by haptic perception. I have to touch everything, turn and so on. When I read through a tutorial I don’t have any problems with the content, but it’s less fun. And since I like to learn when I have fun doing it, I build things.

In addition, our culture is based on the copying of learned processes, technologies or developments. When we are children we copy our parents, when we get older and as teenagers we have to re-orient ourselves every day and we copy young adults from our circle of friends and that goes on all our lives. Someone has seen that it is worth breaking stones to develop a knife like object. The household knife you use every day to peel your apples is a copy of millions of different knives, all based on the original stone knife. When I copy a USB microscope, it is still an in-house development, because it is based on the mother of all microscopes of the Nimrud lens. Copying something is naturally good, because you don’t just show respect for the developer or the idea, you also transport the idea further into the next generation. Good knowledge continues to spread and thus remains a part of humanity. At the age of 8 I was in the Neanderthal Museum and learned how to recognize flints and how to make a stone tool from them in a very slow and difficult process. I carry this knowledge within me and can retrieve it as needed and pass it on to the next generation. You can never be sure when you may need this knowledge.

I also like to optimize ideas. Here you can work a little more with the image quality, at this point you might prefer to use metal instead of plastic. There is no perfect development. What may be incredible these days may seem ridiculous to us in 50 years. Just look at the history of the video game console. Do you remember tapes? I grew up with it. It’s all a matter of perspective. To test if the idea works at all I simply held a lens of an old digital camera in front of the webcam and carried out first tests. I examined the circuit board of an e-reader and a flower from my garden. Even if the pictures are still blurry, you can already see many more than with the naked eye.

Version One

Our main components consist of the lens of an old digital camera and an old webcam. Otherwise we only need the tools listed below. I advise against breaking a digital camera, but to look in the cellar or ask friends first. Sometimes you can also buy a cheap one at the flea market. The same goes for the webcam, where you can also order a cheap one. Webcams are things that everyone had lying around at home in my time, even though most of them were of questionable quality. You can see this later on the photos below in the final test.

First we have to separate the superfluous housing from the board. This is best done with a carpetknife and pliers. In most cases, the housing should be screwed, which can then be unscrewed witha screwdriver. Only asshole companies headquartered in Redmond, Washington, who don’t want to repair webcams themselves, glue them completely during industrial production.

The flexible holder of the webcam is only held inside the housing with a metal pin. We can simplypull it out with tweezers, as it is not glued or screwed. We put the parts aside and pack them into abox after work, where we store them in a safe place. You can use something like that again and again later and you should get used to throwing as little as possible into the trash can while hacking the hardware. This not only saves money, but also protects the environment. Then we have to loosen the black GND cable with a screwdriver. A small piece of metal falls off, which we also store. With our first prototypes we will not connect this cable anywhere yet, but will implement it in later versions. May the electric gods forgive us.

The complete board of the webcam is held by two small screws. One in the upper right and one in the upper left corner. We solve these so that we can continue working. Now we come to a place where we need to be a little more careful. Unfortunately you can’t disconnect the cable from the rear housing without breaking one of them. It would be possible to solder the cable again, but it costs additional work that we want to save. Therefore we cut the rear housing with scissors, sharp pliers. Since the connecting piece of the cable is also glued to the plastic we cut cleanly along the edge. In the first version of our USB microscopes we don’t need a microphone yet, so we separate it from the board with scissors. If we want to solder on a (better) microphone later on, this is no problem, because the solder points are not damaged. Since we should clean the lenses before we stick everything together, we have to fold a toothpick into a glasses cleaning cloth as we see on the photos below. With our fingers we would not reach the small lens and the toothpick helps us to clean the corners properly. This is how we improve the quality of our photos right from the start.

It’s time to glue our lens (with a second lens from the digital camera) to our board. I put the superglue in the place with a toothpick where the button for triggering the camera is located. Idon’t need it anyway, because I create the photos with the Cheese software. The button should not be visible later anyway. If you want to do this differently, you don’t have to stick to my guidelines, of course, but develop your own idea of how you want to solve the problem. I have often read in forums that people towering over it quickly clog up the tubes of cheap superglue. I have a few things to say about that. After this tutorial I will use superglue only in case of emergency, because it attacks and destroys the plastic very strongly. If I have a blocked tube, I simply unscrew the lid and insert a toothpick. Then I take out a tropical glue and apply it to the area that is to be glued. It is therefore not necessary to buy a new tube immediately.

With a plastic clamp the parts are held together as long as they dry. Please do not use metalclamps, as this can damage the board. The clips can be bought in any supermarket and are not expensive and only cost a few euro cents. After everything has dried thoroughly for at least one hour, our construction is screwed ontoanother component from our old digital camera. We kept the screws as a precaution and can use them again now. Now we can already do the first test. For this purpose I took the circuit boards from my e-reader project as an object of investigation. Of course, everything is possible. In my opinion, PCBs are the most suitable because they have many details that can be examined. You can also take something from the garden, e.g. small stones or something. What you can’t so easily see with the naked eye.

The first test went quite well, but I was not satisfied. I could improve the image quality of the lenses, but I didn’t take into account that they also need external light. With a webcam that depicts a person, it doesn’t matter, because there is enough external light. But if you now seal a lens with an (almost) light-tight plastic ring, you can see very little. I had then tried to light it with a blue LED, but quickly realized that this light source is much too bright for such a simple camera. Nevertheless you could see something on a big monitor and that’s the basic idea of the project. I want to display small things in big.

To solve the problem with the missing light I had to remove the black plastic ring from the previous construction and look for something else. I took my time and thought about it for three days until I had a sensible idea. Couldn’t you just use the neck of a bottle of lemonade as a ring? The transparent plastic could allow enough light to penetrate the microscope and significantly improve the image quality. Of course I couldn’t use the microscope at night, but that’s a secondary problem you can deal with in another version at the moment. I measured the width of the ring and bought a matching iced tea bottle in the supermarket1.

The best way to separate the ring from the body is to use a hacksaw (or a similar fine saw), sawing directly under the small plastic projection. To make this easier for me, I have clamped everything in a parallel vice. But it works quite well without it. You just have to be careful not to slip off with the saw, but you should always pay attention to that anyway. Since I don’t have a professional tool for questioning edges yet, I simply use a carpet knife. This is not quite so clean, but should be enough at the moment to be able to work with it reasonably. On the two parts of our lens there are some superfluous pieces that we have to break off carefully.We’ll use pliers for that. Everything that stands upright must be removed and prevent the plastic ring from being properly glued on. First I wanted to screw the ring to my construction (hence the small black dots). Unfortunately I noticed that although I have a very fine drill, I don’t have a tensioner in the right size for my multi functional device. Therefore I had to fix everything again with the impractical superglue.

The final test brought the expected results. I could see a lot of small details on the examined boards and this will make my work a lot easier. This will help me in troubleshooting for broken solder joints or traces, for example, if I want to develop my own circuit board. The labels are also really easy to see.


All in all, I was really surprised what you can get out with so little cheap technology. I’ve seen other tutorials, but either the quality was further below what I had achieved. If the quality was higher, much better lenses (e.g. from digital cameras) were used, but I can’t afford them at the moment. I’m already thinking about how I can improve the USB microscope. But first I want to spend the next days in the forest or park to collect things that I can put under the microscope.

Version Two

During the last weeks I used the first version of my USB Microscope several times. I noticed somedetails that disturbed me. For example, I could only examine objects with a certain height, becauseI had made the transparent plastic ring for PCBs. As you can see on the photo below, I wanted toexamine a dead bee. I couldn’t use the microscope with the ring because it has different planes.The ring also limited the size of the objects I could examine. After I had removed the ring, I couldadjust different heights with my hand, but unfortunately it wobbled then with video recordings.Altogether I was very dissatisfied with the first version and started to implement version two. Inthis article I describe my approach, what kind of materials I used and which changes were introduced during the implementation of the project.

I am also concerned that the development of a prototype can be very simple. I think it’s important to sensitize beginners to the fact that you can get all materials from your own household. Whether this is from an old computer or external hard disks. We all have old devices somewhere that can be recycled. Something that also bothers me a lot is that a lot of people use hot glue for their great ideas and projects. This not only looks bad in many cases, but also shows how little effort you put into it. So I’ll also be showing techniques that work without superglue or hot glue. This makes your project look much more professional.

Except for the tools and the iron-on beds I didn’t buy any material this time. In the year 2018 I noticed that I ordered quite a lot of materials on the internet to realize projects, but I stored everything in the warehouse. For example, I still have solar cells for the Cyberdeck, two Arduino Uno, three Raspberry Pi, and components for a complete CNC plotter. In addition people have monitors, DVBT Reciever, DVD Player and what else I know. An old german Mark VII console, half a Apple Newton MessagePad 130, an E-Reader that needs a new housing and much more. I have enough material for lots projects and have to work on them before I buy new material. It’s like my Steam account. I buy new games all the time although I haven’t played through the old ones yet.

Two weeks ago I was in a One Dollar Store and bought some parts and small stuff for my projects. By chance I found the iron-on beds shown above. I used them as a child in kindergarten, but didn’t notice this product later. That turned out to be a mistake because you can use them for so manyapplications. For example, if you want to screw a CNC shield on a wooden plate, but you don’t want to put the PCB on the wood. Then you can use the iron-on beds as spacers. This is quite practical. Since I was just working on another idea, I put a white iron-on bed on a small metal rod from an old DVD drive and it fitted like a glove.

So that my USB microscope in the second version looks a little more like a professional product, I found a small metal frame. I can’t remember exactly where I got it from. Maybe the frame is from an old external hard drive. From an old printer I found small white rolls that can be used as feet for the frame. These also fit perfectly. A printer is a perfect source for components, you get rollers, threaded rods, motors and a large amount of gears.

Since the wheels already fitted to the metal frame, I only had to find a solution how to put the DVD poles with the wheels together. I didn’t want to use superglue or hot glue as I already mentioned above.

To make sure that I really like the idea with the rolls, I have equipped the frame with four pieces. I had to clean them first, because they were rubbed with grease in the old printer. For this I used a solution of glass cleaner and water.

I measured the iron-on beds with a digital caliper gauge. The outside diameter is 4.8mm. Since the components are made of a very soft plastic you can cheat a little and it is possible to get 0.1mm by warming it by hand.

Since I spontaneously had no better idea how to get the inside diameter of the rolls out, I simply put different drill inserts into the hole and came to the result “3”. So I had to file out 1.8mm. This is fortunately quite simple with plastic and does not pose any problems.

Here I have once again put the two components in front of each other for size comparison. Maybe some people already have an idea of what kind of plan I’m pursuing.

At my first attempt (I prepared the roll with a file. But that took too long, so I tried my multifunction device. It was faster, but also less clean. In the last step I used a drill with a 4 size drill insert. I hoped that the metal rod would need 0.1-02 mm more space and that I would have the right size again. This is what the two interconnected components look like. Since roll, iron-on beds and metal rod all press against each other, no components slip. It’s easier than expected and works really well.

With this pole the iron-on bed has warped, because I polished out dirty (that was my first try). Ma can see exactly how the iron-on bed takes on the other form. I have to test if the beds can be soldered together, maybe with a hot air device. The deformation is no longer visible from the outside and since this is a let’s show prototype, it is not at the top of the list of priorities. It has to look good and show the technical weaknesses. So e.g. these questions should be answered. Can it hold at all with the small rolls or does it wobble too much?, Can you connect components to one another without adhesives, so that they hold even if you use them more often? or Does the design fit with a microscope? It’s about testing what’s possible. So we will see later that an idea was not so good and was exchanged.

Here the whole construction in a side view. We need four of them so that our base can stand safely.

When inserting the metal rod I noticed that the small tube inside the frame still has a thread. I had to drill it out with a metal drill bit. I clamped the workpiece in a parallel vice. I clamped the cardboard on one side to prevent the vice from scratching the metal. So I can clamp the frame tighter and don’t slip off by mistake with the drill.

The rod fits incredibly well into the hole and is very easy to remove. You rarely have such suitable components at hand or have to search for them in your boxes for much longer.

Here I have already put the frame around properly and you can already see that everything fits.

Here we see that everything fits together perfectly. But since I already have some experience with my projects, I often leave them on my desk for a night. Shortly before falling asleep, I let all work steps run in my mind’s eye. I noticed that it’s rather bad with a frame if you want to build a microscope. It is better to use a different base. Also with another color, so that e.g. PCBs, insects or seeds stand out better from the background. I had to revise my first idea and find a better solution.

Sometimes it is better to first create a paper model to implement an idea. You are on the safe side and save expensive material, if something should not look as good as you imagined. For most of my projects I use white cardboard (160 g/m2) which doesn’t cost much and is available in a well- assorted paper shop. Cardboard is also strong enough to be used as a professional case. To create a very simple model, we take all the poles out of the metal frame and place them on the cardboard as a template.

With a sharp pencil we trace the contour of the frame.

When we have completely recorded the frame, we cut it out with a model knife and a metal ruler. If necessary, we can also use scissors, but you can’t work so cleanly with them. I also need a tool to cut out circles better, but I haven’t found anything good yet.

I made the holes for the rods with a paper punch. It’s a bit difficult to determine the exact position, but you can ignore that because it’s only enough for a short test.

Since the cardboard is not strong enough to hold the four rods, I used the metal frame again. It was exactly what I had imagined. Now all I had to do was find a white material (preferably plastic) that I could process easily. By chance I had a quick solution. A few weeks ago a cheap laptop stand made of plastic broke down. I had stored it in a box for broken materials.

As with paper, we can simply transfer the outline onto the plastic. Normally I use a CD marker or another pen with permanent paint. Fortunately I could use my pencil.

After we have drawn the pattern, we saw out the pattern roughly. I used a metal saw that worked best. Of course you can also try other saws. Since there are different types of plastic, I can’t give an exact estimate now, but only speak for my application case.

So that I can saw better, I fixed the plastic to my work table with screw clamps. This prevents me from slipping and from accidentally cutting into my hand. It also helps to work more precisely.

A tip I’ve learned over the years. It is better not to saw out everything perfectly immediately, but to approach the drawn line. With many small steps you reach your goal better.

The technique on the next pictures I taught myself also by several failed attempts. We tighten the metal frame with the plastic on a wooden base. This should not be your table, because we have to drill through the plastic.

Here I drill the holes, into which our metal rods will be inserted later. With plastic I always use a wood drill.

After I had drilled the holes, I looked for all screws with nuts from my tool case and screwed the metal frame with the plastic. This way you can saw cleanly and you don’t cut yourself, because the metal protects the plastic. A very simple trick, but it has a big effect.

Here you can see in a side view how I inserted the screws. Since it was a very soft plastic I could continue after sawing with the carpet knife. With it you can work very clean and the cut surface becomes straight. So you don’t have to grind so much later and you save a little time at work.

Normally I always use sandpaper with a block of wood, but this time I tried something different. In the drug store I bought files for little money, which are actually used for the fingernails. These are very good for working accurately and curves are not as difficult as before. I will use them more often in the future, because I made very good experiences with this project with it.

Here you can see the metal frame and the copy next to each other. Of course I am not satisfied with the result, because it is too inaccurate for me. I have very high demands on my own work and always have to hold myself back with a prototype, because otherwise I lose myself in details. Here’s another tip. Don’t put too much work into an idea, leave perfection to the machines. A CNC miller/laser will always be able to do this better and if you don’t have any machines of your own, just check the internet. There are hundreds of suppliers offering these services. I have done this myself so far and wasted too much important time. Therefore I will use this project to test a service provider. Of course I will report about it in my next articles.


Here I use the drill holes so that I can push the metal rods better into the plastic. If you can’t do it by hand you can do it carefully with a hammer.

Here I inserted the first pole and everything fits perfectly.

To secure the bar, I used an iron-on bed again. At this point you can also see that the edge of the plastic has not become perfectly straight.

From the top the whole construction looks great and I am approaching a real prototype. This also shows that you can work professionally with very few purchased materials. But I must honestly admit that this has a lot to do with experience and many failed projects. So you don’t have to be annoyed if it doesn’t work the first time. The best approach is to implement and practice a lot of ideas.

Even though I’m not quite satisfied with some details yet, the base still looks quite good. Now we have to take care of the lens from the microscope in the next section. All in all, everything is going in the right direction.

Since I will use a new lens in the next versions, I didn’t want to invest so much time and material here anymore. I took three small curved rods and tapped them. With this you could screw them into the small holes in the plastic. This is a good method if you don’t want to use glue. Before, the outer diameter of the metal bar was as wide as the inner diameter of the plastic. But with the new thread we can screw in the screw, because plastic is softer than metal and is therefore reshaped.

I noticed that elementary os own webcam software records the images mirror-inverted, so I used cheesy webcam booth. But I don’t like the size of the pictures, which I would like to convert to 1000 x 562 px. I will also look here for a suitable software. But since the hardware will be changed in the next versions, this will happen later. I tested one of the small rod with thread (3mm) and a Microship on the white base.

Overall I am satisfied with this version. Everything could be implemented quite well, although not everything was built perfectly. But I can already say that the microscope will get a bigger update from the lens. The white base is great and makes the hardware look like a medical device. I also like the shape. It’s minimalistic but not boring. Of course you could make everything rectangular and save material during production, but that would look really cheap. I also like to integrate some round elements into the construction. What bothers me a bit is that the plate is already very tightly connected to the rods. For example, if you want to change the plate quickly, don’t give it that easy. Actually there should be a quick clamping system. For the next version I plan to fix hardware under the plate, but I suspect that the plate is a little too thin.

Version Three

In the last weeks I have invested a lot of time in my USB microscope project and slowly you can talk about a real prototype. I replaced the old webcam with a Raspberry Pi Zero and the microscope now has several layers. I also learned a lot from mistakes in this version. I mentioned this in another article sometime, but you really only learn when you make mistakes. Of course I plan all steps theoretically, but you have to work on the experience yourself. Somehow this is turning into something serious. Actually I just wanted to build a microscope for myself and document the way there, because I do that with all my projects. With this documentation it is a little different, because with every new version I see something that I absolutely want to improve. I also pay attention to cost-effectiveness, i.e. whether the modules/components can be produced cost-effectively. But that would still be a long way to a real release, because the market for USB microscopes is very competitive and there are a lot of (and professional) products.

In this version some more expensive components were used. I wanted to have the plates for the individual layers created by an online service, but decided against it for several reasons. On the one hand I still had enough material from the white PVC and did not have to produce now extra new garbage, but rather use old material first. The supplier didn’t have the right thickness for the plates, which really bothered me. It would also have cost me money, which I would rather invest in a final version, i.e. a professional prototype. This made this version a little warped, because I could correct mistakes I had made in the plates only with great difficulty. After this version I wonder if I should use metal instead of plastic. Later it should be produced cheaply and environmentally friendly. Quite a tough task.

As I have already mentioned above, I will also point out my mistakes. This will help people to learn more about what they have to pay attention to in their own projects. So you should measure screws over and over again on a new project day before you screw them into components. You shouldn’t rely on old memories. You can make a mistake quickly, as we can see right away. On the photo you can see how I hold the drill insert next to the plate. With this I try to estimate how far I can drill. To set a mark for me, I taped the rest of the drill with tape. There are also special rings with locking screws, unfortunately I didn’t have them in the suitable M1 size.

I wanted to fix the Raspberry Pi Zero on the plate underneath, I have simply drawn the holes of the board with a pencil on the white plane. As you can see here the connector for the flex cable is on the left side. We recognize this side by the slight slant, which points to the outside. On a car this would be the bonnet. So that I do not get confused with the project I have decided to orient myself to this side. It is on the left side, so that the cables of the Raspberry Pi (HDMI, mains plug, USB) point to the back. The flex cable goes out to the right, to the trunk side. In order to be able to distinguish all levels better, I called the lowest plate slave layer. In the further course of the documentation I will only call this layer slave layer. On the left it is called bonnet side and on the right trunk side. With the names you can remember the orientation better and you don’t get confused so fast if I count many details at once.

Since I needed eight screws which I didn’t want to buy extra now, I looked in my screw box whether I could find suitable there. I also sorted them. At first only by color (black/copper/silver). If I have more time I sort them by brand or function.

After I found the right screws, I put iron-on beds under the board of the Raspberry Pi Zero again and screwed in the screws. I made my first rookie mistake.

Unfortunately I was not concentrated for a few minutes and had forgotten to measure the length of the screws before screwing them in. I still had the length of the second version in my head. Anyway I didn’t notice how the screws drilled through the plastic on the other side. I had already suspected before that 3mm in the thickness of the slave plane could be too little, but wanted to try it anyway. Even if this meant more work for me now, I find these little insights exciting. This allows you to see how a prototype develops and how errors are corrected bit by bit.

Overall, the idea was already quite good. The external HDMI plug adapts to the shape of the plates, which I liked very much. Also the height was almost perfect to my mental picture.

Here you can see on the photo how I tightened the right screw much too tight. This is a problem with my body, which works rather gross motorically. So I could never become a heart surgeon or watchmaker. I try to pay attention to this again and again, but between my brain and my hand signals sometimes simply disappear in nirvana.

Here you see how the connector plug for the ribbon cable is still on the bonnet side. At this point I really noticed that I don’t like it that way and I have to change the sides.

While looking at the page (with the HDMI plug) I noticed another irregularity. The HDMI connector is 2-3mm deeper than the small column on the right side. That doesn’t sound like much now, but I always work very hard and it has to fit together very well. If I now place the construction on one layer, the connector bends a little, which in turn causes the component on the board to bend (very little). However, this minimal irregularity is sufficient to additionally load the component and thus promote wear. This allows the connector to break more quickly and nobody wants to buy a new Raspberry Pi Zero all the time just because the builder has slammed the construction.

From a diagonal perspective, everything looks good. That’s why it’s important to look at your prototypes from all sides and different perspectives so you can also find hidden flaws.

Everything looks good from the top, too. You can still see the iron-on beds on the poles. I left them out later because they are very hard to get off again. Although I want to make everything as safe and durable as possible, it should also be easy to disassemble. The USB microscope is to be divided into individual modules that can be quickly exchanged. It should also be possible to pack the microscope quickly, for example when traveling in inaccessible areas. I am thinking of bio- and archaeologists. Another target group should be students or kids. If you set up the microscope at the beginning of the lesson, don’t waste too much time before you can do the first examinations.

As we saw, there were some mistakes and problems that I have not seen before. That’s quite normal, because you don’t develop a perfect prototype the first time. Most inventors like Alexander Graham Bell, Nikola Tesla or Guglielmo Marconi needed several attempts to develop working prototypes. They had to deal with difficult technical questions and many of these inventors unfortunately failed until someone else rediscovered their ideas and developed them to maturity. To save me time and unnecessary work, I made many drawings and went through the construction in my head. This sometimes allows me to sort out some stupid ideas beforehand. It is helpful to draw the construction from all perspectives. And yes, I saw the mistake in the drawing but ignored it before I took a picture because I didn’t have a suitable pen within reach.

Since the bottom level should become 6mm thick, so that I can attach the Raspberry Pi Zero better, since I did not have material in this thickness available, I had to improvise there a little bit. I cut out two 3mm pieces in the right size.

In order to join these two panels together, I decided to use wood glue. I have enough of this in stock and it doesn’t matter if I use a little more glue than necessary. Of course you can also use other adhesives, unfortunately I haven’t tested it and can’t tell you how the material reacts. There are e.g. strong industrial adhesives that can soften plastics. This happened to me as a very small child with my Lego bricks, because I wanted to stick a house together permanently. So you have to test for yourself what works well. Wood glue is ok, but not perfect. I remember reading somewhere that the two surfaces stick better if you sand them with a piece of sandpaper before, but I have no idea where or if it’s just a fake memory.

After I coated both plates with sufficient wood glue, these were fixed with clamps. Since I wanted to make sure that everything was glued well, I put the material on the windowsill for 24 hours. I put the heater up so that everything dries out faster.

A technical process that I developed in the second version is the joining of the master and slave level. With this I can saw, carve or grind without removing too much material. Since the master is made of metal, I don’t have to worry about accidentally making a mistake.

This time I had to use thicker material (6mm (slave level) + 3mm (master level) = total 9mm), I bought extra new nuts and a M2 (2mm) threaded rod (1m) and sawed with a hacksaw into suitable pieces. The material was not expensive now, but I could be sure that everything fits exactly.

This is the underside with the screws. As you can see at the ends of the threaded rods, I only sawed them out very roughly. Nothing has to be perfect, just hold and press the plates together.

In the second version of the USB microscope I already described in detail how I worked while sawing out. Therefore there is only one single photo of an intermediate step without documenting the whole process.

In between I noticed that after a few days a bar has massively distorted. The idea with the iron-on bods within the small roles was apparently not so good and I will look for another solution in the next versions. The rolls weren’t completely filled, which I found practical because it reduces the weight of the whole construction. Sometimes these are also points you have to consider. How much weight does a component have and can you reduce the total weight without compromising stability? This question is extremely important when the project is scaled. Nobody is interested in the weight of four small rolls (you should be interested), because what happens when production is increased to 5,000,000,000 units. 5 million parts weigh a little more and that affects transport costs. In addition one should pay attention to the fact that components are always produced Bi (2). 2, 4, 6, 8 is better to produce than 1, 3, 5 and 7. I will explain this in more detail in future versions. Just imagine that your prototype has to be transported to the International Space Station (ISS) with a rocket and you have to pay 1000€ for every single milligram. These aren’t realistic numbers now, of course, but large numbers.

Here you can see the drawing for the construction of the Raspberry Pi Zero. In order to be able to work and plan more exactly, I have dimensioned all thicknesses of the plates etc.. As you can see the total height is 9mm and only for the slave layer. So you can imagine how fast you can reach a really big height, if you don’t reduce a few mm in between. Since I don’t have a wind cutter for sizes above M4 yet, I had to prepare the drill holes with the screws. I simply screwed them in and out with a screwdriver. This creates a simple thread in the drill hole and the screw can grip better when screwing in.

Since I had shortened the iron-on beds and I wanted to avoid turning the screws through the slave plane this time, I shortened them with a pair of pliers. That was about 2mm. If you have screws of a suitable length, you can do without this intermediate step.

In the next step we edit the camera layer. It should be adapted to the other shapes, but in the middle it should carry the camera. In the first version I didn’t come up with a specific style for the two mounts, but simply implemented it. First I have to test if the idea works at all, because the functionality comes before the design.

In order to be able to position the board of the camera exactly, we measure the center. The total width of the board is 2.4 mm. Thus, there is a dimension of 1.2 mm on each side.

I drew the dimensions with a pencil. The areas with the diagonal lines are sawn out. As you can see, I’ve listed myself a few times, but that’s ok. That’s why we’d rather use a pencil than a permanent marker.

Sometimes it is difficult to saw out surfaces within an area. Since I don’t have a special tool at the moment, I had to resort to a little trick here. We take the biggest drill bit and drill a lot of holes in the areas we’ll be sawing later. Please take care to place a wooden board or other old support under the layer so that you don’t drill into your working stitch.

In the next step we pinch the pieces between the holes with pliers. Please work carefully, because plastic can tear quickly and damage the whole component.

Now we have to work on the edges again with a file and sandpaper. The rough pieces were cut off with a carpet knife. Since the material is easy to work with, I was finished within 20 minutes. As you can see from the photo, it’s not quite perfect yet, but it’s enough for a prototype.

Since I revised the idea with the iron-on beds I had to think of something new. So I found the fixing screws (with the rings) in a model shop. These are a little expensive, but they gave me more ideas. For the first moment these are enough, but as we can already see these small black screws are much too small and can easily get lost. In a laboratory this is no problem if you work on a clean white table. But if you work in the wild with the USB microscope, it can get lost very quickly. In my shop it didn’t take two minutes for the first screw to say goodbye. To show a size comparison, I put the locking screw on a bit for a screwdriver.

The rings with locking screws are very practical and can be used in projects where you need spacers that can be moved quickly. I also didn’t think they would fit so well on the rods from the DVD player. That means that they are standardized and you don’t have to worry about the width. That’s a big advantage compared to other concepts.

I did not fix the object layer with the iron-on beds. The first time I had to take them off and that was incredibly impractical and exhausting and was in line with my idea of an easy to use USB microscope. That’s something you have to learn first. Good ideas to question whether they really work. I’ve seen people in you so often wanting to keep something in the project that just turned out to be a bad idea but was from project managers (this happens very often in software development). In many cases they couldn’t let go. Unfortunately, you can’t publish a first-class prototype on the market like this. At the moment I’m still testing all this myself and haven’t even gotten the user feedback. It may be that the complete design has to be reworked because I overlooked important points. But I’m getting to these important points right now, which had already surprised me personally. Just have a look at this video of Adam Savage’s One Day Builds: NASA Spacesuit Parts! and you can imagine how exhausting it is sometimes to have to start from scratch.

Here you can see the camera, how I fixed it to the layer with screws. The shape is still angular and not round, but this will change later. Since I didn’t have any screws in the right size, I had to reach for my screw glass again. When screwing in the screws, make sure that no components are broken off the circuit board. Maybe I should use electrostatic adhesive tape in a later version.

The bottom of the camera. As you can see, the screws are too long and could cause injuries when using the microscope. Here I have to find a better solution, because I don’t want to have it in a finished product.

After everything was installed, I could also put on the master layer and the design looked like I had imagined it to be.

So that the surfaces of my tables do not scratch so fast, I attached some floor mats to the slave layer. At the moment I only had angular ones, but they can be replaced later by round ones. Now let’s get to the part that didn’t work so well because I made some mistakes. When I bought the camera for the Raspberry Pi, a blue/white ribbon cable was included. Since I’m usually one of the early birds buyers on new projects, that was a long time ago. At that time the cable was still used for the first versions of the Raspberry Pi, which I also used here, but I didn’t want to use it for the USB microscope because of its size. Anyway, the cable didn’t fit into the connector of the Pi Zero. So I had to buy a new one. Since there are no reasonable hardware stores in my city, I had to register at Amazon and order the cable. Shortly before Christmas it took an infinite amount of time and sometimes I feel the urge to move to Shenzhen. I hate having to leave my house, but I don’t like it even more when I have to wait for hardware. Since I practice haptic perception (only with objects, not with people), I want to touch things before I buy something. If I have to order online, it creates stress for me. Anyway, after a few days the new cable arrived and I was able to plug it in…to see my second error.

After I plugged in the orange cable, a lot of me thought it was twisted. I had installed the raspberry Pi Zero with the wrong side on the slave layer. So I couldn’t leave it alone, because after a few days it could lead to a cable break and I didn’t want that. I had to disassemble the whole USB microscope and turn the Raspberry Pi Zero to the other side. I still noticed that the HDMI plug was no longer suitable and I had to rearrange the board. For this new holes were drawn and drilled. It stressed me a little and pulls the morale down. Nevertheless I continue to work and could reassemble everything completely with it.

Altogether I learned a lot in this version. I’ll start with the points that didn’t go so well. The deep drilled screws were a typical mistake for beginners and I was not concentrated that day. This allowed me to create work that I could have saved myself. That was stupid. But since I know myself, this mistake will remain very much in my memory and protect me from similar stupid ideas in the future. Then there was the point with the wrong cable. With a real project, that would have taken a lot of time and the deadline would have had to be postponed, which in turn would have cost money for materials, management, etc., which would have been better to invest in important things. In the future I have to get used to completely assembling hardware before I start with the rest of the design. So I can be sure that all components really fit together and no important parts are missing. Because to have to reorder something is pure stress, because then you have to put the project aside although you are motivated to reach your milestone. With a previous assembly I would have noticed how I have to install the Raspberry Pi on the slave layer the right way round. All in all, I did some work that I could have avoided. That was very stupid, but it takes me further for the future.

What also struck me was that the microscope takes far too long to be assembled and disassembled. Especially the object layer caused problems because I had adjusted the diameter of the lace to the millimeter with the diameter of the rods. I used some oil for precision mechanics and watchmakers, but plastic and metal don’t seem to work well together. I have already thought about having later versions made of metal, which is also more environmentally friendly. Also the stability is increased, but unfortunately also the weight. In addition, the surface of the object layer would have to be painted so that the examination material remains visible under the microscope. At the moment there are far too many factors that need to be considered and I would rather concentrate on the concrete development.

What really pleases me is that the project is progressing. I have hundreds of ideas that I would like to implement right now. Unfortunately, I don’t have to wait very long. I like the layering, because I want the microscope to be as mudular as possible. people should be able to adapt the hardware to their needs within minutes, regardless of whether they’re biologists, hardware hackers or students. I have the feeling that the overall concept is already going in the right direction. Unfortunately, so far everything is still crooked and crooked, but these are technical weaknesses that appear in a hobby prototype. I also lack better tools in some places, which I just noticed with this version. In the next step I will adjust the lens and install the software and the rest of the hardware (keyboard, monitor etc.). Maybe I’ll try some more ideas with the layers.


  1. I also realized that I don’t like cheap iced tea.↩︎

Download the microscope.md source file or follow this page with RSS. Some readers will want to know what Equipment I use. Who wants can Support my work or leave a Comment. Please use my PGP-Key for your submission.

x14km2d.com | 2021-2022 | Made for the Web1.1 with a Terminal and Pandoc.