Fluorescence Digital Microscope

I wanted to create something I could be excited about and, by doing that, share knowledge with people. So, I developed this project to involve anyone who wants to know how a fluorescence digital microscope can be made using their own resources.

Let's begin

To understand this project easily, we need to break it down into 4 categories: Motion control, Lighting, Mechanical design, and Image acquisition.

Motion control

For motion control, I wanted it to be reliable and easy to work with. So, I decided to develop the motion control based on the development board Arduino Uno. I added to the Arduino a CNC shield and 3 TMC2209 drivers for the 3 axes (X, Y, and Z)

For the motors, I went for 2 NEMA 11 TR6-12 50 mm motors for the X and Y axes and a NEMA 11 TR6-4 50 mm motor for the Z axis. I wanted a bigger lead in the X and Y axes (12 mm) because I needed more mobility there. The Z axis needed to be more sensitive, so I stuck with 4 mm.

To control the motors I used 2 joysticks. One for X and Y and the other one just for Z.

Now, I just needed power. I decided to go with 4 18650 batteries to power the motors. After connecting everything and programming the Arduino, the motion system was ready.

One last thing: if you don't want to struggle with running out of power every 10 minutes, you will need to install a battery balancer. This will balance the load between all 4 batteries. If you don't install this, one battery will completely run out, leaving the rest almost fully charged.

Lightning

One key aspect of microscopes is lighting. You can make everything perfect, but if the lighting is not good, you will not see anything on your screen.

I used 3 main light sources with two lighting methods: Episcopic and Diascopic.

For the Episcopic light, I used 2 white LEDs and 1 UV LED. The positioning of the two white LEDs was simple; you can see that in the image below (or in the 3D files in the repository), but the position required to properly illuminate the sample with the UV LED was another story.

The challenge was to illuminate a sample with UV light. The issue was that the focal distance was about 1.6 mm, so I had very little room to illuminate the sample. My idea was to adapt an optical fiber to guide the UV light from the LED to the sample. I went through many versions of the fiber holder because the fiber had a diameter of 2 mm, so I needed to introduce the fiber at a specific angle that would allow me to illuminate the sample at the correct focal distance. In the end, I was happy with the results.

I also had to fix the UV LED in order to couple the fiber into the LED, so I designed a solution (I’m very proud of that design). Additionally, I covered the fiber body.

I also added a diascopic lighting method. I placed a white LED at the base of the sample holder. The lens of this LED is different from the ones on top. This unique form allows it to illuminate a wider area.

To control the light intensity I had to design a PCB with 3 PWMs in it, one for each lightning system.

After designing the PCB, I sent the schematic to a PCB manufacturer, and once I received it, I soldered the components. You can check the details of which components I used in the schematic files.

Mechanical Design

The mechanical design was probably the most time-consuming aspect of this project. Different versions of a single fixture were designed, and changes were made along the way, but in the end, I think it came out very well.To design all the pieces, I used NX from Siemens.

I also printed the parts in my own 3D printer. All material used was PLA.

For the mechanical design, there is not much to say. It just took me hours and hours of measuring parts, designing, and printing. Of course, you can access all the designs in the repository.

Image Adquisition

For the image acquisition, I used a Raspberry Pi 3 B+ with a camera module 1. I needed to add a lens so I could magnify the image. I took it from an old DVD reader (DVD readers have high-quality lenses).

After taking the DVD lens out of the cage, I attached it over the camera module using tape (I wish I had a better solution to fix the lens, but actually the tape works perfectly well).

After this, and using Python, I developed an application on Raspberry OS to be able to see the camera image in real time and save pictures of the sample. I'm not very good at coding, so AI's help was very useful.

Final outcome

According to my numbers, the total magnification is about 42x—not bad for a homemade microscope.For the sample, I used a dead scorpion's body because I knew they are biofluorescent. I was amazed when I saw the results.

Using episcopic white LED light, I noticed some transparent protuberances on the scorpion's body.

I was impressed when I turned on the UV light.

As you can see in the image, the white protuberance disappeared! I was wondering why this was happening, and it turns out that those protuberances are minerals that the scorpion uses to harden its exoskeleton. These minerals are transparent to UV light, so when you illuminate the scorpion's body, you will not see the minerals, but the indent that the mineral made in the exoskeleton. What an amazing thing—I didn’t know that.

Assembling all these aspects into the project was a good challenge for me. I really enjoyed working on this because it had the perfect amount of difficulty without being too complicated. I feel proud of all the work I did, and I’m excited to show others my work.

Thanks for reading me.