Outputs

Used tools, methods and technologies:

• Arduino Nano microcontroler
• Microphone, LED Matrix and a OLED display
• Kicad, 3D printing, small CNC machining
• Less sleep than a humming bird

     This project was a effort to combine many aspects of making and signal theory together. A portable device that could perform many audio related tasks and visually display them. It could be a tuner, sound spectrum visulizer or a sound recorder.

     First, I got a better Op-Amp design then before. It was clear that if I want a good audio for recording or analysis, I need filtration and a low-noise amplifier. Basic concept of the architecture is from a video by GreatScott. Thanks for a great video! I drew the schematics in KiCAD and made some costume symbols and footprints for the matrix and OLED display. Circuit parts and explanations can be found at the bottom of this page or here. Now the schematics:

     When it was done, I exported the KiCAD project and opened it in FUSION 360 in CAM mode (manufacture). With the help of my teacher, I fixed dozen of problems and then then after 2 hours of fixing and CAM work it was ready to be made. We did it on a CNC machine Carvera Air which I used in week 4.¨ The material for creating the PCB was a 1.6 mm Cuprexit board with a thin layer of copper.

     Here is a small cut to the second and final prototype I made, so the board is different and also not made alone. With my teacher we also applied a protective mask to protect the copper from oxidizing and shorting by external objects. It also made soldering a little easier. The mask is a photoreactive substance and we used Relife RL-UVH900R.

     Then it was around 2 hours of soldering (the first time, the second I was done under an hour). The circuit worked right, which made me super excited. Then came the programming. I was not able to get the desired output, noise levels were high and the whole circuit was not nearly as sensitive as I expected. I spent too much time trying to make it work in software. The second day I went to an ossciloscope and with the help of doc. Petrucha we identified the problem. The microphone was ¨ overstaruted due to high supply voltage to it, around 4,8V . So I swapped the 1k resistor before it with a 5.6k, which produced around 3,7 V on the microphone and that worked well. We also changed the Op-Amp's amplification from 21x to 45x times by changing the old 47k input resistor to a 20K one. Those were the changes I made to the final design.

     Sound is recorded as a analog signal and then an FFT is used to get its spectrum. The Arduino Nano is not powerful enought to compute the whole FFT, it does not have enough memory for it too. But we can use a specific fix-point FFT to compute the spectrum at only specific frequencies. Precisely 64 of them. That gives us a idea about the amplitude spectrum and then we can use this information to approximate the tone that is being played. A simple scipt then displays the spectrum and tuning interface on the OLED display and LED matrix.

     Designing PCB for CNC manufacturing method requires some extra precursions and rules in KiCAD. I found out that you want atleast 0.6 mm paths and 0.4 mm spacing. You also have to check in Fusion 360 CAM all corners and holes for errors. Also be very careful with brushing the mask on pads away, it can easily brush away the copper layer too. And while soldering be gentle, strong impact can remove copper path and destroy your board. And finally, next time I will take more time than a week for such a big project. It was a rush and my sleep harmonogram was non-existing. Oh and to doc. Petrucha, I WILL learn how to operate ossciloscope, i promise!