Boldly Going Forward

A project log for Electronic 6-hole Ocarina

Play without annoying the neighbors

dehipudeʃhipu 10/23/2022 at 21:400 Comments

It has been some time since I last worked on this project, mostly because the tests now require actually playing some sounds, and that generally annoys people. But I have been home alone this past week, so I managed to push this a little bit forward. First of all, I had to figure out how the holes actually work. So I took the fingering chart, noted the frequencies of the notes, and compared how covering and uncovering of each hole affected the frequency.

Turns out the frequency is roughly linear to the total surface area of the holes (as helpfully stated by this Wikipedia article), but not quite, as it also depends on the total volume of air and other details. Also, most ocarinas are not really tuned according to equally tempered scale, which is why makes them so interesting — each one has its own unique feel.

So I decided to just go with very rough increments of frequency by 200, 150, 100 and 50 Hz respectively for each of the four holes on the front (ignoring the holes on the back for now), and later I might just "snap" them to the nearest clear note, if that's needed.

Immediately I discovered several problems with the current prototype. Since the traces for the touch pads pass under the speaker, and they are pretty sensitive, the speaker interferes with the touch sensing, giving a weird vibratto effect. Also, the traces for different pads are too close to each other, and it's very easy to touch ones you didn't mean to touch. And finally, the speaker was very quiet.

I solved the first and last problem by removing the speaker from the PCB, adding a MOSFET, and making the speaker hang from some wires. Unfortunately the second problem will require a PCB redesign, and I decided to also add a battery holder, so that it can be used without a computer.

I still have one more problem, which is a little bit weird, that I so far failed to diagnose properly. It seems that removing a finger from the pad is detected with some noticeable delay. I'm not entirely sure where that comes from, so I'm holding off on ordering a new PCB yet, in case the fix involves moving the traces again.