Building a Haxophone : Part 1, the whys and wherefores, buying and ordering parts

For a long time I've wanted a good wind-instrument-based MIDI controller. A few years ago I bought an Artinoise re.corder. I like it but it's not ideal for me. Before we get to the Haxophone, I'll expand a little on my issues with the re.corder. 

It doesn't have wired USB MIDI, only bluetooth. This does allow it to connect to the dedicated app and my Mac. But bluetooth is always a bit of a faff and there is bad latency in this case, which may be the fault of the bluetooth, or the instrument itself or both. If it worked over USB, it could reduce the latency to zero. I'm almost always using it in the studio and wouldn't mind the cable.

Maybe this next point is a configuration issue. I remember discussing the issue on forums so I'm sure it isn't just me. You need to blow quite hard to get a response. I helped this issue by partially blocking the air-escape hole (the fipple if using it as an acoustic recorder) but I still don't particularly enjoy playing it. It just doesn't feel as easy or natural to blow as a real recorder. 

Another issue is the capacitive touch. The re.corder is also usable as an acoustic instrument so it needs physical holes. Touch-sensitivity around those holes is the only way to go there. Other similar instruments also have touch-pads. However, this type of instrument gets wet. When that happens, it affects the use of the pads. In particular, the thumb one (which quite cleverly allows for full coverage or 'half-holing') becomes unusable and there's no option but to stand the instrument up and let it dry out. 

Enter the Haxophone

(picture from the open-source project, not my own)

This is an open-source project and it looks that way. I don't mind that. In fact I like the DIY look. I also love that it's "designed to be an inexpensive and fully customisable travel saxophone".  

It was the first time that I'd seen mechanical keyboard keys and caps used on an instrument like this. This is very appealing. I think the positive action of those will make this much more playable than a capacitive-touch instrument. Although this is modelled on a sax, I believe it'll be possible to build one with only the 12 keys needed for recorder-style fingering, and to add both soprano and alto fingering maps to the software.

Buying parts

The designer aimed to keep the cost of this below $100. I was never under any illusions that this would be easy or cheap. You have to buy parts from several different places, often with postage to pay on each order, and sometimes you have to buy multiples for economy or because you have no choice. Sometimes you make the wrong choices or just order things you don't need. So for the money I've spent so far, I think I could have bought a decent readymade. However, the DIY aspect and the customisation possibilities make that worthwhile for me.

Personally I don't need the audio side. I'd far rather just have MIDI out which is really flexible and useful. Therefore I ordered the PCBs without the surface-mount assembly which saved some money. There's a minimum quantity of 5 when ordering from most fabricators. This is fine because I have plans to build some variations. 

At the same time I ordered the 3D-printed parts. These aren't strictly necessary. You could blow into the flexible tube and manage without the thumb-rests, but I like the design of these parts and I'm not trying to save money. 

I don't have a printer but I do order 3D-printed parts regularly for other projects. What I wasn't sure about was the best material for these. PCBWay happens to default to 'default white resin'. Choosing that wasn't so much a deliberate decision as simply going with the default because I didn't know what would be best. As it turns out I think that it might have been a very happy accident. This is the first time I've handled resin parts and I really like the feel. They feel more like ceramic than plastic and I suspect it'll be better for hygiene reasons.

It also turns out that the small self-tappers I happened to have will go into that material beautifully and securely without the feeling that anything's going to crack. 

For the Pi, I've gone for this option - an RP2040 Pi Zero. "All the power of a Pi Pico in the size of a Pi Zero" as someone said. It sounds like the worst of both worlds when put like that but I think this is a great choice. 

Haxophone software is available for the RP2040 - Raspberry Pi's microcontroller. This has the advantage of being ready to go as soon as it's powered up (unlike the full Pi, required for the audio version, which takes a minute or two to boot up) and this is perfectly capable of receiving the user's input and generating MIDI. I use that controller all the time in my own instruments. 

To keep things simple on the Haxophone board, it just has the one standard Pi header and this board fits that. 

There's a lot more on this device - most of which won't be used here but could be useful in future projects. Importantly it has USBC which is way more robust than the micro-USB that the Pi Pico has. 

I bought other components from Amazon (keyswitches and caps) and Mouser (most of the rest of the electronic components and some hardware).

Naturally when I started to build, I found that I had missed some parts and have had to place a further Mouser order. Even though I'm not building the audio version and didn't have the surface-mount parts assembled, it turns out that some surface-mount parts are needed for basic operation, namely the keyswitch diodes, some pullup resistors and an i2c adc. I'll have to handsolder those parts but that's feasible. I've made a start and have ordered the parts that I missed. 

I have to say that the documentation for this project is excellent. The assembly guide in particular is well-written, comprehensive and has loads of pictures. 

to be continued...