EWI mouthpiece with breath and bite (lip pressure) sensors

Fair point Jonathan, I'm running 5v with a pull-up but at 3.3V it may not suffice.

As a note on the optical method, I just found the service manual for the AE-10 online, it appears the cantilever bar is equipped with a tab that simply just sits on top of a reflective sensor (type KTIR0711S) which is much like this:


So, nowhere near as complex as I thought, and one point scored for an optical assembly in case the Hall Sensor does not work out. This system may not even need an amplifier as output swing is effectively a range between 0V to VIN (5V)
Took a while to arrive due to the delayed shipping these days, but I have tested the QTR-1A sensor.

It works at either 3.3 or 5 volts and has no trouble providing full scale output with 2-3 mm movement depending on the height of the reed above the sensor and reflectivity of the reed surface. I am surprised at the sensitivity and no signal conditioning or amplifier would be needed.

This will lend itself well to being inlayed into the mouthpiece beneath the reed and detect lip pressure movement.

Sensor - 1.jpg
Last edited:
I've been using these sensors for a MIDI wind controller breath sensor. It has an appropriate pressure range and has been reliable.
Sourcing was problematic last year - competing with ventilator makers for parts! It's analog, and I can read it as fast as I like. I average the sensed responses for a few samples to smooth the signal.
(I use the RunningMedian library for the math).

I'm interested to learn along with you about potential bite and/or lip-squeeze pressure solutions.
Looks like a good option!

I am using an MPX5010 in my new EVI. The MPX2010 works just as well in this current one (and much cheaper) but it does need an op-amp for full output.

I have made an optical joystick with these reflectance sensors to use in my new EVI project.
I replaced the potentiometers in a playstation remote joystick with small vanes that move toward / away from the sensor and this has proven to be much better than using pots.


Likewise, for the lip pressure sensor, testing shows this to be a viable, simple option. It will just reflect off the inside of the reed and
return ample signal requiring minimal processing.

It would be great to know more about your design!

Last edited:
This makes me wonder if it would be possible to create an instrument that controls the pitch by mouth shape alone without using fingers. My finger coordination is not good at all, but I'm good at whistling precisely in tune.
But I imagine such an instrument would require a reed with lots of movement freedom and I'm not sure how blowing into the pressure would work with mouth wide open :D Another option would be to use optical sensors that reflect from mouth cavity (KontinuumLab had one similar experiment), but it could be tricky to ensure accuracy and hygiene.
I imagine that would be pretty complex and you may end up some a sort of a theremin-like device. Could be a bit frustrating to play.. I made a midi theremin, and let me tell you they are a real challenge to play!
Yes it would, although they have a limited lifetime as far as I know. Optical sensing is easier and avoids a multitude of issues with setting the range, drift, etc.
Sort of a necroposting here, but anyway - for the bite sensor you a can also look into how Nurad's sensor is implemented. It is quite creative, as far I understand you chew some silicon tube, which is a part of the closed pneumatic system, and the tube compression causes pressure which can be read by a yet another pressure sensor.
Hello, another necropost, but might as well give it a shot.
I'm a ewi5000 owner and do my own patches (internal sounds suck) and i'd have been pretty happy with this instrument if not for the problem that its side buttons which give sort of "sustain" pedal functionality, as well as interval (which can be fixed or dynamic, playing the first interval jump afterwards when you activate) work only with internal sounds and instrument doesn't transmit any extra midi data whatsoever. :/
akai seems not to be giving any shit about this market whtsoever, and all the products being released are mostly just for hobbyists, so i've been nurturing a plan to make a custom arduino controller and hook up existing hardware/sensors to the new unit, effectively modernizing this device into a new gen MPE capable controller, would anyone give me some direction where to start? is it even a feasible idea?
i've got some basic coding knowledge and pretty ok with these kind of things, but programming my own hardware controller would be the first, and i'm also a bit afraid of their proprietary stuff, that i simply won't be able to do anything with it
Hey, @arion_s, I'd love to swap stories. I've built a trumpet-like wind controller and I'd like to learn your experiences on the sax-like EWI controller. PM me here if you'd like to chat.