[sdiy] Questions: Akai ewi 3020 controller to CV

[Richard Arntzen]

Hello.

I want to DIY my Akai EWI 3020 controller into a conventional CV controller,
and being a newbie (a measly ChemE) I have some initial scares that I have
to overcome. I could of course go MIDI, but then I have to lug around this
beast of a module. And it wouldn't be nearly as cool as having my own
controller module.

The controller is capacitive (which I still haven't figured out exactly what
means in this intstance). It has four  voltages going in: +/- 12VDC, +5VDC -
and ground. No service manuals available as far as my searches can tell.

1) Have anybody already done this :)
2) Will I fry the controller if I apply +/- 15VDC instead of the existing
+/-12VDC? Then I can use my MOTM-power-supply which would be really neat :)

I have poked around inside the companion module (an EWI 3030m), and there is
a circuit board that handles the signal from the controller and sends it to
the digital synth engine PCB. One of these days I will probably have to
dismantle the thing and check out the controller PCB - with the 7
panel-mounted pots... But I have gotten some readings out from near the
internal instrument connector:

The controller sends 6 signals back to the module:
Key; 0.20 V/semitone (0.15 - 3.55 VDC)
Octave; 0.59 V/octave (0-3.55 VDC)
Vib; 1 - 1.2 VDC - at least with the pressure I am able to apply. It is a
blip-signal, so my DVM may be too slow to catch the highest reading.
Breath; 0.46 - 3.5 VDC, again with the pressure I am able to set up.
Glide; 0.64 - 1.64 depending on the coverage of the capacitive pad.
Pitch; -3.5  - 3.5, -0.26 in zero mode

Now - it would be pretty simple to offset/scale these voltages using OpAmps
circuits (after re-measuring for +/-15VDC first). I have prepared some
schematics, and I am starting to get worried (...) about temperature drift
and accuracy since (I assume that) 1 V/oct means 83 mV/semitone, and so
believe I have to use high-specs resistors, trim-pots, and possibly
temperature-compensation circuits in order to get some stability into this.

3) Is it as hard as I think to get decent reliability by analog circuitry
alone?
4) Does it make more sense to do this digitally - AD-DA?

I hope I don't sound too paranoid.

Regards

Richard A.