Let's Design and Build a (mostly) Digital Theremin!

Is that the one you owned for a while?  If you have a higher resolution image could you send it to me?  Thanks.

Metal Plate In My HeadThe Theremini is actually a hybrid rod / plate design 

Thanks for the insights! Function should be completely like a plate, only the outer design part suggests a rod. 

"Is that the one you owned for a while?  If you have a higher resolution image could you send it to me?  Thanks."  - pitts8rh

Yes I owned it briefly, and that pic is from my unboxing / teardown post back in 2014:

http://www.thereminworld.com/forums/T/29392?post=205118#205118

There are links in that post to the TW photo album, as well as to a ZIP archive of full resolution pix.  Those may have been somewhat curated, if you want all of my Theremini guts pix let me know.

(Read down the page past my post for ILYA's oscillator analysis, and Wilco Boterman's fairly scathing review with which I concur 100%.)

Metal Plate In My HeadThe Theremini is actually a hybrid rod / plate design Thanks for the insights! Function should be completely like a plate, only the outer design part suggests a rod.

I have Theremini, partially disassembled.
I can take some photos. What parts are you interested in?

Not Exactly 100% Touchless

Roger cleverly made some encoder knobs out of wire nuts and sent me some pix.  The idea here is to make the knobs as small as possible so the encoders can be tightly packed together without interference when twiddling them.  This inspired me to search through my extensive collection of rare & exotic priceless wire nuts from around the world to give them a spin, so to speak:

Some of the nuts feel too big, others feel way too small, some feel just about right.  They all have 10 ribs, except for one which has only 5 and feels weird because of that, and the angle from vertical is around 4 degrees.  This coincidentally matches my current knob (10 ribs, 4 degrees) which feels quite good to me (space gray @ lower left).  But the "just right" wire nut (yellow @ center) actually feels a little better.  So I decreased the knob diameter, made the ribs a little deeper, cut 1mm off the overall height, and made the skirt about as minimal as possible, and the result (white @ lower right) feels a little better than the wire nut it is based on.  I think the ribs being rounded by the 3D printer nozzle diameter (0.4mm) helps to make it less painful to grip, and the skirt is actually rather crucial to the feel as well, as it helps position one's grip at a constant height and diameter.  Of course, the main esthetic job of the skirt is to hide the mounting nut and washer. 

Above is a rendering of the new knob and a cutaway view, showing how minimal the knob plastic is compared to the flatted encoder shaft.  The perspective is different for both, which is perhaps a bit confusing.

Finally, above shows 1/2 old knobs and 1/2 new knobs on the control panel.  A definite improvement!

Little Ben(s)

Got a wild hare to try and synthesize a couple of really big bells: [MP3].  Sonic blueprints were bell videos on YouTube.  The usual 8 formants + resonator:

Musical Saw

At least two excellent Thereminists that I'm aware of also play musical saw most excellently: hYrtis and Grégoire Blanc.  Which makes sense really - the playing style, timber, and overall delivery are highly similar.

The simplest saw sound is a sinewave, with some harmonics added during the "bowing", which isn't too difficult to dial up on the D-Lev: [MP3].

The saw preset is based on my human whistle preset, with some vmod harmonics, a really long decay, and a moderate damping point farther out in the volume field (I believe that saw players who don't use a handle can damp the saw by relaxing the 'S' arch of the saw, and I'm sure there are other ways to damp a saw).

O Little Saw

One's playing style has to really adapt to much longer decay times: [MP3].  Playing the musical saw patch I'm using a lot of wide vibrato and infrequent stimulus, rather like playing the dobro patch.

=========

Both axes of prototype 3 (P3) have been really fighting me with ACAL (auto-calibration) - initiated by pressing the lower right encoder pushbutton - and it felt like time to finally do something about it.  The pushbutton is too close to the plates, and the volume axis is extra touchy because the plate is currently positioned so close to my body.  P1 was a joy to ACAL, just no issues at all, but its plates were way out in space.

Yesterday I looked at the code, cleaned it up here and there a skosh, and generally got ready to try some things out.  Had the notion that it could be based on "stability" - press the button and it keeps doing periodic ACALs at 12Hz (or some binary subrate thereof - this is the encoder update rate) until the cal values aren't significantly changing.  But the process might possibly never end due to interference or something, and so would likely need a settable variable threshold, so I've talked myself out of it.

The alternative was a strictly time delay type thing, with just enough time to move one's arms into the cadaver position or similar, so ACAL can get a good reading of environment + body, and indeed Roger has been requesting something like this for like forever.  So I implemented a one second delay for auto calibration.  The foot pedal ACAL input has the minimum (for the new timer) 1/12 second delay, and the encoder pushbutton and power-up each have a 1 second delay.  I think 1 second is around the sweet spot, too much shorter and I'm racing to bring my arms to my body, too much longer and it gets boring / harder to know when it happened.  If there is a significant change to the ACAL I see the LED tuner jump.  The timer is a byte, and so can go up to 255 / 12 = 21.25 seconds if need be, and all three ACAL delays (footswitch, encoder PB, and power-up) are independently adjustable in SW, though they are hard-coded.

With this I'm down to quite minimal values of ~12 for both Vcal & Pcal, and the process is much more repeatable and less squirrely.  Thanks to Roger for suggesting this way back when, and I'm super sorry that I dragged my feet on trying it until now.  (I fought the Roger and the Roger won - yet again! - when will I learn?)  In my defense, my beautiful mind was poisoned to delays here by the wretched Theremini calibration process.

=========

Another thing I tried was +/- level controls for the formants.  I thought that having control over formant phase might be handy, and perhaps give an optional deep notch between opposite polarity formants, but it was a bust.  Much as how bass and treble control boost and cut require two entirely different filtering settings, flipping the phase actually muddies up the transition.  Live and learn.

"The Aliens" Ringtone

A couple of episodes into the SF miniseries "The Aliens" and the main character's cellphone ringtone sounds like an old phone, but somewhat variable, muted, and slow.  So of course I had to try my hand at mimicking it (first is from the show, second is D-Lev): [MP3]

From time to time I try to play the musical saw theme from One Flew Over the Cuckoo's Nest using one of your earlier D-Lev presets (maybe the whistle?).  It takes either fluttery fingers or a very tense arm/elbow, and even then neither method achieves the high vibrato rate.  I've always like that piece though.

I suggest that you try the saw preset with some of these:

The First Time Ever I Saw You Face
I Saw Her Standing There
I Saw the Light
I Saw Three Ships A Sailing
I Saw Mommy Kissing Santa Claus
I Saw an X-ray of a Girl Passing Gas
Have You Ever Saw The Rain (little known CCR version before grammar/tense correction)