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

Posted: 1/2/2024 2:19:00 PM

From: 30 km south of Paris (France)

Joined: 12/23/2022

I'd be so happy to make one for him as soon as I make one for myself !

Is there any reason for this super long pitch antenna ?

Posted: 1/2/2024 10:24:48 PM

From: Portland, Oregon

Joined: 2/22/2018

Is there any reason for this super long pitch antenna ?

He's a tall dude

Posted: 1/2/2024 10:34:29 PM

From: Northern NJ, USA

Joined: 2/17/2012

Maybe You Need More Than Q?

An experiment I finally got around to performing:


Posted: 1/3/2024 11:08:40 AM

From: Northern NJ, USA

Joined: 2/17/2012

Merry Hobbs Mid!

A demo of the latest hobbs_mid preset: https://d-lev.com/audio/2024-01-03_merry_hobbs_mid.mp3

Posted: 1/3/2024 3:30:51 PM

From: Northern NJ, USA

Joined: 2/17/2012


With the Estonian National Symphony Orchestra (from approx. 1:00 to 1:20):


She uses several different voices, and there are a few close-ups of the D-Lev controls & tuner.

Posted: 1/4/2024 5:46:23 PM

From: Northern NJ, USA

Joined: 2/17/2012

If you find yourself near beautiful Boonton, New Jersey on the Sunday afternoon of January 21st, please consider attending Rob Schwimmer's and my presentation of the D-Lev Theremin.  Open to the public, free admission & treats.  We'd love to see you there!

Posted: 1/20/2024 3:21:28 PM

From: Northern NJ, USA

Joined: 2/17/2012

Mmm Donuts!

Cross posting this from JPascal's Basic Experiments thread:

There's also this with a handy AWG list and more practical solving targets:


Links up at the top and off to the side to various calculators.

I can't believe there isn't software, or at least some form of ancient table or other manual guidance to make these things which were pretty common only a few decades ago.  I imagine a ferrite core would really complicate things, with various formulations, distances, etc. but an air core RF choke seems very analytically solvable.

For coupling one might be able to use two single layer windings in INCA.  Even small distances seem to hugely reduce coupling.

Coupling is nice because it boosts the inductance value, and can therefore be used to reduce the DCR and increase the Q.  But beyond that I'm not sure how important coupling between the donuts actually is.  We obviously want the "cold" drive end somewhat physically separated from the "hot" antenna end in order to minimize self C.  I've read that the cross section of the windings should ideally be square, which makes sense in terms of packing the turns together for coupling within the donut, but IIRC it's a broad maxima.

Taking the square donut winding cross section as something of a given, I imagine the optimal number of donuts would be rather high?  Not that it matters, scramble winding on a 3D form (in spindle slots) would be much easier than making a single layer solenoid, particularly when the AWG gets above 30 or so.

I guess I could select an AWG that makes sense for a given DCR / Q / mH, then experiment with these programs.  Physical tests could perhaps scavenge the wire back to do many such tests without creating a lot of waste.

There's the possibility too of counter-winding half of the turns, which would give negative coupling, but would be something of a "humbucker" for AC magnetic fields.

Posted: 2/7/2024 1:57:08 PM

From: Northern NJ, USA

Joined: 2/17/2012

2 Wire Mutual C?

I'm getting caught up in stupid stuff.  Wikipedia and others say the capacitance of two parallel wires is:

 C = pi * e0 * L / acosh(D/d)


L = wire length
d = wire diameter
D = distance between the wires

The thing is, you can put in absurd amounts of distance and the capacitance doesn't diminish by much due to the acosh.  For instance, two 1mm dia wires of length 1m and separated by 1m = 3.66pF.  Separated by 2m = 3.35pF.  Separated by 2000m = 1.83pF.  Separated by 2000000m = 1.26pF.  Separated by 8.8x10^26m (the diameter of the observable universe) = 0.4pF

I can't believe this is just mutual, mutual should be inverse to the distance, and particularly so at large distances.  But if it were mutual + self it seems like it should drop to a constant pretty quickly over a reasonable distance.

Posted: 2/7/2024 11:04:17 PM

From: Occitanie

Joined: 3/4/2012

I'm not sure about anything, but isn't there an assumption that L is very large compared to D ?

Posted: 2/8/2024 3:29:16 PM

From: Northern NJ, USA

Joined: 2/17/2012

"I'm not sure about anything, but isn't there an assumption that L is very large compared to D ?"  - Mr_Dham

Perhaps.  Though there is an equality sign in the equation rather than one of those wavy things that indicates an approximation.  Indeed, there is an approximation version to be found elsewhere which uses ln(D/d) rather than acosh(D/d).  But the plate capacitance equation has an equals sign too, and it doesn't take fringing into account.  To my mind, everything should revert to the plate equation given sufficient distance (in the far field everything is a point source).  But I'm reaching the point where I don't even trust the basic plate equation.

I never noticed this before, but in the INCA program ZIP file (https://www.coe.ufrj.br/~acmq/ see the "other programs" page) there are two academic-type papers written by the program's author that discuss the techniques involved in calculating self and mutual inductance, and self and mutual capacitance.  There is even a C solver for types of coaxial configurations that may be constructed of lines and arcs rotated about the common axis.  No source code, unfortunately.

Kurt Nalty's web page is worth a couple of weeks of entertainment, now only accessible via the wayback machine: http://web.archive.org/web/20190407233301/http://kurtnalty.com/.  His incredibly useful "Helmholz.pdf" paper (http://web.archive.org/web/20190407233301/http://www.kurtnalty.com/Helmholtz.pdf) solves self and mutual inductance of a torus and 2 coaxial torus system using 2 different methods.  Theremin type coils can be constructed from these, so they are the keys to the city.

From Kurt's paper, the INCA inductance paper, and the Coil64 program (https://coil32.net/) discussions and source code, my Go program now successfully calculates the self inductance of air core single and multi layer solenoids and chokes, which is fantastic. But the INCA capacitance paper is a bit over my head, and it doesn't explicitly show exactly how to perform the calculations.  David Knight's capacitance paper (https://www.g3ynh.info/zdocs/magnetics/appendix/self_res/self-res.pdf) is entirely devoted to single layer air core solenoids, which doesn't help with chokes and such.

With inductance, self and mutual are really clear things, but this isn't the case with capacitance, hence all of the squabbling in physics forums.  If one end of the coil is driven or otherwise essentially grounded, then the self C of that end is moot, though the self C of the other end isn't, so the total effective C of a coil depends on how it is connected in circuit.  In multi-layer coils there are clearly inter-layer C effects due to phase differences, because there is a 90 degree phase difference between the ends of the coils, and I assume this is more or less linear down the coil length.  Like the Miller Effect in transistors where the 180 degree voltage inversion and gain from base to collector magnifies the base to collector capacitance.  This article:


discusses how to perhaps qualitatively compare different winding arrangements, which might be enough to fuel some coil design experiments.  I wish I had equations for the self C of a torus and mutual C of two coaxial torus.  I can get numbers from INCA that I more or less trust, so maybe I can come up with some polynomials or similar that give rough approximations for reasonable dimensions.

You must be logged in to post a reply. Please log in or register for a new account.