1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Theremin Hand Capacitance Approximation Needed!

  1. Feb 23, 2013 #1
    I've been working on a Theremin Excel simulation for the past couple of months. For those who don't know what a Theremin is, it was one of the very first electronic instruments to be invented, and has two "antennas" that independently change the pitch and amplitude of a tone via hand capacitance. The capacitive changes "seen" by the "antennas" during play are in the femto Farad range so it's quite remarkable that it works at all!

    The pitch antenna is vertical, ~500mm in height, ~10mm in diameter, and the player's hand moves horizontally nearer to and farther from the antenna on a line roughly perpendicular to the midpoint.

    I have an equation that gives me the free space capacitance of the pitch antenna and it seems to jibe with lab measurements I've made. But I don't have a good equation that will estimate the capacitance of a human hand near the antenna.

    For modeling purposes I'm thinking the antenna rod could be represented by a 2D rectangular metal plate, and the player's hand represented by a 2D square metal plate facing the antenna, moving along the playing centerline. Inputs to the equation would be antenna height, antenna width, hand height & width (the same since it's a square), and distance between the plates. Output would be capacitance.

    I've spend a couple of days pawing the web and trying to make sense of my old college texts, but can't find this problem or one like it solved. Could anyone please help me?
  2. jcsd
  3. Feb 23, 2013 #2
    It might help you to know that
    At theremin frequencies the dielectric constant of the human hand is about 80
    The ISOLATED human hand might be a bit like a fist-shaped sphere
    It is connected to a large body (conductor) capacitively-coupled to ground.
    So it depends on which floor of the building the player is on and how near wiring and reinforcement metal construction mesh.

    You could MEASURE this using a digital voltmeter as sensitive microammeter.
  4. Feb 23, 2013 #3
    Thanks Drmarshall,

    I was thinking sphere for the hand but maybe that makes the math more difficult? I assume the solution will be a double integral, but don't know the exact approach. My physics book says to assume the antenna and hand are oppositely charged with +/- Q, then find V between them, then use C = Q / V. Even though it's a second year book it's kind of sketchy as to how to do that second step.

    It's true that grounding can very much influence the response of the Theremin, but I'm just looking for a parameterized formula that describes the capacitance formed by the hand | antenna better than the equation I'm currently using.

    My source for these formulas is "Physics of the Theremin" by Kenneth D. Skeldon, Lindsay M. Reid, Viviene McInally, Brendan Dougan, and Craig Fulton, 1998:

    Antenna capacitance above a ground plane (this gives pretty much what I can measure):

    Cant = 2*pi*epsilon0*L / (ln(2*L/D)-k)

    L = length of antenna in meters
    D = diameter of antenna in meters
    epsilon0 = permittivity of free space = 8.85E-12 Fm^-1
    k is a constant depending on how far above the ground it is mounted
    k = 0.4 for almost at ground level.

    Capacitance due to a hand in proximity (this underestimates capacitance near the antenna):

    Chand = pi*epsilon0*L / (10*ln(4*x/D))

    L = length of antenna in meters
    D = diameter of antenna in meters
    x = distance from hand to antenna in meters
    epsilon0 = permittivity of free space = 8.85E-12 Fm^-1

    I have measured antenna and hand capacitance for three different antennas, but my desire is to have a good enough model of things so that I can do more of the optimization up front and easily on my PC, rather than painfully, slowly, and expensively in the lab. I'd also like the reassurance that my measurements aren't out in the weeds as these are exceedingly tiny capacitances. I can do the heavy lifting but am in need of some hand holding and guidance.

    Can anyone point me to articles or texts that clearly describe the process and maybe give a few examples? What is the best physics text out there for this (that maybe doesn't use vector math for everything, I'm an MSEE and never really got the hang of curl and such, and I kind of don't like the text I was taught with)?
    Last edited: Feb 23, 2013
  5. Feb 24, 2013 #4

    See my entries above.

    Excellent books are
    Langford Smith : Radio Designer's Handbook
    Reference Data for Radio Engineers: publisher Howard Sams
  6. Feb 24, 2013 #5
    Hmm, I'm not exactly seeing how the self-capacitance plays into capacitance between two objects?

    The equation I'm using for hand capacitance seems to be way off (underestimated) in the near and mid field of the hand to the antenna (compared to capacitance data I've taken). Far field seems maybe OK.

    My spreadsheet sims are quite flexible in terms of oscillator circuit topology, so all I need is a decent hand/antenna inter capacitance formula to examine perceived linearity between hand position and pitch.

    It's a student paper, sorry I should have been clearer.

    The antenna often has a few hundred pF in series for various reasons as well, and the spreadsheet sim takes this into account.

    I guess the hand model not taking the attached arm and body of the player into account is a rather large simplification, but probably anything will be better than what I'm using now.

    Ha ha! Thanks!

    Thanks for this as well, will check them out.
  7. Feb 24, 2013 #6

    The capacitance of ANY capacitor depends on the area of each plate and the distance apart.

    The potential v depends on the charge Q per unit area.
    Thus the self capacity of a sphere gives you its area and the potential V falls off as 1/r
    All you need to do is add the Q/V for the other plate of the capacitor - the "antenna" - for which the formula you give may be accurate enough

    Yes the midfield region always gives trouble, for here the field strength does not build RADIALLY (as for a sphere, 1/r) nor cylindrically (as from a straigt wire - logs).
    BUt for any situation you consider LINEAR, the result of A + B is simply the sum of A and b if tasken alone. This is very powerful - but of course few things are linear despite what schoolteachers tell us!

    It is easy to write a computer prog that DOES cater for the midfield also - use R V Southwell's "Relaxation of Constraints"

    If your simulation is good enough use THAT to tell you how much hand capacitance changes the oscillator frequency 1%

    The "few hundred pF in series TUNES the (impedance of) the antenna.
    A straight wire length L, has an impedance of say 300 ohms times tangent (2Pi times its length in wavelengths) . So a wire open at its far end (L) looks like a short circuit at its other end where we desperately TRY to feed power in.
    This is VERY difficult! Resonance helps no end
    Short antennae are capacitive and those longer that 1/4 wavelength are inductive (until L=1/2 wavelength - and so on)

    So the series few hundred pF CHANGES the impedance SEEN by the energy source.
    We'd best MATCH it to the "internal impedance" of that source!
    The few pF of "hand-capacitance detunes this resonance - more than just "leaking away current".

    My capacitance (total to ground) is so BIG (100 pf) and in SERIES with your hand capacitance from antenna to my hand, that ONLY that hand capacitance really counts!
    (For capacitors in series 1/C=1/c1+1/c2 where c2 much greater than c1)
  8. Feb 24, 2013 #7
    OK, self-capacitance:

    For a sphere, my physics book says:
    V = Q / ( 4 * pi * epsilon0 * radius )


    C = Q / V = 4 * pi * epsilon0 * radius

    For the antenna, that paper says:
    C = 2 * pi * epsilon0 * Length / ( ln( 2 * Length / Diameter ) - 0.4 )

    How do I combine these linearly so that I get mutual capacitance, and the distance between them is a factor?

    My sim is definitely good enough for that and that's what I'm using it for.

    This is an exceedingly short antenna, nowhere near 1/2 wave, which is why I put "antenna" in quotes. It's really just a way to sense capacitance via change in the resonant frequency of an LC tank.

    If anyone is interested, my spreadsheet is here (along most of the project files including verilog - I am designing a mostly digital Theremin that doesn't use heterodyning per se): http://www.mediafire.com/?w36b3brqyg2g3

    Snag the latest one that has the file name "Theremin_simulation_*.xls. The phase criterion for the "Linearity" worksheet is a bit off, I've got a better one in the works but am holding off until this hand & antenna capacitance model is improved.
  9. Feb 24, 2013 #8
    The field due to two things ( Flux density, potential and charge) is merely the sum of the field of A PLUS that of B
  10. Feb 24, 2013 #9
    I'm confused. If I simply add the two self-capacitances there is no distance-between-them factor (mutual capacitance should increase with decreasing distance between them). Could you be more explicit or perhaps show me the formula you're thinking of?
  11. Feb 25, 2013 #10

    Andy Resnick

    User Avatar
    Science Advisor
    Education Advisor

    Here's an alternate method, making use of a circuit diagram which you may be able to easily alter for your specific circuit.

    The output tone is generated by subtracting the resonant frequency of an LC circuit from a ‘reference’ LC circuit operating at 173.4 kHz. The tunable LC circuit has L = 598 uH, C1 = 1.41 nF and C2 is variable, based on the hand position. C1 and C2 are in parallel. When your hand is far from the antenna, fvar = fref. When your hand is very close to the antenna, fvar = 171.6 kHz. This tells you the capacitance contribution from your hand/antenna.
  12. Feb 25, 2013 #11
    Thanks Andy!

    I've already measured hand | antenna mutual capacitance using this method for three different length antennas, so I have some data. Now I'm looking for a mathematical function for hand | antenna mutual capacitance that takes distance between them, the physical dimensions of the antenna (length & diameter) and perhaps hand size as inputs, and gives me capacitance. It can be a rough approximation. I just need it to be good enough to have some predictive power.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook