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

Posted: 2/28/2014 5:54:22 PM

From: Northern NJ, USA

Joined: 2/17/2012

"If there is a 40 mH coil in series with the antenna current and voltages are phased and the antenna works well as emitter and receiver of electromagnetic waves.

If there is not the 40 mH tuning inductor current and voltages are 90 degree phased and the antenna does not emit and receive any electromagnetic wave."  - livio

Your two simulations are not comparing the same things.  For the EW simulation you should just look at the voltage across & current though the antenna capacitor, and not the combined coil and antenna.  If you do you'll see the same 90 phase as in your design ("ELI the ICEman" - current lags voltage in an inductor, voltage lags current in a capacitor, and both by 90 degrees).

Posted: 2/28/2014 6:15:22 PM

From: Eastleigh, Hampshire, U.K. ................................... Fred Mundell. ................................... Electronics Engineer. (Primarily Analogue) .. CV Synths 1974-1980 .. Theremin developer 2007 to present .. soon to be Developing / Trading as WaveCrafter.com . ...................................

Joined: 12/7/2007

"But the electric fields behave differently, do not travel much distance, do not cross the walls. The electric fields can be stopped by a thin layer of material, only slightly conductive." - livio

I did say that im not going to argue - and im not going to!  ;-)

But you are wrong when you say "do not travel much distance" - Capacitive (electric) fields "travel" infinite distances - if there were two 'plates' charged with respect to each other anywhere in the universe, they would have an influence on each other, and this influence would pass "through" any objects - conductive objects acting to distort the (spread out) the lines of the field, and other objects behaving acording to their dielectric properties.

There is no end to the influence of electric fields - a theremin will be indluenced by a changing ground object a mile away - if one has a large metal object placed on insulators a mile away, and were to have this connected to a switch which connected it to ground, and were to pulse this switch, it WOULD be detectable by a sensitive enough capacitance sensor (which is what a theremin is) a mile away..

The ONLY thing that will block this is another GROUNDED object between these two.

Put a theremin in a room whose walls are lined with grounded aluminium foil, and another on the other side of the wall 50cm away, and the two theremins wont interact with each other... If it was EM, they WOULD interact with each other!

"The difference in behavior is similar to that between the alpha rays and gamma rays."

This is something I know a little about - I worked as a medical physicist for 5 years, and was graded as master of science (Msc) level by the IEEE following an examination in this discipline in order to legally be qualified to develop instruments for radiography (being upgraded from Bsc to Msc without having to do much extra study was a real treat! ;-)

There is no similarity whatsoever - electric, magnetic, EM are not even loosely comparable to alpha and gamma - alphas are particles, gammas are rays (EM), electrical fields are not particles - Alphas are large, which is one reason they dont go through paper - electric fields are charge distributions, they are size-less (or at a stretch one could say their 'size' is that of an electron or two ;-)  and they do 'go through' objects (these objects or materials become or are dielectrics in the field, and the effect of the field is to to 'bias' the orbitals of electrons of / within the dielectric) - over long distances this biasing is extremely slight, but can be seen easily when two capacitive plates are in close proximity and a high potential  difference is across them - the electrons are 'stretched' towards the +ve plate, and the nucleus is pulled towards the -Ve plate, and this effect will occur in all material between charged plates - thats what an electric field is - its a distortion of the electron 'clouds' and a distortion of their orbitals with respect to their nucleus.

With a HF electric field, the 'mode of transference' is the changing of electron / nucleus orientation .. but a lot of factors are not fully understood - in particular, there are differing ideas about how / where charge is stored - I tend to think of it as the dielectric being an elastic 'substance' -

but the bottom line is that these fields are not stopped, not by paper, not by brick, not by vacuum - wherever there are two charged plates, there will be a electric field between them - and all sorts of exciting things happen if you charge plates and then relocate them with respect to each other, because the charge will remain, even if the capacitance changes - which translates to the potentials on these plates changing as their distance changes...

I am actually involved in an electrostatic motor development at the moment, which uses shaped dielectrics, moving plates, and astronomical voltages, to provide rotation - all using electric fields! - this is being developed to fit a massive capacitive collector that will 'catch' lightening strikes, and use this extremely high voltage to power this motor, which will turn a generator to ptovide usable electricity.... LOL ;-) I think the only reason people come to me with jobs like this is because Im the only person crazy enough to take them on, and because I risk not getting paid if they dont get the funding... I am only involved in a tiny part of the project...

"Yes FredM, oscillators can be coupled also via an electric field, also in the complete absence of electromagnetic waves."

Youve got it the wrong way 'round! - Coupling is almost exclusively by electric (capacitive) fields - EM DOES provide some 'coupling' but this is so miniscule that it is COMPLETELY IRRELEVANT for all practical purposes!

I am not able to do the sums (am too stupid!) to give you any numbers regarding how insignificant EM is for theremins - but its many orders of magnitude less significant than capacitance, and certainly not worth even thinking about with the kind of theremin designs any of us (including you) are developing.

I have steered away from RF all my career, and the whole area of electromagnetic radiation is, to my mind, on the edge of my comprehension - but I am sure that you dont understand EM any better than I do, simply from reading what you are saying about the nature of EM "waves"..

So all I can advise is that you go back to stuff you can really comprehend, or have a chance of comprehending - get a full real understanding of capacitance and electrical fields - you will find that when you fully understand these, you wont need EM to explain how theremins work or interact.

Otherwise you risk becoming one of the theremin developers who loses their way in this EM nonsence, and worse, leaves a trail of misinformation which causes confusion.

If you dont believe me, do what I did - When I am not sure of something fundamental, I pay (or paid - dont have the money to pay anyone anymore ;-) someone who is specialist in the science to go over my designs and / or answer my questions - I also have a friend who works at the Diamond Lin-Ac in Oxford, and can contact some of the most renowned physicists in the world - to them, my questions are innane, but they give me answers I can trust!

Also - May I suggest that you be more scientific with your assertions.. Real science does not assert anything, it postulates ideas, tests these, and regards them all as "until disproven" - nothing is "sure" - you cannot absolutely prove anything, all you can do is say "it seems to work like this and we have done everything we can think of to disprove this idea, but not been able to.." .. Oh, we all tend to make "absolute" statements, and all get egg on our face sometimes - Everything I say could be wrong, everything Newton or Einstien said could be wrong - but I am probably more likely to be right if what I say agrees with what they said! ;-)

You are going out of your way to prove your ideas, and ignoring anything that contradicts your ideas - thats the way to bad science.. Oh, its natural - we all want to be right! ;-) -- And no matter how bad our science, there will always be people who will be fooled by our enthusiasm - but like inventors of perpetual motion machines, their happy delusion rarely gets them anywhere!

Fred.

Posted: 2/28/2014 6:41:08 PM

Joined: 2/2/2014

Our CapSensors can work very near one to each other instead EW can not (not tested personally but many people says this) for me this is due to the tuning inductor. If the tuned inductor is not the reason, please explain you to me another reason for this.

However, other reasons are recommending to not use tuning coils:

1) So large coils are not tunable and prevents easy interchanging antennas with very different capacitance.

2) A coil of such great value type does not exist "for oscillators" but only "RF-blocking" and the impedance will change very much with temperature

3) A large coil add much parasitic capacitance

4) A large coil suffer from dimensional changes that cause mechanical stresses and occasional noises

- - - - - - -

There are also other mechanism we are not considering:

1) A 40 cm antenna composed by a thin rod has an impedance about 100 times more than a plate similar to a 10 x 10 cm square. And therefore it emits and receives much better.

2) The tuning bobbin increases voltage and current on the capacitor and "adapts" the impedance to the oscillator.

- - - - - - -

Really you think that a 120 mW oscillator with a perfectly tuned loaded-antenna, emits the same as a 2 mW oscillator with a non tuned square plate?

Posted: 2/28/2014 8:16:33 PM

From: Northern NJ, USA

Joined: 2/17/2012

"Really you think that a 120 mW oscillator with a perfectly tuned loaded-antenna, emits the same as a 2 mW oscillator with a non tuned square plate?"  - livio

If all the power were being dumped into the antenna, OK.  The power you're measuring in simulation is being dissipated by the mechanics of the oscillator (static bias and dynamic current) not the antenna.  You're calculating and comparing different things, and none of which take into account RF losses / radiation.

Posted: 2/28/2014 9:45:57 PM

From: Eastleigh, Hampshire, U.K. ................................... Fred Mundell. ................................... Electronics Engineer. (Primarily Analogue) .. CV Synths 1974-1980 .. Theremin developer 2007 to present .. soon to be Developing / Trading as WaveCrafter.com . ...................................

Joined: 12/7/2007

"Really you think that a 120 mW oscillator with a perfectly tuned loaded-antenna, emits the same as a 2 mW oscillator with a non tuned square plate?  " livio

You cannot "tune" the "antennas", regardless of their form, plate or rod or whatever! - I realise that this statement is an error - not an error when it comes to theremins with 1/4 wavelengths of between say 25 meters and 400 meters, but wrong as a general statement.

The first fundamental misconception comes from the idea of "antennas" - These are NOT "Antennas" in the radio sense of the word  - they are a plate of a capacitive sensor. I believe (may be wrong, cannot be bothered to check the patent) that Lev Theremin specified what we call "antennas" as capacitive plates, and I am fairly sure his patent references electrical and/or capacitive fields, not "radio" or "EM" fields.

Yes - there are some radiated EM fields - if this was not the case then AM radios would not be able to "pick up" signals from theremins.. But these fields are TINY, as one can determine by the fact that the reciever must be really close to the theremin to pick up signals - Also, I suspect that most of the EM radiation is not being radiated from the antenna but from the inductors - but as I said, radio isnt my forte.

"Tuning" antennas is (as I understand it) a process of matching the antenna dimentions to the wavelength of the signal one wants to radiate or recieve.. Something which is NOT done on theremins (for obvious reasons).

With theremins the only tuning which occurs is to match the capacitance and inductance such that one gets the required change in frequency as the capacitance "seen" by the capacitance sensor (plate) (which is wrongly called an antenna) changes when the player moves their hand, and that this translates to a usable frequency shift on the oscillator - The only other tuning is completely unrelated to the antenna, and this is the setting of the local (reference) oscillator.

The above can be done with either parallel configurations, where the plate acts as a capacitor across the tank inductor (often in parallel with a small fixed capacitance), or with series LC (as Dewster is doing) or with the standard scheme invented by Lev Theremin, where there is a parallel tank having a small inductor and large capacitance, connected to a series large inductor connected to the antenna capacitance..

But IN ALL cases, the tuning is a combination of inductor/s capacitor/s and antenna dimentions - One can change any of these, reduce or increase antenna dimentions (resulting in a change of capacitance presented by the antenna, and a change in sensitivity to players movements) and make proportional changes to inductors and/or parallel / series capacitances and "tune" for the operation one wants...

Tuning is about tuning the RESONANT CIRCUIT of which the capacitive sensing  "plate" is a part, it is NOT about tuning this "plate" or "antenna", or tuning to "match" this plate in any way other than to produce the wanted oscillation frequency for the capacitance "seen" by this plate.

When it comes to linearization, there are good reasons why the rod "antenna" is superior to a flat plate "antenna" - The plate tends to increase capacitance more rapidly as a hand approaches it, giving far worse linearity at the treble end, the rod tends to spread the capacitive sensing to the arm, giving less dramatic increase in capacitance as the hand approaches the antenna..

If one is doing linearization through computation (either analogue or digital) then the above may be of little concern, but for classic analogue heterodyning instruments its of major importance.

The simple truth is that Lev got it right! Everything he did, rod antenna, loop antenna, heterodyning, hand away from loop = louder, were all the optimum configurations - His oscillator design was IMO the best for the application, and it is only now all these decades later that we are perhaps getting close to bettering his instrument.. (The only thing Lev may have got wrong was, IMO, that the theremin whilst a brilliant bit of science, is possibly one of the stupidest musical instrument ever concieved from a perspective of user interface and playability!)

But IMO we will only do this if we understand the science Lev was saturated in - Oh, we may now be able to linearize as well as he did without using double resonators - but if we get mislead by flawed terminology (like the word "antenna") and start re-engineering the theremin with flawed ideas about electromagnetic radiation, we will go nowhere.

And thats it - my last word on this subject.. Really! ;-)

Fred.

1) So large coils are not tunable and prevents easy interchanging antennas with very different capacitance.

Changing antennas isnt a common requirement - thereminists are quite happy with one antenna AFAIK ;-)

It is (IMO) true that the lack of ability to tune the series (equalizing) network limits precision - but one can tune the tank for optimum linearity and then adjust the local oscillator - On some of my prototypes I did make the series inductance electrically tunable (rather than the simpler common method of adjusting the local oscillator) - but realised that the added cost and complexity wasnt worth the effort.

2) A coil of such great value type does not exist "for oscillators" but only "RF-blocking" and the impedance will change very much with temperature

You dont understand the function of the antenna inductance at all! You dont understand that with care, all the problems you speak of above dont apply! Go look at Lev Theremins patent, look at the postings here about how the series LC works, because you really dont have a clue! - "RF-Blocking" LOL ;-) Look at the large (30mH for example) inductance, calculate the series resonant frequency with antenna capacitance, and get some understanding from the many postings here which explain whats going on before you start comparing the operation to "RF Blocking" !!!

Apart from which, the inductors DO EXIST! You can buy suitable Hammond or Bourns inductors from many suppliers, and you can wind an air coil.

These inductors are fitted in many theremins - If you are looking for a single 30mH inductor with suitable low capacitance / high SRF, you may not find one - but you can certainly get suitable 10mH inductors you can wire in series - these are not "blocking" inductors - they are pi wound and segmented specifically for HF operation.

3) A large coil add much parasitic capacitance

if you use a coil designed for a SMPS running at 40kHz, you may well get "much parasitic capacitance" - but when you use coils with high SRF and low capacitance, you dont get that problem - And if you use well constructed air coils you certainly dont get that problem!

4) A large coil suffer from dimensional changes that cause mechanical stresses and occasional noises

Really? First time ive ever heard of that! - I would imagine that after 80+ years all we would hear from RCA's with their massive ancient coils would be these "noises" if that was the case! - then theres the Moog 91 and the Ether-vox and tube theremins and Wavefront modified EW's all of which have large air coils and ive never heard of "mechanical stress" or "occasional noises" problems, so please let us know where you came across these problems!

"Our CapSensors can work very near one to each other instead EW can not (not tested personally but many people says this) for me this is due to the tuning inductor. If the tuned inductor is not the reason, please explain you to me another reason for this."

I dont know. It may be that the series inductors radiate magnetic fields, or it may be a multitude of things.. The test would be to have a grounded sheild between the theremins - if they still interact, then its magnetic or electromagnetic coupling.. then move one theremins orientation, if this changes the coupling, its likely to be magnetic (radiated from the inductors) but if its EM radiated from the antenna, then coupling wont be affected by orientation of the coils with respect to each other..

It may also have something to do with the way data is extracted from your theremin - It may be that the integration and low resolution of your theremin simply cannot resolve the interactions which are occurring in the way that an analogue heterodyning theremin which outputs all nuances of the interactions (as audio) can.

"But this requires a complete redesign, if you simply remove the four 10mH coils from a EW, its oscillator become so deaf, that it will be impossible to play it."

Been over this before! Yes - You would need a complete redesign! Thats because the EW oscillator was CORRECTLY designed for use with an equalizing inductor! .. It works like this - there is a small tank (oscillator) inductor in parallel with a LARGE tank capacitor - completely unsuitable and "deaf" if connected directly to the antenna, because the large capacitance utterly swamps the antenna capacitance.. This oscillator is not designed to operate with a varying capacitance - it is designed to operate with a varying INDUCTANCE across its small tank inductance!

Thats what the antenna inductance in combination with the antenna capacitance does! It is effectively series connected across the tank, and operates in its (nonlinear) inductive region - Effectively, changes in antenna capacitance are converted to a "virtual" change in the inductance across the small tank inductance, and it is this change in INDUCTANCE which adjusts the oscillator frequency.

A redesign would involve increacing the tank inductance to probably about 1mH, and decreacing the tank capacitance probably to between 100 and 200pF, one could then connect directly to the antenna without the series inductor, and the oscillator frequency would be directly controlled by the antenna capacitance, and because the tank capacitance was low, antenna sensitivity would be high ..

But doing that you would lose the increased amplitude on the antenna (go from perhaps 180V P-P down to perhaps 20V P-P) and lose the linearizing effect provided by the non-linear transfer function of the series LC combined with the parallel tank.

The EW is designed in much the same way as Lev Theremins design, take away the series inductor and one lowers sensitivity hugely - but one doesnt do that unless one doesnt understand how they work..

Just as adding a series inductor to a design like the Silicon Chip / Jaycar theremins (which have large tank inductor and small tank capacitor, and are "designed" for direct antenna connection)  or your theremino design, is completely pointless and probably detrimental to their operation.

This whole issue is one where hobbyists are exposed! And, alas, some of these "designs" find their way even ito somewhat reputable magazines - examples of such bodges are found all over the place - the "strange apperatus" theremin, and the EPE 2008 theremin (a Jaycar/SC theremin with a series inductor tacked on) are examples of such folly.

Its EASY to make this mistake - I never grasped the beautifull and brilliant idea of Lev's for a long time - and was fuc*ing about with tacking series inductors onto anything that oscillated, and wondering why I was getting strange results ;-).. But when I saw it, when I looked at the LC curves and it clicked that the series resonant section was behaving like a capacitance controlled variable inductor, and I saw that the non-linearity of this transfer function compensated for the inverse-square antenna capacitance - distance relationship to produce the required exponential relationship - well, I think that was the moment I recognised Lev's genius, and probably was the moment I saw how it all fitted together - I understood!

Others get understanding through different means - If I had comprehended the mathematics, I may have understood through that mechanism.. But the "virtual inductance" visualisation is the one that suits my "style" best - The frustration is that its real difficult to convey such a simple idea to others - lots of discussions have gone on the lines that series antenna inductance is "pointless" is a "waste of wire" or, as you have presented "a RF Blocking" component -

Think about it! Why would designers like Lev and Bob Moog add an expensive component just for the fun of it? Both were trying to minimise their costs, and if these inductors werent needed they would never have been fitted! - Less competent modern designers fit series antenna inductors to completely unsuitable oscillators, and this misleads people - but go back to competent designs, and the reason for this topology becomes clear!

Ok, the above was my last word! ;-) .. actually, the last quite a few posts shouldnt really be on this thread IMO, a new thread should be started and if possible some of these posts moved there - this is becoming a real hijack of Dewsters "digital theremin" topic!

Posted: 2/28/2014 11:04:44 PM

From: Northern NJ, USA

Joined: 2/17/2012

livio, I think (but I'm not sure, I'm not an RF guy) that you are right about one thing: I believe that coil loaded RF antennas work the way they EQ coils work on Theremins: short antennas look capacitive, so you match the coil inductance to the antenna capacitance to make the LC resonate at the frequency of interest, thus producing a large voltage swing on the antenna.  The larger voltage swing somewhat makes up for the poor RF coupling of the antenna to the "ether".  But the antenna is still poorly coupled in an RF sense because it is very short compared to the operating wavelength.

So the comparison method I was talking about before is probably valid: forget about series inductance, look only at the antenna voltage swing after the inductance and the antenna geometry at the operating frequency.  For your design this is ~30V @ 3MHz.  For the EW this is ~100V @ 300kHz.  But the geometries are different.  I'd say the RF emitted is probably close to a wash, if there was a 1,000:1 or 1,000,000:1 effect going on in reality I'd be completely amazed.

And just to be clear, my only concern with RF is for compliance, non-interference, and that sort of thing.  Like FredM, I don't believe the RF emitted / received has much if anything to do with the way Theremins interact with humans.

Posted: 3/1/2014 8:15:33 AM

Joined: 2/2/2014

------------------------------------
In response to FredM
------------------------------------

Capacitive (electric) fields "travel" infinite distances - if there were two 'plates' charged with respect to each other anywhere in the universe, they would have an influence on each other, and this influence would pass "through" any objects "

Yes, any field travels infinite distances, but the electric field:

- can be emitted with zero power

- can be blocked by a 0.05mm thin film like a kitchen aluminium foil (grounded)

- is completely blocked by walls (I have personally tested this)

- does not pass trough wood because it is slightly conductive

- does not pass trough the humid air, for distances more than some meters

- has a real power and the oscillator see it like a dissipating resistor

- easely passes through kitchen aluminium foils (grounded or not)

- easely passes through wood, walls and humid air

"There is no similarity whatsoever - electric, magnetic, EM are not even loosely comparable to alpha and gamma - alphas are particles, gammas are rays (EM), electrical fields are not particles"

I said only "similar"...  "The difference in behavior is similar to that between the alpha rays and gamma rays."
I
t was only an example!

"Also - May I suggest that you be more scientific with your assertions.. Real science does not assert anything, it postulates ideas, tests these, and regards them all as "until disproven" - nothing is "sure" - you cannot absolutely prove anything, all you can do is say..."
"You are going out of your way to prove your ideas, and ignoring anything that contradicts your ideas - thats the way to bad science.. Oh, its natural - we all want to be right! ;-) -"

You're right, I'm sorry. I am very sorry that my sentences sound bad and contain assertions. For me it is already difficult to explain my concepts in English and I can lose control of the "tone" easily. I hope you will try to understand me.

"You cannot "tune" the "antennas", regardless of their form, plate or rod or whatever!
The first fundamental misconception comes from the idea of "antennas" - These are NOT "Antennas" in the radio sense of the word  - they are a plate of a capacitive sensor..."

The EW antenna is composed by a 40 cm rod and a 40 mH inductor. This complete system forms a "charged antenna" end emits an electromagnetic field with strong real power (my best guess is: about 25 mW)

I will explain this better in the next post responding to Dewster.

"When it comes to linearization, there are good reasons why the rod "antenna" is superior to a flat plate "antenna" - The plate tends to increase capacitance more rapidly as a hand approaches it, giving far worse linearity at the treble end, the rod tends to spread the capacitive sensing to the arm, giving less dramatic increase in capacitance as the hand approaches the antenna.."

Based on my calculations, exactly the same linearization can be more easely accomplished with a single capacitor, of some pF, in series. (with our CapSensors this is possible, with EW maybe not)

"...because you really dont have a clue! - "RF-Blocking" LOL ;-)  ...  You can buy suitable Hammond or Bourns inductors from many suppliers, and you can wind an air coil.  ...  they are pi wound and segmented specifically for HF operation..."

More than 10 000 turns of thin wire, hundreds of meters of copper that extend with temperature... no comment !

"...the EW oscillator was CORRECTLY designed for use with an equalizing inductor! ..
A redesign would involve increacing the tank inductance... But doing that you would lose the increased amplitude...
...your theremino design, is completely pointless and probably detrimental to their operation....  This whole issue is one where hobbyists are exposed! And, alas, some of these "designs" find their way even ito somewhat reputable magazines..."

We have found an alternative way, simple, inexpensive, with zero tuning components. We can demonstrate this with working devices, already in use by many happy performers.  ;)

Posted: 3/1/2014 10:36:57 AM

Joined: 2/2/2014

--------------------------------------
In response to dewster
--------------------------------------

"...I think (but I'm not sure, I'm not an RF guy) that you are right about one thing: I believe that coil loaded RF antennas work the way they EQ coils work on Theremins: short antennas look capacitive, so you match the coil inductance to the antenna capacitance to make the LC resonate at the frequency of interest, thus producing a large voltage swing on the antenna.  The larger voltage swing somewhat makes up for the poor RF coupling of the antenna to the "ether".  But the antenna is still poorly coupled in an RF sense because it is very short compared to the operating wavelength."

"So the comparison method I was talking about before is probably valid: forget about series inductance, look only at the antenna voltage swing after the inductance and the antenna geometry at the operating frequency.  For your design this is ~30V @ 3MHz.  For the EW this is ~100V @ 300kHz.  But the geometries are different.  I'd say the RF emitted is probably close to a wash, if there was a 1,000:1 or 1,000,000:1 effect going on in reality I'd be completely amazed."

"And just to be clear, my only concern with RF is for compliance, non-interference, and that sort of thing.  Like FredM, I don't believe the RF emitted / received has much if anything to do with the way Theremins interact with humans."

I was building (and using) loaded antennas since the 70s, when Volkswagen vans and Citizen Band radios were in use, so I know by my own experience, that there is a big difference, between them and a metal plate.

What you (and FredM) are forgetting to consider, is the strong magnetic field produced by the 40 mH coil.

Doesn't matter if the coil is 10 cm away from the "antenna". The distance between these two components is practically zero (compared to 1000 meters wavelength). Coil and rod are tightly coupled and form a single component called "loaded antenna".

EW antenna and coil together, form a "loaded antenna" that (if tuned) emit a strong electromagnetic field, with actual real power. (my best guess is around 25 mW)

In contrast, an untuned plate does not absorb energy from the oscillator, does not emit energy (best guess: 50 uW) and also as a receiving antenna, behaves bad.

Posted: 3/1/2014 4:19:23 PM

From: Northern NJ, USA

Joined: 2/17/2012

"What you (and FredM) are forgetting to consider, is the strong magnetic field produced by the 40 mH coil."  - livio

What you are forgetting to consider is the inductor in your oscillator, which is also emitting an very real electromagnetic field.

This discussion has been very interesting for me because I now believe I understand how coil loaded antennas work:

Think of an antenna, which has a certain geometry, but that geometry is poor (small) compared to the wavelength we want it to function at.  So the antenna looks like a capacitor in series with a large resistance to ground (RF emission).  Books tell you to put an inductor in series with this to "cancel" the antenna capacitance, and to make it look like a pure resistance.  But what you are really doing is making a series resonant LC tank that boosts the voltage at the antenna, which makes it dissipate more power through the large resistance.  Small voltage swing into the coil gives big voltage swing at the antenna (at LC resonance, and if the Q is high).

In your design there is a parallel tank, but this only differs from the series tank above by the way you drive it.  The drive is a current source in parallel with the tank, rather than a voltage source in series with it.  There is no voltage increase, but this is made up by allowing the parallel tank to swing outside of the transistor power supply rails.

It's all the same.  Series and parallel tanks differ only in the way they are driven, not in the way they behave.

There is no magic going on in the combination of an antenna and a coil, just a voltage boost at resonance that helps overcome an impedance mismatch (antenna to the "ether") to some degree.

==========

I've seen on my bench that air core coils have a noticeable electrostatic field around them, particularly the end connected to the antenna.  I believe this is a big reason you see big air cores in Theremin's own designs, and why they are positioned near the antennas.  I'm not sure how much air-gapped ferrite cores emit in the way of RF, but the magnetic field is more confined to the core so I would guess less?

==========

What would be nice to know is the actual RF emission of a Theremin given a certain operating frequency, antenna geometry, and voltage swing.  I'll probably look into this at some point.  I'm currently trying to find an accurate formula for capacitance between widely spaced circular plates.

Posted: 3/1/2014 4:54:48 PM

From: Northern NJ, USA

Joined: 2/17/2012

Interference

I have a conjecture:

Common wisdom seems to be to avoid the AM band for pitch and volume operating points because strong broadcasts can/will negatively interfere with the proper operation of the Theremin.

My feeling is the way in which the volume side is often implemented (e.g. EWS) is mostly to blame in this regard.  With no heterodyning, and that diode located right at the tank, it looks kind of like an AM radio!

Whereas the pitch side, with heterodyning, acts more like an FM radio because what we hear is the difference of two frequencies, not the amplitude.

Anyway, the AM band is kind of a sweet spot for digital oscillators, and it allows for reasonable sized air core inductors, so I'm tempted to use it.  But I'm afraid of possible interference (and I guess the FCC wouldn't like  emissions there either).