"Tanks" For Nothing!

"When you go to the extreme, grasping the pitch antenna and thus adding a hand capacitance of >=100pF, an un-EQed oscillator will go down by 30kHz or even more while the deviation of a well EQed oscillator will be limited to 5 or 6kHz. That's a huge difference!"  - Thierry

I imagine that some players now and then intentionally touch the pitch antenna for effect.  Not having the pitch go super crazy when this happens accidentally would likely be seen as a plus by most players.  So, yes, I agree this feature could make a huge difference in that respect.

"And yes, the range within that phenomenon does not only happen, but give musically useful results is very small..."

I hope you don't think I'm shoving your words back at you, and not to belabor the point, but different designers / engineers / players  / people in general can have legitimate differences of opinion on these types of features.  I personally don't think this feature is worth the added fiddling headaches that accompany it.  If it could give ruler flat linearity with not much fuss I'd likely have a different opinion of it.

"The idea of almost zero factory tuning kills totally the artisanal aspect of traditional music instrument building."

If one is building a one-off, it can (and most likely will) take on the form of a prototype, with various adjustments and such as temporary expedients, stop-gaps, to vary parameters for "what if" type experimentation in the lab, etc.  If one is designing for production one must necessarily work like a demon in order to nail everything down and remove all of the production variables one can.  There can still be real artisanship in a design in terms of the oscillator topological details and in the possibly hand wound coils, and in the major logical structures and how they are stylistically implemented (in my case the processor and code).

[EDIT] The existence of factory adjustments and final product variation can be a source of anxiety for the buyer as well.  For instance, I'd normally never buy a guitar sight unseen, but I might if I knew it had run the plek gauntlet:

Chris wrote: "I have been explaining this the best I could for ten years. You would think my consistency would make up for the fact I am not an engineer, musician, nothing to prove and nothing to sell. Maybe I should have called it Chris' Antenna... I don't because I did not invent anything, this is all natural."

Chris, I think perhaps what escapes the engineers in this forum is that there is a heavy, but largely unstated, philosophical/metaphysical component to your posts. You may be unaware of this because to you it all seems entirely natural....even obvious.

I presume you are aware that you live in a reality that is slightly out-of-phase with that of most of the people around you - and I don't mean just in this particular forum - I mean everywhere. 

To be fair, in the category of possible usefulness, a parallel LC tank is a bandpass filter.  This response, when coupled with the peaky (high Q) low pass response of the EQ / antenna series tank, could perhaps help lower environmental noise and so increase SNR in a Theremin application.  But I have no idea how this applies with the two resonances so distinct and close to each other, pulling on each other as it were.  Even defining what resonance actually means in this scenario (in terms of phase detection) is problematic.

"By replacing the link with a low Z mA meter, I can monitor current drawn by the oscillator - As the majority of this current is determined by the antenna current, and as this current is primarily determined by proximity to resonanace, this has been a useful tuning and diagnostic aid."  - FredM

Very interesting Fred.  Do you know how this qualitatively compares to placing scope probe tied to a short test lead ~6" from the base of the antenna and tuning for maximum voltage swing?

"Do you know how does qualitatively compares to placing scope probe tied to a short test lead ~6" from the base of the antenna and tuning for maximum voltage swing?" - Dewster

Hi Dewster,

LOL, No - I dont know how it compares! - the probe would probably compare extremely poorly against monitoring the osc current, I would expect! - ;-)

I have never done your "placing scope probe tied to a short test lead ~6" from the base of the antenna and tuning for maximum voltage swing?" - With about 10pF across a probe, this doesnt work unless you compensate for the probe capacitance.. and as one is looking at a few hundred fF change where one gets closest to resonance..

I have used a scope probe with additional capacitor (3.3pF) strapped across it (no antenna) just to get ball-park check on voltage levels at / close to resonance.. But its been of extremely little use to me.

To be honest, I have never managed to get satisfactory results from any antenna-in-line monitoring scheme unless this was actually built into the theremin as a permanent circuit which remains in circuit while the theremin is played.

I did lots of experiments using saturable reactors in the equalizing coil (for tuning and linearization) and was monitoring both the oscillator frequency and the antenna current, in attempts to get control over linearity - I found that the most reliable place that I could obtain an analogue of the antenna current from was the supply current to the oscillator - this despite having built the antenna-side monitoring into the circuit.

One of the really nice things about this method is that its DC / Low frequency.. On antenna-side monitoring, one is dealing with HF - you dont want to put a big sensing resistor or introduce anything which will affect stability in any way - you need to rectify and filter the signal - and this leads to using expensive opamps and/or tricky high gain transistor / FET circuits.. And the signal change is extremely small - one wants to accurately and actively linearize over a change of only a few kHz of oscillator frequency, )which equated to about 1mA variation or about 5% in my circuit) but I never got it to work reliably / repeatably.. I abandoned the above approach - it was just too fiddly and expensive.. But the supply monitoring was easy and reliable, I could just put a small resistor in the supply, and use a low cost R-R opamp to amplify the Vdrop - Null it, and get a good voltage swing corresponding to antenna current. (and here I have given away one of my earlier schemes for direct-to-voltage AFE, LOL ;0)

Fred.

ps - please feel free to email me if I vanish - Any day now I will be gone.. it may take a while for me to reply to email but I will eventually..

ps.. "Even defining what resonance actually means in this scenario (in terms of phase detection) is problematic."

Conventional theremins do not operate at antenna resonance - they operate between the "natural" (free running) oscillator frequency and the (variable) antenna resonant frequency..

So yes, there is a problem if you are using phase detection - if you lock onto the antenna resonant frequency, you will get no equalization effect - you would need to lock onto the oscillator frequency I think, because this frequency is being "pulled" and is the one which is linearized.. but there wont be any reference phase to compare against (?) or perhaps the phase of the antenna resonator could be a 'reference' against which the oscillator is compared (?).. I havent thought about this at all - may be nonsense!  ;-)

Since this thread has a lot of terrorist chatter about parallel tanks, I want to bring up something I've been thinking about a little.

Given a parallel tank Theremin with series EQ coil(s):

1. The combined series EQ coil and antenna capacitance has a resonant frequency.

2. The tank LC has a resonant frequency.

3. There is a third frequency used for heterodyning.

4. The 1 & 2 frequencies must maintain a critical relationship to improve linearity.

It seems that unless one accomplishes 4 by altering the series EQ L (FredM's saturable reactor trick), or via some small stray capacitance added to the antenna (FredM's variable antenna length trick) then the interactions become rather complex and tricky.

Has anyone ever tried using an AM radio dual ganged capacitor for the tank and heterodyning oscillator, effectively lashing both together?  This way frequencies 2 & 3 would track, the heterodyned result could be used more directly when tuning, and neither the EQ L nor the antenna C would need to be altered.  Those little trimmers that are generally provided for free on these types of variable capacitors might come in handy as well.  Gear it down / add series capacitance to lower the sensitivity to adjustment.

Hey dew,

There are only a few of us left so my chatter must be considered input.

The post by the coal miner above stated I live outside the box or something like that. The theremin antenna is something I have made interesting observations, for over ten years.

In math it would be better to define it with geometry rather the algebra.

If the inside pitch field is too cramped you need to reduce the diameter, if the outside low musical notes are spread out you need to make the wire longer.  If you want the outside low notes even tighter you bow the the top and bottom of the antenna towards the Thereminist. It really is that simple.

Your product has a market if the Awakening occurs. That brings up another thought, there is someone else in the TW group that is able to keep a very good secret. (-'

Christopher

"Has anyone ever tried using an AM radio dual ganged capacitor for the tank and heterodyning oscillator, effectively lashing both together? " - Dewster

I see no theoretical problem with this, but do see a practical one.. as for "Has anyone ever tried" I dont know.. ;-)

The practical problem I see is with regard to the preset ratios available on variable capacitors - Good dual variables are hard to come bye anyway, and the cheap plastic mini tuners are just horrible.. You need the change in tank resonance to track the change in antenna resonance - but in order to get the equalizing effect the antenna resonator must be acting in its inductive zone on the tank resonator - which means the tank C must be extremely large WRT the antenna C (2n2 vs 12p in EW ?) , and the tank inductor must be small WRT the the antenna inductor (100uH vs 30mH in EW ?)..

If one could get a dual variable capacitor with one varying by say 2pF (antenna side) while the other varies by say 100pF (tank side) then perhaps.....

Thing is, you dont need this - If one has a fixed (trimmable) variable oscillator, and tunes the theremin by varying either the antenna capacitance (affects sensitivity) or the antenna inductance (easier with saturable reactor - but these are bothersome as they must be ovened) then after manufacture calibration, tuning will not affect the theremins INTRINSIC linearity...

But there is nothing you can do about linearity distortions caused by grounding / objects / moon phase etc.. And, as I have recently discovered, some people cannot play a linear theremin! - They dont like it - Epro or Skywave, they hate it - but they can play a standard EW and less linear theremins... I cannot play the EW, but can play the Epro, Skywave and Tvox..

And I have concluded that the search for a universally acceptable distance <-> pitch response is a completely futile one - Obviously if notes are spaced too close together the instrument becomes harder to play for everyone - but the prefered actual linearity seems to be a personal thing..

The latest feedback was my final confirmation - I have had numbers of people giving me feedback which I doubted, but this particular theremin I was almost sorry to ship - it just felt absolutely 'right' to me.. I was told by the recipient that it was unplayable - "just like the Epro" LOL ;-) "too linear"...... So Im happily taking it back! ;-)

What can you say ???   ROFLMAO!!!

 Oh - this particular Skywave has a directional antenna which confers greater linearity than the standard - its a passive directional antenna, not particularly directional, about 70% "attenuation" of 'wrong side' capacitance - the shield is connected to ground via an inductor about 1/3 the inductance of the EQ inductance - so there is an even quirkier antenna network than usual! - Simulation was a nightmare - and the irony is that I wasnt even trying to optimise the linearity!

"If one could get a dual variable capacitor with one varying by say 2pF (antenna side) while the other varies by say 100pF (tank side) then perhaps....."  - FredM

I was thinking more along the lines of ganging together the fixed local oscillator and the parallel tank.  They often have identical inductors and capacitors.  Though there would still need to be some kind of fine offset adjustment between them to get the audio null right.

This way the EQ coil could remain fixed and antenna capacitance could vary all over the place.  Tuning for linearity might be much more intuitive.

"And I have concluded that the search for a universally acceptable distance <-> pitch response is a completely futile one - Obviously if notes are spaced too close together the instrument becomes harder to play for everyone - but the prefered actual linearity seems to be a personal thing.."

My prototype seems to have just one null point setting where mid field and far field linearity are quite good.  Setting it anywhere else (e.g. to change tone spacing) throws this completely out of whack.  I wonder how much this trips up users?

"My prototype seems to have just one null point setting where mid field and far field linearity are quite good.  Setting it anywhere else (e.g. to change tone spacing) throws this completely out of whack.  I wonder how much this trips up users?" - Dewster

I may well be wrong, but IMO this is the primary flaw in the whole tuning / linearity issue - Tuning the reference oscillator rather than tuning the variable oscillator throws out a large % of any fine tuning done to get best linearity.. this is why I went the route of tuning the antenna rather than the far simpler manual adjustment of the reference frequency.

"I was thinking more along the lines of ganging together the fixed local oscillator and the parallel tank."

Yeah, good idea! - I hadnt thought of that option.. But I am not sure that doing this would improve linearity..

To be honest, I am no longer sure about anything WRT optimising linearity on a distance based long range capacitive sensing system - I was sure about lots of things, but the majority of my experiments have now proved unrepeatable, ( and I am inclined to walk away from the whole matter as my 'ribbon' style sensor solves all the problems.... Oh, its not the dramatic visual control system where the audience sees a player playing thin air.. but everything else is sonically identical, and the instrument is easy to build, easy to play, insensitive to long range capacitive or EM influences, and doesnt require hours of calibration to get it linear ) .

My hypothesis has been that there is one variable oscillator combined with one antenna resonator tuning which gives optimum linearity over a usable (player controlled) variation of antenna LC, and that if changes outside this optimum antenna C occur due to environment or whatever, it is the tuning of the antenna LC which needs to be changed in order to return the theremin to its optimal operating point.

For this reason I went the route of changing the antenna L manually (or in my more ambitious ventures by automation) and having both pitch and reference oscillators set at initial calibration and fixed at those tunings.. In my view, if the antenna (background) C changes due to whatever, it will cause the response of the antenna circuit || Oscillator circuit to deviate from whatever operating point was set on calibration - Adjusting the reference frequency does nothing to correct this, all it does moves the beat frequency or null point to a usable position.

Thats my hypothesis anyway - As to whether its valid, well, I thought my experiments showed that it was.. but like all hypotheses, they need testing - and some of my tests have been (to say the least) inconclusive.

Fred.

"I was sure about lots of things, but the majority of my experiments have now proved unrepeatable..."  - FredM

Perhaps you could publish your findings in The Journal of Irreproducible Results.  ;-)