Goals for a TW Theremin

I agree. I'd mentioned modular a bit earlier. Another advantage is in troubleshooting.

Also, my (free) PC board CAD software can't handle a big board :)

I cant see a need for anything more than a radio frequency probe connected to a good volt meter for adjustment. A scope will be needed during design and prototyping, but not for final adjustment.

A cheap one would come in handy for repairs. I'd recommend having one on your bench anyway. They are just too useful.

3 feet might be too big. Mine will do about 20 inches. I think an EW is near 18 or so.

As for other design ideas, I like buffering the oscillators with an emitter or source follower to improve bass, and maybe take it a step further and add a coupling control, like Clara's theremin.

My etherwave can do around 3-4', i usually tune it to about 2 feet and a bit to null. Which i think is fine, but there are bigger men than I. 18" would only give about 3 or so octaves if they were perfectly linear and playable.

Really? I don't remember where I heard 18". I'm old and I get confused easy :)

Mine will stretch out further, but I keep it down to 20 or so to shut it up while I'm standing near it. That's actually a guess. I never measured it.

To be honest, most of my experience is in wrangling radio frequency's and sound as a ham radio operator, broadcast engineer, tube guitar amp builder and guitarist.

A theremin uses exactly the same operating principals as my radio equipment, including the antennas, but I have only played with my Kustom and a Keppinger.

I'll try, from now on, to only comment on things in my field of experience.

1: "We also discussed adding harmonics, and I remember you commented here (I don't remember where) on Bob's use of IF transformers to get harmonics."

2: "Additive synthesis at RF sounds good to me. I have a partial prototype on my bench, but I've set it aside while I work on my Kep."

3: "The volume section could be made adjustable for range and "snappiness" with op amps."

4: "Linear range is assumed in a quality instrument."

- Above from w0ttm.

1+2: I cannot help on these points for commercial reasons - other than give hints.. This area is quite mind bending - but this schematic http://www.thereminworld.com/files/Pages/25/images/schem.jpg gives a clue - look at the mixer!

The coils are the clue - but its the principle which one can derive from this, and some lateral thinking, which can eliminate the coils and replace them with quite simple electronics.. However, even doing that, going to a full additive engine is, I think (no - im sure ;-) way beyond the scope of the TW theremin.

5:"If the theremin is constructed very modular-ly, say on 6 separate pcbs, Each function of the design on a separate board, it would allow the design to evolve better, and be more customizable. " - Nieradka

Yes, this is a good approach - it allows sections to be 'upgraded' without requiring a complete rebuild.. and allow several 'standard' variants - one could have a CMOS oscillator variant for example.

3: "Snappiness" is related to the exponential nature of volume hearing - it isnt simple to alter the law without some non-linear circuitry.. Log VCA's can be useful here, as can multipliers or log amps..

Fred.

Abso-dad gum lutly!

I guess I wasn't clear on the word "linear". When speaking theremin, it means something different to me than in the engineering sense. I refer to the pitch field as an example. It should be linear on the musical scale, but an engineer would not see that as linear at all :)

It was also in he middle of the night when I said that, so I ask forgiveness....

If memory serves, a reasonable log amp can be made with a transistor in the feedback loop of an op amp.

I'll be back after the doctor removes my foot from my mouth....

One more thing....

I think we can get decent sound by adding just a few harmonics.

In simulation, and real silicon, I got a reasonable, if a bit lumpy ramp with just the second and third.

@dewster and nieradka

Please e mail me.

randkg (at) att.net

Nothing serious. I'd just like to have you in my contacts.

From: http://www.thereminworld.com/Forums/T/28786/design-development-rules-and-tools?last=True w0ttm said:

" I'd also like to suggest the boards be single sided for cost, service ease, and home builder simplicity. I guess I should add that to the design goals thread."

May I suggest a mid-way solution (one I often use) - Design the board as double sided, but keep the tracks on the component side straight.. Do not use component pads to carry connections from top to bottom layers, use seperate (large - enough to drill easily and solder TC wires into) vias in every case.

With the above, the board can be manufactured as a double sided PTH board.. but it can be prototyped or built as a single sided board as well - simply by placing straight TC wire where the top (component side) tracks would be. If these wires are placed before any components are fitted, they can be tacked to the PCB with adhesive, giving a good reliable double sided board equivalent.

I drill out the holes for the wire links first, tightly fit and solder the links, spread a thin layer of epoxy over the component side covering the links, then when the epoxy is hard I drill the remaining holes and fit the components. Sometimes I print a component overlay on paper which I saturate with epoxy and then place this on the board over the links, when set I then drill the remaining holes.

From a production perspective, limiting the design to single sided is, I think, too restrictive - There is almost no price difference between single and double sided boards, and wiring links is a royal pain in the backside - takes a lot more time than fitting components, and unless one has done it many times before, is likely to look untidy at least, and be a potential source of problems at worst.

I agree that for home made boards, double sided boards are quite a lot more effort, and one needs (unless one is extremely well equipped with PCB making kit) to hand wire the vias, which is nearly as much bother as putting links on the board.. I make my DS boards by etching two seperate boards (one thin flexible and the other an 0.8 or 1.0 mm) and bonding these with epoxy prior to drilling - but its a lot of work, and registration of the 2 layers must be perfect.

Fred.

 While on the subject of layout - test points are an important thing often overlooked

"Several professional theremin designs achieve a decent but not perfect linearity without linearization coils : the tVox tour and (even better) the Henk theremins. In both cases, very low oscillator frequencies were chosen (135kHz for the tVox and 145kHz for the Henk). The rest is done by an intelligent choice of the oscillator's L/C ratio and (here comes the huge difference between both instruments) the ratio between the pitch rod's static capacitance and the hand capacitance. The tVox works with a big metal ground plate which greatly increases the static capacitance and it has an unusually thick pitch rod to "see" more hand capacitance while the Henk has an unusually long and thin rod to minimize both, the static and the hand capacitance.

I think (but that's my personal opinion and experience) that a design in which the pitch rod is part of the main tank circuit and thus without separate series resonant circuit would be more robust and easier to build and to tune."  - Thierry

Fascinating Thierry, thanks for that info!

One thing that is bothersome in my (mostly) digital Theremin AFE is the double resonance of the tank & linearizing inductance - touching the antenna can make the PLL jump out of range and stick there, and I suspect there are perhaps heating issues with the linearizing coils inside the antenna. 

I zeroed out the linearizing inductor in my simulation spreadsheet and found roughly the same DFOM (digital figure of merit - useful sensitivity, essentially) with 39mH (Bourns 8250-393K-RC) & 10pF in the tank operating at 177kHz.  It produces a couple of hundred volts at the antenna with 4mA drive and has a single quadrature point.  The inductor is around $1, I think I'll order a few.

As far as specifications baseline, it seems like that needs to be nailed down first, so, if I may, here are my proposals:

Modular

Heterodyning

4+ octaves (and the bass octaves playable like a modded etherwave.)

At least, tunable to a 75cm playing field. 

At least 3 of those octaves can be tuned to a near perfectly linear pitch spacing. +/- 10%

Timbre is consistant across those 3 octaves. 

Volume control should be adjustable from a 7cm to 25cm response range, or wider. 

Thermally stable after a ten minute warm up, if the ambient temp changes by < 5 degrees C.    (should be playable in ambient temps from -10C to 40C) 

Daily tuning should not require more than one knob. 

Design is rugged enough, for performance and travel. Should be built for a 25 year service life. 

Capabilities to add pitch preview, effect loop, mute switch, CV. Selectable output from line level to instrument level. 

Cost less than $400 USD to build. 

Solid state, analog. no surface mount. Build-able by someone with moderate experience, and tools. No double sided PCBs unless unavoidable. 

But most importantly, what do we want it to sound like? Adjustable, sure, but it should do one tone very well, do we want it to sound like: an RCA, a tvox tour, an epro, or? Perhaps we should come up with the waveform we want first, and then think about implementation and everything else.