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

Fixed the coil winding worksheet of my Theremin simulator spreadsheet (wouldn't converge for some inputs) but I'm not releasing it until I fix the main sim parts too to use a transformer tank. 

Playing with it: using a PVC schedule 40 2" (OD=2.375"=60mm) wound with a single layer of AWG 32 (~0.2mm dia) wire, 380 turns takes 72 meters of wire and gives 5mH inductance and 39 ohms.  The winding is 76mm long on the pipe (probably a bit longer due to wire insulation).  I'm reluctant to go smaller than AWG 32 due to handling trouble and resistance (Q) and I don't want to go above 60mm diameter because I can get heatshrink tubing to fit over this.  A single winding layer minimizes capacitance (highest SRF) and keeps the thing from needing bobbin flanges.

Questions:

1. Is heatshrink over the windings a bad idea?  Is varnish better?

2. Is PVC a bad coil winding base?  Would cardboard or similar be better in terms of temperature stability?

3. What do people use for terminals?

4. I can get extra thick varnish insulated wire (diameter, double=0.0094" vs single=0.0085") is there any benefit from thicker or thinner varnish?  Should I be worrying about voltage breakdown if there's ~100V or so across the winding?

5. The tank coil will likely be a transformer.  I'm wondering what the best way to include the ~0.1mH tank winding should be.  From the Inca simulator software it seems like the winding should have dimensions as close to the other winding as possible, and in as close proximity as possible for best inductive coupling.  Should I have a thin layer of something between primary and secondary due to the voltages involved?  If so, what should that be?

I guess I'm getting tired of shopping for coils that don't exist or have too many missing specs.  Winding my own may be less painful in the long run (famous last words).  Only gonna do this for the pitch side though, Bournes 6300 or similar should probably work fine for the volume side where drift isn't such a big deal.

Dude, just varnish it.   Dont cover up all your beautiful hard work!- also then you can make them any size you wwant/need.

PVC should be fine. so are cardboard mailing tubes.     For terminals, I like to  solder the wire to a brass washer and attach it to the form via a machine screw and nut (on outside) to create a lug (brass of course)- add a second nut to secure your slit, ring (or curled bare wire ;) terminal connector.

For my RCA replica osc coils, i use paper-backed fabric for bookmaking and scrapbooking.  Dick blick has it for sure, I imagine most craft or art stores would too.  I precut everything, hit the backs of the fabric strips w/spray adhesive and then wind em up- wrap it twice to equal the the fabric layer on the original RCA's.

Personally I'd get single build(1 coat) wire.   Someone more educated may recommend otherwise.

Welcome winder.

Also you will want to make a simple winder w/ a counter.   I wouldnt go nuts - having it motorized is completely unnecessary.   A hand crank works perfectly. 

Thanks Chobbs!  I'll give your DIY lug a try.  Was thinking of hand cranking it, and a counter probably isn't necessary because the actual inductance value is uncritical (I can probably use a micrometer and tape measure to estimate windings).

I can't go too much larger in diameter and still fit in the case I'm thinking of using (4" vinyl fence post cover).  Playing with the winding formulas, 60mm diameter is fairly optimal (in terms of minimum wire used & highest Q) for a range of inductance values using this wire gauge.

What's the paper-backed fabric for?  Does it form the tube itself or is there a tube underneath?  Do you use anything to secure the windings while you are doing the winding?

Do you use special varnish made for coating magnet wire (one eBay supplier sells both wire and varnish)?  Theremin used some kind of clear plastic tube to cover the coils, correct?  I'm thinking heat shrink might protect the windings from nicks and such during final assembly.   There should be a fair amount of empty space inside the case on the pitch antenna end, plenty of room for a couple (tank & EQ) of air coils.

"What's the paper-backed fabric for?  Does it form the tube itself or is there a tube underneath?  Do you use anything to secure the windings while you are doing the winding?"  - me

I see from this page that the paper-backed fabric is used to space coils wound on top of each other.

Measuring Inductor SRF

Seems notoriously difficult to measure!  You're looking for a tiny capacitance (~pF range) across a low impedance.  Stimulating the coil with a function generator & series resistor and looking for resonance across the coil with a scope loads the coil with the scope probe capacitance, confounding measurement.

Interesting site where they measure Inductor Self Resonant Frequency (winding capacitance) via several methods, I particularly like the grid dip meter approach:

http://www.cliftonlaboratories.com/measuring_distributed_capacitance.htm

Extremely interesting paper on SRF:

http://www.g3ynh.info/zdocs/magnetics/appendix/self_res/self-res.pdf

He shows the historical angle, and how much crap there is out there in terms of theory.

I connected one end of an inductor to the hot output of a function generator, with the other end of the coil just hanging out in space unconnected.  I set a scope probe several inches away from the free end, and connected the scope probe ground to the function generator output ground. 

With a Bournes 6310 (50mH) I see a very distinct resonance at 460kHz, which works out to a bit over 2pF.  (Spec for this part is 330kHz.)

When I do the same experiment with 200 feet of 30 AWG on a roughly 1" diameter by 16mm high plastic spool (it was sold on) I don't see any obvious resonance.  It measures 12.36mH on my LC meter.  Hmm.

Another measurement method I saw recommended in a forum somewhere was to put a largish capacitor (> 10x the SRF capacitance) in series with the coil and use a capacitance meter to read it.  I get 1pF this way with the Bournes 6310, which I believe is too small.   For the 30 AWG coil I get 77pF!

[EDIT] Second try with the 30 AWG coil shows a slight resonance at 160kHz, which translates into about 80pF.

"Seems notoriously difficult to measure!" "He shows the historical angle, and how much crap there is out there in terms of theory." - Dewster

HeyHo.. Looks like I might have been responsible for putting some cr*p "out there" with my simple oscillator "SRF tester" circuit..

Only glanced over the articles, but it seems I have missed the complexities - Surely, by determining the resonant frequency of an inductor (with its parallel capacitance), factoring in the capacitance of your test probes, one can get a reasonable "hit" at its SRF ? .. Or driving the inductor from a swept signal, and monitoring current, one gets the impedance resulting from the sweep.. from this (looking at X below resonance and X above resonance and X at resonance) one should be able to deduce resistance (X at resonance) and capacitance (From X above resonance) and inductance (from X below resonance).... also, the fact that Xl increases linearly with frequency whilst Xc decreases exponentially with frequency, it should be possible with some computation to be able to recognise and seperate the component impedances and therebye deduce the values of all the components - (?)

Perhaps I should read those articles.. but I will leave that for some other day! ;-).. Dont want to get back to coils-et-al right now.. I have given up on that in frustration and am resorting to the sledgehammer of analogue computation and dynamically varying the oscillator frequency.

fred.

"HeyHo.. Looks like I might have been responsible for putting some cr*p "out there" with my simple oscillator "SRF tester" circuit.."  - FredM

Oh not at all!  The crap out there is earlier bad papers that Medhurst had to resolve,  people thinking interwinding capacitance in single layer coils exists, and not taking coil form dielectric into account.

"Dont want to get back to coils-et-al right now.. I have given up on that in frustration and am resorting to the sledgehammer of analogue computation and dynamically varying the oscillator frequency."

Finding appropriate Theremin coils has been quite a frustrating experience for me.  Months of shopping and investigation with almost nothing to show for it.  I'm now considering winding my own air coils and transformers, which is not where I want to be but where almost everyone seems to end up (even those not interested in building historical reproductions).  If only the magnetic permeability of free space were a bit larger...

" If only the magnetic permeability of free space were a bit larger... " - Dewster

If it was, I think we would have a lot more to worry about! LOL.. Lets see.. Earths magnetic fields for starters.. ;-) ... Then, for theremin designers, having iron anywhere near your inductors might pose a problem - not to mention having to keep your Ref and Var oscillator coils a lot further apart.. ;-)

For analogue theremins, the Bourns or Hammond inductors are, IMO, good enough - I have been through the "wind your own air inductors" phase but really saw no advantage over the Bourns 6300 series.. As I see it, the only advantage is less thermal drift, and this can be compensated when using ferrite (Bourns 6300) with careful selection of capacitors used in the tank.. The only reason I will not be using this method is that I have entirely changed my oscillator topology, and am not using equalizing inductors at all - thermal drift is now a function of the 42IF106 transformers and semiconductor junctions, and I must actively compensate this by monitoring the temperature and/or by ovening the circuit.

"people thinking interwinding capacitance in single layer coils exists"

It doesnt ?! .. But... Wires next to each other must have capacitance.. Take 3 turns - the middle turn must have capacitance to the turn on either side ? - Surely the equivalent schematic nust be a load of tiny inductances in series, each with a tiny capacitance across it........... Oh yeah! I see it! - The tiny inductances add, and the series capacitances each reduce to 1/2 that of one turn in a compound manner .. So the total "interwinding capacitance" reduces to far less than the capacitance between two turns.. !! -  Well ill be stoned!

"not taking coil form dielectric into account" - Hmm.. This would surely apply to single layer coils as well (?) Lets say that a cross-sectional angle of 45 degrees exists through the coil former coupling one end of the coil (capacitively) to the other - and lets say the former was solid ceramic with high dielectric - surely one would effectively get a capacitance due to the distributed angular couplings through the former... Not sure this would be "interwinding capacitance" - But it would be a capacitance in parallel to the inductance.. (?)

Really not sure I know what im talking bout here - just thinking out loud....

Fred.

"people thinking interwinding capacitance in single layer coils exists"

"It doesnt ?! .. But... Wires next to each other must have capacitance.. Take 3 turns - the middle turn must have capacitance to the turn on either side ? - Surely the equivalent schematic nust be a load of tiny inductances in series, each with a tiny capacitance across it........... Oh yeah! I see it! - The tiny inductances add, and the series capacitances each reduce to 1/2 that of one turn in a compound manner .. So the total "interwinding capacitance" reduces to far less than the capacitance between two turns.. !! -  Well ill be stoned!"  - FredM

Not to mention each winding is essentially shorted dead out to adjacent windings by a loop of wire quite a bit shorter than the operating wavelength - where's the separation of charges that causes capacitance?  And theories often don't take the magnetic field into account.  Etc.  That paper by Knight is a total hoot to read (on top of being super informative).

"not taking coil form dielectric into account" - Hmm.. This would surely apply to single layer coils as well (?)

Yes, definitely.  This is one of the problems with Medhurst's findings.

[EDIT] "thermal drift is now a function of the 42IF106 transformers and semiconductor junctions, and I must actively compensate this by monitoring the temperature and/or by ovening the circuit."

My problem is I can't tell if I'm on a wild goose chase after having my mind poisoned by articles like this, or if the hams have a lot to teach me and I'm crazy not to listen.  I'm guessing not asking for drift in the first place (use NP0/C0G caps, air core inductors, and low drive current) is a good place to start, particularly when I can't count on two fairly identical oscillators drifting together and cancelling it out.  But like you I intend to bludgeon anything else that might stay dead with modern technology.