DIY Alternatives to Bob Moog's EM Etherwave Antenna inductors

Posted: 7/19/2018 3:10:06 PM
Corehbola

From: São Paulo, Brazil

Joined: 7/19/2018

Hello TW folks,
My daughter shown some interest in the Theremin, and me, as an electronics tinkerer, thought... Why not building one?
Soon got to know ThereminWorld and read as much as I could to make an educated choice of what Theremin to try to build.
I decided for what it seemed to be an authentic sounding, and yet not very complicated in design, the EM Bob Moog's Etherwave.
From the beginning, I had already a guess that it wouldn't be easy to get the exact antenna inductor parts, but I didn't think it would be THAT! difficult.
It seems I decided to jump on this wagon just a few months too late... as those inductors apparently went out-of-production by Hammond at the beginning of 2018... and I just started this July (7 months too late, to my misery).
Now, I think it's already too late to step back and choose another design and I don't like to give up on anything...so I'm really decided to find a solution for the unavailability of those inductors.
My first idea is to build them (the inductors), as similarly constructed as the original ones as I realized from reading here that Theremins are tricky all by themselves, so it seemed to me a good idea to keep as close as possible to the original 3-pi construction over a ferrite rod.

So, after this long introduction here is my starting point:
1. I got a bunch of 24mm long, 6mm dia ferrite rods. I don't know their actual composition and their permeability, but I can figure it out by trial and error, winding a few simple 1 layer inductors, and measuring its inductance.
2. I don't have a honeycomb winder, so I'll have to wind by hand the 3-pi coils on a form.
3. At this point, I don't have any Litz wire, but I got 250g of AWG 37 enameled wire.
4. I`m planing to machine a few plastic forms (from nylon or teflon rods), to allow the hand-winding of the n-pi coils

And here are my initial questions:
A. do you think that it's feasible to obtain reasonably working inductors, using the cores above, hand-wound AWG37, over the plastic forms? What I mean is: will I be able to achieve the low self-capacitance needed for getting the self resonant frequency high enough for them to work?

B. Do you think that some deviation from the original 3-pi construction could help? For instance, I think I can machine as much as 5 narrow channels, of about 1.5mm wide, separated by four 1.5mm walls (making up for a 5-pi construction).

C. What about the 37AWG wire...Do you think it would help if I twist it in a pair or a triad, making up for a poor man's Litz wire? Or maybe trying to do the same (a triad), using thinner wire (like 40AWG).

D. About the number of inductors... To help alleviate the build up of self capacitance as the number of turns increase, would it work, or would it help to build more inductors, of lower inductance each one? (I know ahead that I would need to find ways to minimize effects of mutual inductance between them).

E. In another attempt to allow for smaller inductors (and avoid the self-capacitance problems inherent to the high value), would it be valid to:
E.1. Increase the operating frequency (say... in about 50%).
E.2. Add some capacitance to the antenna? (e.g., I noticed on some pictures of Moog's Etherwave plus, that it has a square of grounded aluminum foil, and the Pitch Antenna connecting wire passes over it... I suppose that this wire alone is probably adding a few pF to the antenna, allowing for the same given resonant frequency using less inductance? Does that have a too big impact on the pitch field?

Sorry for the long post and too many questions...

Regards, from Brazil,

Fabio


Posted: 7/19/2018 7:37:51 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

You don't need Litz wire, the frequency of operation for analog Theremins is low enough that the skin effect won't substantially lower the Q.

There's nothing magic about the Hammond inductors, other than that they're fairly small physically.  As a result there is some temperature dependence, some magnetics loss, and some self-capacitance, but all-in-all they work OK for this purpose.  Though several have to be used in series to lower the bulk self-capacitance, and therefore keep the self-resonance sufficiently below the operating frequency.  You could do better with a single layer air-core, though it will be much larger.  One way around that is raising the operating frequency and lowering the absolute sensitivity with C padding on the antenna side, or through the use of digital dividers.

If you wind your own coils, try to use an AWG that isn't too fine to handle, otherwise it will drive you crazy.  Get a single coat type (for tighter turns and higher L) that the coat melts with solder (so you don't have to manually strip it).  Unless you know the formulation of the ferrite (or took it from an AM antenna or something) I'd avoid it.

Maybe try to get your hands on 2nd hand chokes, or some new ones from NEOSID (German firm) part number 00612232 (4.7mH, 700kHz) - I have no experience with those, though the few specs they provide look promising on paper.

The EW isn't the thing I'd build for my first Theremin, nor maybe for any of my Theremins, as it is difficult to tune and therefore own.  Either find a design that doesn't use series EQ inductors, or stick a tiny-valued variable C on the front panel from the antenna to ground.

If you really want the EW circuit buy the kit or bare board from Moog, it will save you a lot of trouble and heartache.

Theremins are kind of a can of worms, common magnetics aren't designed for them, and you make one small decision and it can paint you into a very difficult corner.  You have to keep all your options open, even the ones you don't know about :-).

Posted: 7/19/2018 9:37:25 PM
Corehbola

From: São Paulo, Brazil

Joined: 7/19/2018

Dear Eric,

My replies / comments, right after your quoted ones.

>You don't need Litz wire, the frequency of operation for analog Theremins is low enough that the skin effect won't substantially lower the Q

Thanks for that... it gives me some reassurance to know that I have some good chances to DIY those inductors without needing much more than what I currently have.

>As a result there is some temperature dependence, some magnetics loss, and some self-capacitance, but all-in-all they work OK for this purpose. Though several have to be used in series to lower the bulk self-capacitance, and therefore keep the self-resonance sufficiently below the operating frequency.  You could do better with a single layer air-core, though it will be much larger.

OK. I've seen in other discussions that the temperature coefficient is an issue with the ferrite cored inductors...

I'm taking your words into consideration... Nevertheless, since I'm just started and cannot say I`m already frustrated, I will invest a little bit more time on this path... (not much).
For the moment, I will try at least to build one of each of those inductors and test them on the bench. I have an oscilloscope and a function generator so I think I can characterize them well enough.
If I see that I can get my DIY made inductors to perform reasonably well, I will stick to my original plan (and to my already purchased parts :-)).
If I see I cannot get the self resonance high enough without too much hassle, I think I will abandon this wagon altogether and go for another design.
If that will be the case, I'd love to hear a suggestion of a good sounding Theremin that won't have all this high-value inductor trouble.

>You could do better with a single layer air-core, though it will be much larger.

I will take that in mind... it may become my plan B! Not that I`m dismissing your suggestion, but right now I don't feel desperate enough to engage in winding a big pipe, not only for its big size and having to build a large cabinet to fit them all inside, but also because it's difficult to wind such a large tube by hand (and well, I cannot think of a practical way of winding them on a machine either).

>One way around that is raising the operating frequency and lowering the absolute sensitivity with C padding on the antenna side

This sort of answers my questions E.1 and E.2... It's good to know that it's feasible to minimize, at least by some amount, the pain of sourcing such big value inductors.
I`m already happy if I can get them down to 2/3rds and, if I can cut them by half, that would be like winning the lottery.
Of course I don't assume this won't have its side effects... I`,m old enough to know that things usually are the way they are for a reason, and specially in an art where so many people already put so much effort for so long a time, if there was any magic to be discovered, it would have already... and I wouldn't be having this conversation...
Still... this is an evolving scenario... Until not so long ago the coils were readily available so, why one would have bothered to change something that was known to work? There still may be some room for tweaking, without too much compromise in quality.

>or through the use of digital dividers

Now you got me curious... I fail to grasp where digital dividers would be of any help here... Can I conclude (from this suggestion) that the main reason why this circuit is not - usually - operated at a higher frequency (thus allowing for smaller inductors), is that the beat products would become too wide (too much absolute frequency change for a small C change) and therefore, too difficult to control (by the player)... and by using digital dividers I could operate the V.P.O at a higher frequency, then dividing it, and then mixing it with the F.P.O? Or operating both the VPO and FPO at higher frequencies and then dividing the beat frequency?
Or did I get it completely wrong?

>If you wind your own coils, try to use an AWG that isn't too fine to handle, otherwise it will drive you crazy.  Get a single coat type (for tighter turns and higher L) that the coat melts with solder (so you don't have to manually strip it).

Noted...It makes every sense.
Still, since I`m not giving up (yet) on trying the Ferrite-Core, and since I`m using a form to hold the bobbins from collapsing, I plan to wind them just like if I was winding a spool of kite line. I even suppose that the "fuzzy" winding can help somewhat to keep the self capacitance a tad smaller (than nicely stacked side-by-side turns).

>Unless you know the formulation of the ferrite (or took it from an AM antenna or something) I'd avoid it. Maybe try to get your hands on 2nd hand chokes, or some new ones from NEOSID (German firm) part number 00612232 (4.7mH, 700kHz) - I have no experience with those, though the few specs they provide look promising on paper.

At the risk of having to hear: "I told you" down the road, I will insist a little bit with those ferrites I've got (despite I don't know anything about them).
I hope I will be able to characterize the inductors well enough, quite early in the process, before even trying to build the rest of the Theremin.
If I see the inductors aren't promising, I`ll re-plan the whole thing, follow your advice and probably try a different design.

>The EW isn't the thing I'd build for my first Theremin, nor maybe for any of my Theremins, as it is difficult to tune and therefore own.  Either find a design that doesn't use series EQ inductors, or stick a tiny-valued variable C on the front panel from the antenna to ground.If you really want the EW circuit buy the kit or bare board from Moog, it will save you a lot of trouble and heartache.Theremins are kind of a can of worms, you make one small decision and it can paint you into a very difficult corner.  You have to keep all your options open, even the ones you don't know about :-).

Well, I`m on it more for the journey than for the destiny itself... But of course, I don't intend to be punching knives just for the "fun" of it so, sure enough I will keep my options open... After all, so far I only spent some bucks on the LM13600s, and a few passive parts... all of which could have good use in any other project.
Looking around, I found an interesting approach for a digital design... This one from the Glasgow university: http://www.theremin.us/Glasgow/schematic.gif
Is that one any good?

Thanks for all your input and Rgrds, Fabio


Posted: 7/19/2018 9:51:17 PM
Corehbola

From: São Paulo, Brazil

Joined: 7/19/2018

I don't know what happened to my previous post... The quoted text shown up in gigantic text.
I tried to fix it, by editing, but it didn't get right... Apologies to all.

Posted: 7/20/2018 2:33:58 AM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

You have to edit it with the "reveal codes" button on the far right (less than / greater than symbols or "sharp" brackets).  It's a pain.  All kinds of formatting is inserted when you copy and paste (but not when you drag and drop), and the default insert font is gigantic.  You can always go back and edit your posts here, which is nice.  

Pro Tip: I've noticed that if you paste with "reveal codes" active you don't get all the extraneous formatting!

Sounds like you know what you're doing and have the proper equipment, so ignore much of my waving off.  Though the EW is a rather touchy design, and you might want to track one down just to be familiar with how it works when it's working correctly.  I've tuned a few perfectly fine ones and they can seem almost broken if you don't know what you're doing (and sometimes even if you do!).  Make sure you're using the latest schematic with 3 inductors on the pitch side rather than 4, the old Big Briar (with 4) isn't as good a performer as the more modern EW (with 3).

The problem with any Theremin that has series EQ inductors (such as the EW) is that they have two resonances fighting each other.  This is to increase linearity and perhaps lower relative sensitivity.  But it also means the oscillator has more undesirable modes it can enter into, the EW transistors can reportedly burn up (!) if you don't tune it correctly.  The two resonance frequencies mean there are three overall resonance if you include the fixed oscillator, which can make tuning difficult.  What you often want to adjust for is environmental C that the antenna sees, but you're stuck with adjusting the fixed oscillator instead, so the EQ L / antenna C and variable tank L / C ratio isn't ideal.  This is why Theremin's designs had a variable pad C at the antenna, so you could adjust the issue at the source.  Also, getting the EQ L / antenna C and variable tank L / C ratio right can be rather involved, hence the poorer Big Briar, and the bendable wire and ground plane.

I've heard the Glasgow schematic is crap.  And except for the XOR mixer it's not really digital either as it uses logic gates in linear mode for the oscillators.  Nothing wrong with that BTW if you pick the right logic family (they make logic inverters specifically to form crystal oscillators with this topology).

I believe ILYA designed a Theremin recently that uses digital dividers to allow higher frequency operation, and thus much smaller inductors.  I haven't investigated it much, but I think you can use this approach to mix octaves and to octave switch.

Once you get the operation around 2MHz or so the inductors aren't all that difficult to wind as medium sized single layer air core solenoids, which have fantastic stability, high Q, and very low self C.  But my stuff is almost purely digital and I don't have a lot of experience with analog Theremins.

Posted: 7/20/2018 5:38:46 PM
ILYA

From: Theremin Motherland

Joined: 11/13/2005

hi Corehbola,
hi dewster,


The DIY inductors can really be fit in the EW type circuit design.
Here is the corresponding thread (link).
The most difficult thing is to find a suitable multi section frame and (in general) force myself to do something.
I made my ones from plexiglas.

12 sections at 44 mm total length for EQ coil and 8 sections at 29 mm total length for volume coil. All data are in the table.

My inductors are core free. I guess  the SRF will not be affected too much if we'll add the ferrite road and reduce the number of turns to keep the target inductances (i guess the reduced wire length will be compensated by reduced wave propagation velocity).

Again -- link to article, the Google translation (dont know about Brazilian language, but English translation is more or less correctly).

Posted: 7/23/2018 2:59:03 PM
Corehbola

From: São Paulo, Brazil

Joined: 7/19/2018

Hello Dewster and Ilya,
@Dewster...
Thanks for the tips on ways to use the Web interface editor of TW. I'll keep the advice for when things get complicated and I really need to include inline quotes... Otherwise (like now), I'll just avoid reply with quotes.
It took me a while to grasp the whole idea of the antennas' series inductors and I see your point when you say they are difficult to tune.
I hope I don't burn too much transistors in the meantime :-).
It will still take me a while until I get something assembled and be able to start playing around with it and learn a little bit from practice (as I guess it will be a little bit too much for me to try to master the theory completely at this stage, with little to none practical experience).
I managed to wind my first 10mH coil and it seems I`m going to be successful down this road (although I just saw Ilya's approach and found them tempting...).
This first coil prototype, I wound in 3 sections, over the 6mm dia. ferrite rod that I have, yielded a 550kHz SRF, with 31 ohms DC resistance.
I wrapped the rod with two layers of shrink tubing, and used four squared section o'rings as forms for the three sections.
To get to the 11mH, it took 200 turns per section (600 turns in total), of my AWG37 wire.
I hope that using 4 sections I'll be able to get the SRF up to the 700kHz whereabouts of the original Hammond coils.
Since the 10mH is the most difficult to get right (due to the large number of windings), I think the smaller ones will be easy.

@Ilya,
Thanks a lot for your suggestions and for the link to your project.
I see that it's not JUST an approach for the original EW, but it already incorporates the improvements to allow for lower keys (I understood that the original design doesn't allow very low tones, as the VPO and the FPO locks themselves to the same frequency when their frequency are too close).
Taking the signals out of the oscillators from the other transistor is a clever way to minimize the interlocking effect.
I just wonder if it only minimizes the interlocking, or if it kills the interlocking completely (which is somewhat undesireable, as the Theremin will be always producing some very low key when the hand is far apart from the pitch antenna).
I`m tempted to try your approach to the inductors as well... I`m surprised that you managed to get the values with Air core, and yet the coils are small (much smaller than some picutres I`ve seen, on 4 inch tubing?). Congrats for that.
I`ll be trying to machine the forms and report back here of the results...

Thanks for now, to both of you
Rgrds,
Fabio

Posted: 7/23/2018 4:43:32 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

"Thanks for the tips on ways to use the Web interface editor of TW. I'll keep the advice for when things get complicated and I really need to include inline quotes... Otherwise (like now), I'll just avoid reply with quotes."

It also happens when you copy & paste your own text using CTRL-C & CTRL-V (but not if you have "View Source" toggled on).

"I hope I don't burn too much transistors in the meantime :-)."

The EW oscillators use small signal transistors that can't take much at all in the way of current.  If I were using that topology I'd look into using beefier types, such as 2N2222A or similar.  It's my conjecture that Moog used that topology to "blast" the series inductor with a fairly powerful drive signal, so when the tank LC resonance "fought" with the EQ L / antenna C resonance oscillation could be maintained.  I prefer oscillators that don't draw much current and rely on a high-Q tank to get the voltage swing high, but those may not work well with series inductors loading them down?

I don't know, but I imagine beefier transistors might absorb / withstand ESD events better as well (a definite failure mode for many Theremins).

Random thought: you might try winding a single, scramble-wound "donut" on a short piece of ferrite, and then wiring several of these in series.  The least temperature sensitive ferrite formulations are unfortunately also the least effective in concentrating the magnetic field, so there isn't a ton of coupling going on with several donuts wound together on one longer rod (I measured this at one point and could probably find those results if you like).  Also, if you look at a lot of Theremin home brew designs you'll see multiple EQ inductors in series often mounted / electrically connected via screw-type terminal strips.  This allows easy experimentation and rough "tuning" of the series EQ L value, which can strongly influence the pitch response (the "shape" of the pitch field).  The EW design itself changed the EQ coil count at one point.

Posted: 7/24/2018 11:47:39 AM
pitts8rh

From: Minnesota USA

Joined: 11/27/2015

If you are serious about winding your own Hi-Q inductors, do a YouTube search for  "universal coil winder" or "Gingery Morris coil winder".  It is very easy to build a machine yourself that is capable of winding pi-wound coils in nearly any form-factor you wish.  The winder can consist of something as simple as a few wood blocks and some threaded rod, or you can go all out and make a more versatile mechanical or Arduino/servo-based machine that allows you to dial in the coil dimensions and stand back while it winds the coils for you with minimal intervention on your part.

If you have the capability to easily wind multiple deep pi-section coils, you may be able to avoid the need to use ferrite cores altogether.

Actually, inductors used at theremin oscillator frequencies in the 200 - 400 KHz range would indeed see a mild Q benefit if litz wire was used, and litz is more flexible and holds together better in the pi structure due to the friction of the insulation.  But the higher cost and limited availability of litz wire may make magnet wire the more likely choice, and we know that it works.  Generally a special solvent-based wire coating is used to make magnet wire tacky as it is wound to enhance the stability of self-supporting pi sections, although you can run standard magnet wire over a block of rosin or even over a fast-drying coating such as shellac or lacquer. Alternatively, if you have a lathe or the capability to machine grooves into a suitable form material, winding the coils into the grooves will prevent any danger of the walls of the pi-structure collapsing. Wall-less coils made with standard magnet wire can be strengthened with a few drops of thinned shellac, lacquer, or Q-dope. 

Examples of winders, from simple to fancy:

https://www.youtube.com/watch?v=Jv7Gsx-kuUw&t=2s

https://www.youtube.com/watch?v=FIOocMoRsYQ&t=133s

https://www.youtube.com/watch?v=tg67CvbcHaU&t=234s

https://www.youtube.com/watch?v=SD8is-Lwkmo&index=2&list=FLrR-dCa1akB8C307dxAoUWA&t=0s

       

Posted: 8/31/2018 12:58:20 PM
Newbie

Joined: 8/31/2018

Hello. 
I have similar problem. I read many posts where the only recomended coils are J.W. Miller 6306 or Hammond 1535G. Unfortunately both of them are obsolete and no longer produced. I don't want wind coils on my own, so I made some research and found Bourns' inductors:

8250-103K-RC L=10 mH, Q=70, SRF=1300 kHz, RDC=93 Ohm
RL262-472J-RC L=4,7 mH, Q=80, SRF=800 kHz, RDC=32 Ohm
RL875S-272K-RC L=2,7 mH, Q=95, SRF=1000 kHz, RDC=8,5 Ohm

Are they good enough? Original inductors had higher Q but lower SRF. Because of the serial connecting total inductance is roughly the same.
If these don't match, then what should I pay attention looking for another ones?

You must be logged in to post a reply. Please log in or register for a new account.