Completely newbie planning to build EM theremin

Ok -

The advice I gave about CMC's is probably not that good - particularly if one is looking for more than 20mH.

With Murata sectional wound CMC and series wired inductance of about 25mH, the SRF was about 120kHz.

I only have one bigger Murata CMC (47mH / winding - measured at 56mH) and the SRF for one winding was about 50kHz.

In my saturable reactor experiments using PLA10's, the CMC's were much smaller - except for the AC blocking CMC, which was being used as a CMC..

Fred.

Once again, the Bourns / JW Miller 6300 series looks like the best bet - a bit expensive to get 100mH though.. the 50mH part is rated at 65mA 127R, two in series..

"With Murata sectional wound CMC and series wired inductance of about 25mH, the SRF was about 120kHz." - FredM

If the inductance truly is 25mH, and f=1/(2pi*sqrt(LC)), then the SRF equivalent capacitance is around 70pF.  Does that seem kind of high?

If that's the 7.7mH part, the data sheet shows a sharp "insertion loss" dip around 550kHz, which I was kind of thinking/hoping would correspond to the SRF (~11pF per coil).

Coils are kind of driving me crazy.  No one seems to make anything even remotely ideal for low frequency Theremin use.  Either the SRF is too low, the current handling too low, the inductance is too low, or the cost is too high (or all of the above).  Not to mention temperature dependence.

"If the inductance truly is 25mH, and f=1/(2pi*sqrt(LC)), then the SRF equivalent capacitance is around 70pF.  Does that seem kind of high?" - Dewster

Yes - Shockingly, surprisingly high.. And I am not sure its right.. But its what my ATE is telling me!

With the results I am getting (been working my way through a box of parts) I am surprised (and cannot really explain) how several of my previous experiments worked at all - But when I put the Bourns / JW Miller 6300 series parts under test, the results are spot on..

The results from the PLA10's dont make any sense really - up to 1.5mH per winding they have high SRF (often higher than the insertion loss dip) but above this value the SRF drops rapidly - well below the "dip"..

Its driving me crazy too - When I saw the results I thought something must be wrong with the test set-up, so dug up my EM test oscillator - in most cases this gave a lower SRF than my ATE.. But on both, the 6300 series parts were spot on - Tried 2.5, 5 and 10mH parts (dont have any 25's or 50's)

I also went over some coils I hand wound on ferrite rod, and these were fine (higher SRF than required) even though I am not a neat coil winder (these were wound by putting the rod into a drill !)

As for temperature dependence - in conventional theremins, the tank inductors are usually the same type, so frequency shift is likely to occur similarly on both oscillators and counteract the drift - I think the problems mainly occur if one is doing something "special" with the antenna inductances - as in, they are "disproportionately" large and changes to their value is not compensated by the same configuration on the reference oscillator... I suppose one could put "equalizing" inductors on the reference oscillator down to a fixed "antenna" capacitance to balance the drifts.. But there are other ways (particularly if one tunes the reference oscillator, and adds compensation at this point)

100mH is unusually high, and 70kHz unusually low! - I honestly thought the CMC method would be sure to work for your application - and Im not sure that it wont - I have used 7.5mH PLA10's at 333kHz (albeit as saturable reactors - but this should make matters worse, not better) - What I am seeing today doesnt make sense - my reactors shouldnt have worked!

I am a bit stumped! I will have a chat with the engineer at the company who made my reactor samples (have emailed him) - And perhaps with CoilCraft who have been helpful in the past. But I have been on this path with variable inductors before - Could not find anyone able to give me the data I was requesting .. Got so piss*d off that I started exploring electronically tunable inductors, and found a whole pre-tube world of saturable reactors and magnetic amplifiers, LOL ;-)

Fred.

ps - what are the currents you expect through the inductors?

I actually thought about putting the whole front-end into thermally isolated unit, and having a peltier device keeping this at a constant 25 degrees C - as in, heating or cooling as required.. OTT or what ? ;-)

"The results from the PLA10's dont make any sense really - up to 1.5mH per winding they have high SRF (often higher than the insertion loss dip) but above this value the SRF drops rapidly - well below the "dip".."  - FredM

The datasheets from other manufacturers who make CMCs, like this one from Epcos:

http://www.epcos.com/inf/30/db/ind_2008/b82733f.pdf

often graph the impedance vs frequency for the windings in parallel.  I wonder if that is the case also for the PLA10's?

"ps - what are the currents you expect through the inductors?"

For >100V swing at the antenna, I'm seeing 1 to 2mA at 70kHz (in simulation, and controlled via the series drive resistor).  The old AFE had roughly 10x current and frequency.  I like the lower operating current because it likely reduces heating and battery power is more of an option, and the operating frequency is way below RF, but the inductor specs scale such that it's probably something of a wash in terms of temperature stability.  Since my AFE doesn't use an LC reference oscillator (the reference is the FPGA demo board crystal oscillator, though not in a heterodyning sense) I'm kind of stuck with whatever temperature dependence exists in the tank & linearizing inductor.  Would love to find a pot core made out of a relatively flat tempco NiZn ferrite, like Fair-Rite 61 or similar, but no dice so far.

Kind of strange that there is little standardization among ferrite manufacturers, and that the core material is almost never mentioned in inductor datasheets.  The last thing this shaman needs is yet more black magic.

" I'm seeing 1 to 2mA at 70kHz (in simulation, and controlled via the series drive resistor)." - Dewster

For the kind of currents you are talking about, I think the JW Miller 6300 series will be your best bet - they behave superbly on every test I have ever thrown at them, and for 100mH you only need 2 - $4 total or thereabout..

"Kind of strange that there is little standardization among ferrite manufacturers, and that the core material is almost never mentioned in inductor datasheets.  The last thing this shaman needs is yet more black magic."

Yeah - I would say that by far the biggest "black magic" area I have come across is with regard to ferrites / coils / transformers - I dont think it was always like this - I think a couple of decades ago there were people who understood them and they were used well (I have been buying old RF transformers and stuff over the years, and the engineering which went into some of them is wonderful, obviously done by clever engineers who knew about ferrites etc) but these days they are mostly used in SMPS and the like.

Fred.

"For the kind of currents you are talking about, I think the JW Miller 6300 series will be your best bet - they behave superbly on every test I have ever thrown at them, and for 100mH you only need 2 - $4 total or thereabout.."  - FredM

I'm thinking that too.  50mH in the tank and 100mH (2x) in the equalizer branch.  PI wound and everything for $1.85 a pop at Mouser.  Thanks for helping me off this pot Fred!  Now on to tuner LEDs (I wish Cree had diffused lens offerings).

For what it's worth, JW Miller is a survivor from the golden age of radio and they really have a good grip on what a quality inductor is supposed to be.

Hi Rob,

My expierience with radio is minimal - Was into it in the late 60's (did my C+G and RAE but never bothered getting a licence - had moved to synthesisers by then..) And at that time parts were easy to get.

Where do radio enthusiasts get their parts from these days? Is it mostly salvaged parts or "new old stock" - or are there still manufacturers supplying this niche market?

For me, the most frustrating aspect of theremin development (and the reason why departing from the "old ways" becomes attractive) has been procuring parts - one can find small quantities reasonably easily, but finding a reliable source of some components (particularly coils and variable capacitors) for anything approacing production, is a nightmare.

The Xicor (Mouser) IFT's have been my main source of variable inductors / transformers - But I notice that some of these are being discontinued (last orders on available stock) - Fortunately the 42IF106 is not in this group.. some Toko variable inductors are dissapearing rapidly - And if the JW Miller 6300 series were to dissapear (they can be hard to get sometimes, with long lead-times) they would be nearly impossible to replace economically.

At times I wonder if buying an old RF coil winder might be a good investment.. But even these need the right ferrites to wind the coils on.

 I suppose there is always Charlie, and a return to beautiful big coils.. .

Fred.

"For what it's worth, JW Miller is a survivor from the golden age of radio and they really have a good grip on what a quality inductor is supposed to be."  - w0ttm

Their 6300 coils do have the best specs I've seen.  And of course after simulating the 25mH part all morning I see it's out of stock at mouser.  They do have the 50mH part though.

"At times I wonder if buying an old RF coil winder might be a good investment.. But even these need the right ferrites to wind the coils on."  - FredM

There are some fairly flat (around room temperature) ferrite materials out there with moderate permeabilitites, though I've yet to find any in pot cores or other easily wound form factors.  Perhaps I should pick one that is high permeability but temperature dependent, then compensate with tempco caps and/or air gaps and/or temperature sensors w/ software algorithms?  Blea.  If it weren't for the high voltages generated I might be looking into gyrators.

Too bad the permeability of free space is so low.  Nature abhors a low frequency Theremin.

"I actually thought about putting the whole front-end into thermally isolated unit, and having a peltier device keeping this at a constant 25 degrees C - as in, heating or cooling as required.. OTT or what ? ;-)"  - FredM

The Peltier is a bit OTT, but not the rest.  My AFE is in an aluminum box all by itself (though the next version will likely be in a plastic box) mainly to keep air currents from disturbing the frequency determining components.  Hams do the same with their home brew VFOs.

"The Peltier is a bit OTT, but not the rest" - Dewster

Actually, Peltiers are quite cheap, and have a huge advantage - With conventional "ovens" one needs to maintain the temperature above what is possible through circuit heating and ambient variations etc.. One has a heater and circuit in a thermally insulated containment, and the current drain is high (particularly if one does what Moog did with the Source, where the whole unit was effectively heated by the oven).

With Peltiers, one can set the temperature wherever one wants - Ideally this will be the normal circuit temperature under normal ambient conditions - and when conditions are "normal" minimum power is consumed.. Power is only consumed when the circuitry requires cooling or heating.

The electronics is in a thermally conductive enclosure covered with thermal insulation, except where the peltier comes into contact with the thermally conductive inner enclosure - a heatsink (which can become either hot or cold) is attached to the out-facing side of the peltier.

I designed one of these for a small sensor many years ago (the whole unit was the size of a big matchbox) and it was great - got months of battery life from it wheras the previous "oven" configuration discharged its battery in days.

But, with all that, I still want to avoid the complexity - Its bad enough getting the layout optimum for theremins without adding thermal enclosures, heatsinks etc.. And even with low cost peltiers, once one factors in the other stuff and difficulties with manufacture and calibration, its just not worth thinking about.. Ok for a small sensor on an oil-rig which one can charge £2500 for (of which £500 goes on certifications and another £500 on insurance, LOL ;-) but most thereminists probably wouldnt pay an extra £50 to avoid a couple of semitones drift over an hour..

Fred.