Moog Etherwave - Problem with Pitch Oscillator Variable Inductors

Hey all, I'm a student building a Moog Etherwave as an electronics project. I'm using the schematic provided at http://www.theremin.info/-/viewpub/tid/10/pid/58 with L14 being 22 uH instead of 33 uH as pictured (I believe this is correct.) For component specifications, I used the list iin the EM article provided at http://www.emusician.com/emusician/web_clips/Theramin%20Feb96_scrn.pdf.

So far, I have only constructed the power source and the pitch oscillators. I decided to test the oscillators with an oscilloscope and discovered something disturbing.

When I take the variable inductor (L5 or L6) out of the circuit (leaving only the fixed 47 uH (L12 or L13), the frequency of the oscillator is as expected (about 400 kHz), but when I put the variable inductor back in, NOTHING happens. It should be about 286 kHz, but it is virtually unchanged from 400 kHz. I tried tweaking the screw, but didn't really notice any change. If the frequency did change at all, it wasn't more than 1%.

I tried two fixed 47 uH inductors in series, and the output frequency was as expected (286 kHz). So it seems that the problem lies in the variable inductors.

I got most of the parts specified in the article, but instead of the extinct TOKO coils, I got some from Coilcraft. For L5, L6, and L11, I used the Coilcraft "Slot-7" 7M2-473 (Data Sheet: http://www.coilcraft.com/pdf_viewer/showpdf.cfm?f=pdf_store:slot7.pdf).

What could be the problem?

Are these coils acceptable to use in this design? I didn't notice anyone recommend them in any thread. I did, however, notice people recommending the "Slot-10" coils. Is there a difference here?

Could the coils be broken? I tried both, and they behave the same, so I doubt it.

Are the inductance values correct? They do say 7M2-473 on the side... so they should be 47 uH, but they're WAY smaller than the fixed ones. I don't know much about manufacturing inductors, but this is a little startling.

If anyone could help out a newbie in need, I would greatly appreciate it. Thanks!

The smaller 7mm variable inductors have a lower Q-factor (thus higher DC-resistance) and a higher parasitic capacitance than the recommended 10mm variants TOKO EZ 1017 or COILCRAFT SLOT TEN-5-10 and are thus not usable in the differential pair oscillators of the Etherwave and EM design.

There are reasons (among them painful experience) why I don't stop here to recommend ONLY the two types cited above.

Just believe in the words of an old theremin warrior...

Thierry, could you elaborate a bit?  Other than adjustment range and Q the coils don't seem all that different though they definitely seem to be behaving differently in real life.  I'd like to understand what's going on here because I'm about to make a coil order myself.

I don't see DCR or self-resonant frequency for the SLOT TEN-5-10 on the Coilcraft datasheet, do you know what they are?

Also, I'm looking at this paper:

http://www.intusoft.com/articles/inductor.pdf

where the author says Q isn't determined by DCR but by core losses, which are modeled in that paper by a parallel resistor (whereas DCR is modeled as a resistor in series with the inductor).  The EWS uses the coils to bias the drive transistor, could DC saturation be the reason why the replacement coil isn't working as expected?

Afaik Q is defined as Zimg/Zreal, so the DCR does play a role. 

The 7mm type has a Q of 20 at 2.5MHz, the 10mm type has a Q of 32 at the same frequency.

Zimg=2 x Pi x 2.5MHz x 47uH = 738 Ohm

Thus I find a Zreal of 738/20 = 39.6 Ohm for the 7mm types and 738/32 = 23.06 Ohm for the 10mm types.

I measured a smaller capacitance (30% less) between the windings and the shielding for the 10mm than for the 7mm which made me deduce a higher SRF. That was the theory part.

The point is (and proven by painful experience as I said before) that the 10mm types work and the 7mm types definitely don't. That's practice! :-)

Thanks for the quick response, Thierry! So if I'm planning to use the Slot Ten in series with the fixed inductors, should I go for the Slot Ten-5-10 with Lmin = 25 uH and Lmax = 60 uH? That seems like it would be hard to adjust precisely. Will this be a problem?

As you could already see, the oscillators are very stable and will work in a broad frequency spectrum. The absolute frequency plays only a role when it comes to optimizing tone spacing and pitch range, there is also a point of resonance with the antenna coils somewhat lower which makes tuning somewhat more difficult.

So adjust both pitch oscillators to about 292kHz first and make the whole circuit working. If all is fine, you should install the circuit in a cabinet with a fixed antenna holder, so that the pitch antenna will "see" a more or less constant capacitive environment. This will allow you to decrease the operating frequency slowly step by step until you reach the magic 286kHz.

Thierry, thank you, I do understand your points, and I do respect your experience, what is the root cause as to why these different coils don't work?  If the circuit was not oscillating at all I'd maybe think the Q was lowering the loop gain below 1 at resonance, but that doesn't seem to be the case.

The different coil is acting like a wire, and adjusting it produces little or no change in frequency, so it is:

1. Shorted,

2. The core is saturated due to the DC bias through it,

3. It is a much smaller value than we all believe (mismarked or something),

4. The adjustment range isn't sufficient for it to correctly drive the linearizing coil & antenna capacitance.

I'm thinking 3 or 4 are the most likely, but I'm kind of new to inductors.

[EDIT] Just saw your latest post - so it's 4 then?

Hmmmm,

#2 is theoretically possible, the cores of the 7mm coils are much thinner than those of the 10mm variant, but I'm not sure if 150mA peak current (but almost no DC bias) will drive the smaller cores already into saturation.

#1 is also theoretically possible, it would not be the first time that I see someone using the two shield pins as active connections...

I must admit that I never invested much time in finding out why the smaller coils didn't work. When confronted with fixing someone's DIY efforts, I replaced the coils by the working types which are known and everybody was happy.

Hi!

I've been using a coil craft "slot ten" 5-11, 10mm variable inductor of 68uH as per the instructions of 

http://www.cs.nmsu.edu/~rth/EMTheremin.pdf

I am wondering if using the series inductance of a "slot-ten" 5-10 inductor in series with a fixed inductor of 47 uH eventually worked out for you laughing.walrus? I would appreciated and answer as soon as possible!

Thanks,

Elsa

I mean the fixed inductor will be of value 22uH