# What type of oscillator is this? (Theremin Project. Variable Pitch / LC Oscillator)

Posted: 4/15/2017 9:07:41 AM

Joined: 8/25/2016

I found this oscillator at a Theremin project pdf. I simulated the circuit, and found that the output (left wire) oscillates at 236.5 kHz.

This circuit is referred to as "Variable Pitch Oscillator", but that doesn't seem to be the best reference for it.

So I'm wondering: what circuit is this, and how do I calculate the oscillation frequency??

Posted: 4/15/2017 9:56:47 AM

From: Colmar, France

Joined: 12/31/2007

This push-pull oscillator circuit is used (in different variants) in all Moog Etherwave (Standard, Plus, and Pro) theremins. The variable inductor has a tuning range from 72 to 103uH. This allows to set the oscillator frequency around 270 to 290kHz, the frequency range used in the Moog theremins. In most designs, the antenna circuit with a huge linearization inductance is connected on the left wire, which makes (if everything is correctly fine-tuned) the oscillator frequency go up by 6 to 8kHz, before it goes down again with increasing hand capacitance.

The left wire is a bad point to pick the oscillator's output, since capacitive load will de-tune the oscillator. A better point to pick the signal is the base of Q2 (as done in the Etherwave Pro).

Basically, the oscillation frequency is calculated from C1 and L5, in reality, the junction and Miller capacitances of the transistors add to C1 and the effective inductance of L5 is reduced by the DC component flowing through it. The result is difficult to predict, that's why Moog made the inductor variable (but in the E-Pro, the inductance is fixed and there is a small variable capacitor in parallel to C5 for fine-tuning) to compensate these effects and to tune it for the desired frequency.

Posted: 4/15/2017 10:04:23 AM

From: Colmar, France

Joined: 12/31/2007

In the theremin domain, where 1/100 of a pF might play an important role, and where "optimally" designed circuits under engineering aspects do not always give the feeling desired by the musician, there is only one way: Build the circuit, play it (only if you are at the same time an excellent thereminist who has much experience with many different theremin models, so that you may judge it from the musical side, too) or let it test by a professional player (you might need to travel a few hundreds of miles to meet one), optimize it, let it check again, and so on, until everybody who tries it starts smiling.

That's how I developed the famous ESPE01 add-on module.

Posted: 4/18/2017 12:38:25 AM

Joined: 3/9/2017

I disagree about not breadboarding. Breadboarding allows a creative individual to explore possibilities without committing to a hard design. If one understands the limitations of breadboarding, then the final design does not need to greatly depart from the breadboarded version. An emitter coupled oscillator is still an emitter coupled oscillator, whether running on a breadboard or on a chassis or phenolic circuit board.

The real question is, 'Is it the best choice?" Just because Moog used it does not eliminate other possoble solutions. For example,  Termin did not use cathode coupled oscillators, and his Theremin sound is the 'gold standard'.

The oscillator shown is the solid state version of the Butler cathode coupled osclillator. Unfortunately, it suffers from frequency instability due to heating effects of DC current through the inductor. This can easily be improved by using a shunt-fed tank circuit which precents DC current from flowing through the tank resonator circuit.  The Butler oscillator is good if you need a low impedance output to drive audio cables of 50 metres or more. Otherwise, one really shouldn't worry about impedance. Is the drive voltage at the next stage appropriate and the loading light enough not to cause voltage sagging?  Those are the only questions.

Mark

Posted: 4/18/2017 2:39:00 AM

From: Northern NJ, USA

Joined: 2/17/2012

Theirry, I also strongly disagree.  Breadboarding and simulation provide extremely useful insight, which is why they are basic tools of any competent EE.   One obviously needs to be aware of the few pF or so between breadboard rows, and that simulation often doesn't perfectly mirror reality (duh).

You're in witch doctor territory if you're advocating against their use.  As a mathematician would you seriously listen to anyone telling you to abandon algebra?

Posted: 4/18/2017 6:05:09 AM

From: Colmar, France

Joined: 12/31/2007

I wouldn't abandon algebra, but I finally abandoned the slide rule and am working with a HP-35s now... :-)

And I agree that breadboarding and simulation might be helpful tools - for the experienced EE who understands the impact and can estimate the discrepancies between that and reality.

But judging a circuit just from simulation as the thread creator did, without having experienced/played/fixed hundreds of instruments using that circuit, seemed "dangerous" to me. That's why I dared to write my "warning".

Posted: 4/18/2017 2:04:34 PM

From: 60 Miles North of San Diego, CA

Joined: 10/1/2014

#### Christopher

Posted: 4/18/2017 2:54:43 PM

From: Northern NJ, USA

Joined: 2/17/2012

All I'm saying is everyone starts someplace, and there's no shame nor harm in that, nor in exploration of any type.  Those with more experience can and should help them along, but there's a danger to both sides when that help is too restrictive and quashes inquiry outright.  Different people come at these things from different angles, and I often find this to be an unexpected source of inspiration that I wouldn't get otherwise.  And through explanation, the teacher is given the opportunity to really learn the fundamentals of the subject.

Posted: 4/18/2017 5:03:21 PM

Joined: 3/9/2017

As a teacher, I found I had to succeed at explaining a concept to an understanding audience before I really understood the concept myself.

Thierry said:  "Basically, the oscillation frequency is calculated from C1 and L5, in reality, the junction and Miller capacitances of the transistors add to C1 and the effective inductance of L5 is reduced by the DC component flowing through it. The result is difficult to predict, that's why Moog made the inductor variable (but in the E-Pro, the inductance is fixed and there is a small variable capacitor in parallel to C5 for fine-tuning) to compensate these effects and to tune it for the desired frequency."

My experience with breadboarding has shown a remarkable ability of formulae to precisely predict component values and resultant frequencies.  I have a known value inductor and a known value capacitor that I use to measure the unknown component. When a tank circuit is properly isolated from DC, the typical theremin oscillator is relatively insensitive to Miller capacitance, and stray capacitance and inductance. For example, 10pF added to 3900pF coupled with a 100uH inductor causes an error of only 1/10 of 1 percent (320Hz). Properly designing the impedance loading of the suceeding circuit can reduce Miller capacitance by several orders of magnitude.

The real challenge, and the genius of Termin, is found in the antennae coils.