coil tuning

Fred & dewster I do not lack enthusiasm and have great admiration for the intelligence you both bring to the board. I have had my soldering iron fired the whole week while we chatted. I have a stand alone volume control circuit board design I have been adding finishing touches that solves many theremin issues and it works awesome. Any theremin line level can plug into it and have click free Muting, Pitch Preview, Total Shading Control, Reversible Volume direction or place it on your right side, a benefit to the left handed Thereminist.

Now for that other subject.

The data sheet for the TLC555 indicates it can oscillate up to 2 MHz. On an experimenter board I doubt it. My experiment for the next week is tune it to 900 kHz, attach the antenna in question and heterodyne with a standard LC oscillator. I will see if linearity can be found on the digital oscillator side.

If I get sound I will use fist popping throughout the pitch field to give a hearing sense of the linearity found, it may well not be found.

For me this should be straight forward, if I need extra power my independent 900 kHz tube/valve oscillators pc boards have extra kick.

There is a harmony between the pitch antenna and oscillator but I have observed that they don't like each other. It is kind of like middle school when the teacher forced you to dance with a girl!#$% ewh

Christopher

What would be the wave shape heterodyning a square wave with a sine wave?

"The data sheet for the TLC555 indicates it can oscillate up to 2 MHz."

Why not just use CMOS 4000B series gates and do it LC instead of RC?  Put a ~20k ohm resistor on the output of an inverter (NAND with the other input tied high, or NOR with the other input tied low) and use this to drive one side of an inductor.  Connect the other side of the inductor to the inverter input, and at this junction also connect a small capacitor to ground, and connect your antenna there too.

Things to try:

- A small resistor (~1Meg) across the inverter input and output. 

- A small capacitor from the 20k / inductor node to ground. 

- Connecting the antenna to the other side of the inductor.

- Using a linearizing coil in series with the antenna.

Regulate the supply.  This would give you square waves though, not sure what the implications are for that.  Maybe use a CMOS XOR gate for the mixer?  Or pick a sineish wave via a small cap off of one of the LC points.  You could make a fairly inexpensive analog Theremin this way, but I guess as usual the coils would be the most costly.  At the very least it gets you away from that single transistor stuff, and removes any DC bias current through the inductors.

A (mostly) digital could give you everything you might want pretty inexpensively, but you'd be working in a different medium.  When you get enough significant digits, digital signals can start behaving in very analog ways.  Once the signal steps are smooth enough, nobody can tell the difference.

Hi there,

My experiment is to eliminate LC as I believe the antenna trick has a relationship with LC. This gets back to the theme of this thread of using no coils, well sort of.

My approach may reveal my limited knowledge but I am sure something interesting will reveal itself. I am not looking for a good sound but to compare the response to what I know. If I found linearity then I will scrap my theory of why it works and consider it is strictly capacitive, this is something I need to know after all this discussion.

Single transistors are "sturdy" and can be made very drift stable. What do you find objectionable as they can have a nice voice with only six passive components to support it. I could post a transistor sound byte but that is not what the experiment is about.

Yes I have CMOS 4000B around here but now you introduce static vulnerability issues.

As recognized early on, our goals are different as every theremin designer should be.

PS: "If you wanted to land on earth", is an excellent answer but only relative to the speed you're traveling. The Milky Way spins onward!

"Yes I have CMOS 4000B around here but now you introduce static vulnerability issues."

The TLC555 has CMOS I/O - no?  I'm using a littelfuse SP721 ESD device to manage this in my AFE, but I haven't actually (knowingly) zapped it yet to see if it works.  I don't like the static & CMOS thing either, but the 4000B series is so darn flexible it hard to avoid.

"My experiment is to eliminate LC..."

RC oscillators are easy as well with a couple of 4000B inverters (as I'm sure you're aware).  And that construct make more sense to me than the 555 in this application.

When it comes to RC oscillators, I have not found anything which beats the TS555C (all other 555s are comparatively innacurate) for simplicity and reliability - The big advantage of the 555 chip is that its thresholds are preset at 1/3 and 2/3 Vcc, so one is gauranteed to get a waveform with consistant amplitude at the antenna - with CMOS gates, the thresholds are highly variable, and the waveform amplitude (the triangle wave at the antenna) is usually much lower than 1/3 Vcc.

The TS555C can clock happily at 2.7Mhz or below - it is CMOS, so not insensitive to ESD (well, actually, nothing is!) but one can easily reduce the risk by careful design..

This circuit shows a TS555C simulation – NE1 is not simulated, but will go into conduction (short circuit) when the voltage across it exceeds about 90V. NE1 has about 2pF capacitance.

R2 limits the current into the TH and TR pins of the TS555CN – Note, these inputs have built in diode protection anyway, and are connected together so are quite robust against normal ESD levels. The addition of R2 greatly decreases the chances of ESD destruction - the value of R2 is set quite high in this simulation - 10k is probably enough, limiting the current to 9mA for an ESD event - in fact, even 1k would probably be enough - the protection diodes in the device could handle a brief splat of 90mA.

C2 is an added capacitance to place the simulation in a “worst case scenario” state – The capacitance of the pins is included in the TS555C model – I have added an extra 5pF to account for sloppy layout. The effect of this capacitance is to lower the oscillator frequency, increase the antenna voltage, and slightly decrease the sensitivity..  

I also have this TS555C circuit http://www.thereminworld.com/Forums/T/28556/new-theremin-circuits which I really want to test ASAP – So if anyone feels like playing with this circuit and letting me know if it does improve linearization, I would be most pleased!

 Fred.

One other thing to mention about the TS555C (and for that matter, other 555's) is that, with the circuit I have shown (Where the Output drives the RC) is that the discharge pin (DC pin 7) is not used.

One can run the 555 with a VCC up to 16V, and pull this  (the discharge pin) up to 5V with a 1k resistor - this gives you a good clean logic level output to interface to digital circuitry, and allows one to have a higher antenna voltage swing - this gives better 'noise' immunity than if one runs the oscillator at 5V with the resulting lower antenna waveform amplitude.

Fred.

"What would be the wave shape heterodyning a square wave with a sine wave?" - RST

With true heterodyning (using a multiplier), the answer is that, for a theremin application in the audio range, the output waveform will be a sine wave.

I know you hate maths (I do too!) but it is quite simple:

A true 4 quadrant multiplier (663 chip, MC1496 etc) will give an output consisting of the sum and difference of the two input waveforms, and the sum and difference of all harmonics of these waveforms.

We can forget the 'sum' components, as these are always beyond hearing and filtered away..

So we are left with difference frequency and harmonic difference frequencies.

A square wave has odd harmonics, but actually, it wont matter what waveform one uses -

Assume your reference oscillator is 900kHz sine, and your VFO is 899kHz square..

the 2nd harmonic of your VFO is 1.798Mhz, if it was present.. But your reference oscillator has no harmonics being a sine.. So the harmonic (difference) between the reference and VFO is 1.798Mhz - 0 = 1.798MHZ ! - This difference will be lost in the 'sum' filter.

As can be seen, no harmonic will appear on the difference output unless it appears in BOTH the reference and variable oscillator waveforms!

- BUT - The above only applies to a true multiplier.. Other simple mixers (like that in the EW) distort the incoming waveforms, and the harmonics generated by these distortions make analysis of the actual output waveshape / spectrum much more difficult to compute.

There is something quite important for you in the above.. You ask about the rationale for running theremin oscillators at lower frequencies.. One rationale is that the higher harmonics extend well into the Mhz - and these harmonics will get attenuated as a concequence of circuit layout and component limitations.

The 8th harmonic of a 285kHz oscillator is 2.28 Mhz, and this is getting into the territory where serious attenuation can occur - For your 900kHz oscillators, the 8th harmonic sits at 7.2 Mhz.

This is a serious rationale for running theremins at lower frequency, despite some obvious advantages to running the oscillators higher - this is particularly true for analogue theremins where the waveshape from the oscillators forms part of the final audio waveshape.

If you want to add harmonics to your square / sine mix, simply clip the sine wave (make the clipping adjustable!) - you will then add odd harmonics (you cannot get even harmonics, because the square wave does not contain them).

But for comprehensive waveshaping, you need to use mixed-signal techniques, like those shown in my posting (same link as the 555 oscillator), as these techniques do not depend on oscillator harmonics and are just as useful to 900kHz theremins as to 250kHz theremins.

Fred.

Fred that is a fascinating oscillator schematic of the TS555C. I never would have figured out using the output pin 3 in the configuration. I was just going to tap a frequency counter connection there.

In that EM antenna thread you said:

"Don't change the coils parameters - doing this will severely affect linearity - the oscillator is designed to work with the loading provided by the tuned antenna circuit - unless you really know what you are doing, and have the test equipment required, leave the antenna circuit as specified!"

In my Lev Antenna explanation I have always said it was the hand loading effect de-tuning the antenna configuration connected to the pitch oscillator that un-cramps the musical notes at the antenna. As I believe they counter balance one another. I refer to de-tuning the antenna as changing the Xc of the tank circuit. "This is why hand capacitance does not work alone. (static capacitance + Xc)"   <= The level of my math ability

Also more current in the antenna give more linear octaves but linearity remains at even low currents, just two or three octaves. In my pitch field all octaves are linear whether two or five show up. There are no non-linear octaves on the outside parameters. There can be a rapid tightening at either edge.

Are you slowly moving in a new direction of antenna theory thinking?

As far as audio harmonics and wave shape my preference is a soft ramp or saw tooth any day.

Christopher

Edit: The 555 experiment seems less important now! We refer so loosely to the word linearity when in our use it is relative or a ratio to the man made musical scale, that is what's mystical about all this, not how it comes about!

"Are you slowly moving in a new direction of antenna theory thinking?"

Christopher, unfortunately, No!, I think it is about the words we use, and the fact that we understand different things by them. I am starting to think though that perhaps you are a stuborn genius who enjoys teasing people.. Probably reading the formulas and thinking "what a load of plonk*rs" while winding us up by pretending they are beyond you! LOL! woudnt that be funny! ;-)

I also think that the time I have spent away from my lab, with my laptop while keeping my legs up, without any pressure to produce anything because I know I am not able to, has been useful - I have been able to 'waste' time reviewing theory and running simulations, brushing up my maths a bit, and generally preparing myself for the next 'phase' - whatever that is.

The complexity of how the oscillator and equalizing circuits operated together meant that, whilst I had understood their operation, this understanding was purely 'visual' - I was not able in any way to apply maths to it - I now have a much better understanding of the loading effects of the series LC antenna circuit on the parrallel LC oscillator circuit - but it is still all capacitive / inductive - it can all be simulated using capacitor, resistor and inductor models - and the "antenna" in these models is the junction between the (equalizing) inductor and the capacitance to ground "seen" by the antenna "plate".

My "Don't change the coils parameters - doing this will severely affect linearity - the oscillator is designed to work with the loading provided by the tuned antenna circuit "

Refers to the components (47uH + ~47uH + 3.3nF) in the EW oscillator circuit - Any change in the resonance point of the antenna will require a proportional change (ratiometric) in these components if the relationship (effect of the loading by the antenna circuit) is not going to change the linearity and/or sensitivity.

"Fred that is a fascinating oscillator schematic of the TS555C"

Yes, its amazingly simple - and the only way to get 50/50 MS from a 555 that I know of - But its not without its down side - The Q output on some 555s is not equal impedence for sourcing and sinking, and on some the voltage from this output does not go rail-rail - and worst - the high and low output levels can be temperature dependent.

This circuit can only be used with any accuracy if the impedence being driven is high - orders of magnitude higher than the output impedence of the 555. Results of all 555s other than the TS555 in this mode are appalling - only usable where stability and accuracy are not an issue.

The standard 555 circuits maintain accuracy by using the supply voltage to directly charge the capacitor, and as the reference voltages (threshold and trigger) are derived from the supply voltage, accuracy is reasonable.

I think it is the combination of low current and good output drive capability which makes the TS555 suitable for operation in this mode - good enough for simple theremins. (the drive problems apply equally to any and every circuit which directly charges / discharges the timing capacitor from output transistors - and this includes CMOS gates.. If I use CMOS gates in an application like this, I stack as many as possible outputs in parrallell to boost the drive capability)

Anyway - it will be real interesting to see where this thread leads us all ;-) Some discussions have a surprising effect on the future - I think this may be one of them.

Fred.

"As far as audio harmonics and wave shape my preference is a soft ramp or saw tooth any day."

Then use this circuit - you get a lovely ramp, as well as triangle (which turns into nice sine with LPF) and square

I have deleted my original reply and re-drawn the schematic I posted there so as to make things easier - my cut/paste from one to the other got messed up - so if things here dont make sense, too bad! Too tired to read this now..

Fred.

Two RC oscillators for VFO and REF, and you have a pitch-only theremin with great sound (but probably completely unplayable because the linearity is as crap as every other theremin using RC oscillators) - But at least it will have a great sound! LOL ;-)

And who knows - If my compensated 555 circuit actually works, I may just have given the world the uber-simple linear pitcho theremin which finally vanquishes the Glasgow crapomin and its evil brethren...

ROFLMAO! 

Ok, made it easier for those wanting a simplified version of my waveshaper..

R3 = 10k, C2 = 4n7.. You could add a 50k variable resistor in series with R2 to give adjustable ramp harmonic content (variable low pass filter)

R1 = 10k C1 = 4n7 .. Same with R1.. You could add a 50k variable resistor in series to give adjustable triangle harmonic content (variable low pass filter)

XOR gates = 4070 or 4030, D latch = 4013

Requires reference and variable oscillator inputs - these can happily be any normal theremin frequencies from 100kHz up to 1Mhz, ideally square wave, vith amplitude going from ground to supply rail ideally, but certainly from at least <0.25*supply to >0.75*supply.

 The FULL set of schematics can be found in this thread at Element-14 and there is more discussion in this New (?) Theremin Circuits thread here at TW.

Note - The ability to control the amplitude of each waveform and mix them under voltage control (which is shown in these other threads) is well worth the few extra components! - Using the waveforms here and trying to mix them with a conventional mixer will probably work out more expensive and inferior to the CV method presented in these other threads.

If you only want one waveform (or to switch between these waveforms, selecting one at a time) then the schematic here is the simplest and cheapest.