# Colpitts oscillator for theremin project

#### Monow

Joined Jan 19, 2016
10
Hello there,
I'm gonna try to build a simple Theremin, pitch only with only two oscillators and amp. Well I've got some schematic and information about LC oscillators at school, the problem is that only theoreticaly, I don't have a clue about practical stuff, what values should I use, which oscillator is better or just anything.

So my question is which design would be the best (good sine wave, stable frequency) and if you could help me also with values.
I definitelly need frequency about 400kHz, voltage supply about 15V, capacitors in the tank circuit in pF and it would be good to have one 'end' of the tank circuit connected to ground so I could change it's capacity by my body (hand) between a antenna and the ground.

Actually frequency is only thing I can get from C and L values and as I've tried to simulate one schematic which I found on the internet, with frequency which suits my requirements, but it had nF capacitances so small change by hand. So I've tried to change the values to pF and at the same time used 1000x bigger coil, but still it stopped oscillation. Sorry if I something wrote wrong, but I have just some basics from school.

I'd really appreciate someone's help, thanks a lot
Monow.

#### GopherT

Joined Nov 23, 2012
8,009
Hello there,
I'm gonna try to build a simple Theremin, pitch only with only two oscillators and amp. Well I've got some schematic and information about LC oscillators at school, the problem is that only theoreticaly, I don't have a clue about practical stuff, what values should I use, which oscillator is better or just anything.

So my question is which design would be the best (good sine wave, stable frequency) and if you could help me also with values.
I definitelly need frequency about 400kHz, voltage supply about 15V, capacitors in the tank circuit in pF and it would be good to have one 'end' of the tank circuit connected to ground so I could change it's capacity by my body (hand) between a antenna and the ground.

Actually frequency is only thing I can get from C and L values and as I've tried to simulate one schematic which I found on the internet, with frequency which suits my requirements, but it had nF capacitances so small change by hand. So I've tried to change the values to pF and at the same time used 1000x bigger coil, but still it stopped oscillation. Sorry if I something wrote wrong, but I have just some basics from school.

I'd really appreciate someone's help, thanks a lot
Monow.
Please post the schematics you have been using (working or not).

#### Monow

Joined Jan 19, 2016
10

This is it with values I found on the internet and the schematics correspond to the one I've learned at school (except the L2 coil, we used a resistor why there is a coil then?) so I get it a little. Wave seems fine to me, but when I've changed values to adjust a higher freq. 300-400 kHz and as I said with pF capacitors it didn't really work since then.

PS. actually it works but with a high distortion, doesn't even seem like a nice sinus wave and the freq. jumps in kHZ when values changed

#### Monow

Joined Jan 19, 2016
10
Any ideas how to fix it, make it work with nice sine wave with changed values in the tank circuit, please?

#### GopherT

Joined Nov 23, 2012
8,009
@Monow
What you are trying to do (change some things but not completely redesign everything), you will always run into the kinds of problems you are having. When you
- change some capacitors or tank and then
- change resistor values to bring a single frequency back to 'normal',

you run into the same problem as if you
- changing the slope of a line,
- grab the line (like you would in a drawing program) and slide it to the left to bring a single point of the new line to cross the existing line at a desired point

Nothing about the new line will be similar just like nothing about the Theremin will be the same. As you said, it may jump in k-Hz values instead of 10s to 100s of Hz. Therefore, the usable range becomes very small and suddenly gets into either sub-sonic or ultra-sonic frequencies.

#### Monow

Joined Jan 19, 2016
10
Thanks for nice mataphor.

So to get where I want I need to change the whole circuit, values right? But there's my problem I don't know how to to those. As I said we've only learned how it looks like and that's all. So I thought I could grab this one and only change tank values to get what I wanted, which is apparently totaly wrong.

Do you think you could help me a little bit with design or at least direct me some example oscillator with my requirements?
I'm sorry but I really don't know how to design it right

#### RichardO

Joined May 4, 2013
2,270
The way I remember a Theremin working is this:
A high frequency oscillator is changed in frequency by the proximity of the player's hands. This frequency change is small so the high frequency is mixed with a stable oscillator to create a difference frequency in the audio range.

More details are here:
https://en.wikipedia.org/wiki/Theremin

#### Monow

Joined Jan 19, 2016
10
That's what I know, how the whole theremin works, how it looks like on paper. But what I need as I said is to design a working Colpitts oscillator to start those frequencies, once I have this I can just do everything else by myself (the build)

What I ask for, once again, is just a working LC oscillator with frequency about 400-500kHz :/

#### GopherT

Joined Nov 23, 2012
8,009
That's what I know, how the whole theremin works, how it looks like on paper. But what I need as I said is to design a working Colpitts oscillator to start those frequencies, once I have this I can just do everything else by myself (the build)

What I ask for, once again, is just a working LC oscillator with frequency about 400-500kHz :/
That is half the problem. You cannot just plug a number into a colpitts calculator and expect it to work in a real circuit. Simulators are happy to spit out frequencies for a given set of values but whether or not they really work in the real world are another thing. An inductor usually has zero resistance and a capacitor has no resistance in a simple simulator of ideal components. In reality, there is DC resistance in a coil and series resistance in a capacitor. The gain and speed of a transistor is limited. So, you end up having only a certain range of capacitors work for a certain coil inductance and, therefore, only a certain range of frequencies can be achieved.

At the same time, you need to know (usually by experimentation or looking up published values) the capacitance of a human hand near the antenna. With that value, you need to design something that will cause the frequency shift to be in the appropriate range based on your fixed oscillator. The higher the frequency of the fixed oscillator, the less % need for a shift. At 400k Hz fixed frequency, then a shift of only 0.015% will give you a low note (~60 Hz) and 1.25% will give you a high-high note (10k Hz). Keeping that narrow range is critical for achieving audible tones. if your base frequency is much lower (e.g. 60k Hz, then the notes will need a much bigger capacitance shift.

In either case, you need to go back and do the math for what base capacitor sizes you will need for 400kHz and for what size capacitor will allow the presence/absence of a human hand to shift the capacitance by 0.015% to 1.25% (assuming the 60 to 10k hz range is your desire. If not, you have more math to do).

#### Monow

Joined Jan 19, 2016
10

Ok, so I've changed some values that I thought they could influence my output. Probably it worked, frequency is stable and the signal doesn't have any distortion ( I know it's only a simulation, but I want to know what could be the best before I try to build it for real ).

Now two questions, is there anything to do in circuit to smooth a little bit the sinewave?
and why the singal on the "top of tank circuit" has faster rise time than fall time (while it should be just the same signal, just inverted because of capacitors divider?)

Thanks again

#### GopherT

Joined Nov 23, 2012
8,009
So, our circuit seems to work with C2 + C1 in series making about 125 pF. However, there is going to be some distortion when these two are the same value. You can often remove that distortion from a this type of oscillator by making C1 about 1/3 the value of C2 while still maintaining about 125 pF of capacitance. Look up how capacitors in Series are calculated to get you the answer. So, the easiest first attempt is to make C2 about 3x the value and see what happens.

#### Monow

Joined Jan 19, 2016
10
I've changed as you said C2 capacity to have it 3x C1, so now I have there 600pF and 200pF, which makes 150pF in series, still fine for me.
The fall time looks better and it also have better frequency stability (the range was like 10-20Hz before and now it's around 3Hz so I'm so happy with that ), still the amplitude seems distorted at this zoom ( 20V/div it looks decent of course - it's 5V/div in the picture ). So I'll try a little bit to adjust that and I hope something helps.

So far could I actually use this design, i mean if there's something that would cause it would work at all ( I know there will be maybe different freqency in real because of the connections etc.)

#### Sonoran Desert Tortoise

Joined Oct 30, 2014
79
@Monow,

You are getting some type of pixilation because your simulation frequency is too slow vs. your signal frequency. Or you are zoomed in instead of making other sweep range or what ever. I don't know how your simulator workis but it is a question of simulator, not circuit distortion.

#### Monow

Joined Jan 19, 2016
10
Oh, I thought it could be this, so just the virtual oscilloscope isn't fast enough here, right? THen it should be fine in real, if it works.

btw. is there any way how to remove R5 in feedback? Because the signal is too high without it so I just added a resistor to lower it down for the base of the transistor, but I suppose it's not really the best solution, is it?

#### Sonoran Desert Tortoise

Joined Oct 30, 2014
79
Y
Oh, I thought it could be this, so just the virtual oscilloscope isn't fast enough here, right? THen it should be fine in real, if it works.

btw. is there any way how to remove R5 in feedback? Because the signal is too high without it so I just added a resistor to lower it down for the base of the transistor, but I suppose it's not really the best solution, is it?
You should be able to adjust the virtual scope - otherwise simulation would be useless if it is limited to signals of 400 Hz and less. Go ahead, give it a try.

#### KL7AJ

Joined Nov 4, 2008
2,229
Hello there,
I'm gonna try to build a simple Theremin, pitch only with only two oscillators and amp. Well I've got some schematic and information about LC oscillators at school, the problem is that only theoreticaly, I don't have a clue about practical stuff, what values should I use, which oscillator is better or just anything.

So my question is which design would be the best (good sine wave, stable frequency) and if you could help me also with values.
I definitelly need frequency about 400kHz, voltage supply about 15V, capacitors in the tank circuit in pF and it would be good to have one 'end' of the tank circuit connected to ground so I could change it's capacity by my body (hand) between a antenna and the ground.

Actually frequency is only thing I can get from C and L values and as I've tried to simulate one schematic which I found on the internet, with frequency which suits my requirements, but it had nF capacitances so small change by hand. So I've tried to change the values to pF and at the same time used 1000x bigger coil, but still it stopped oscillation. Sorry if I something wrote wrong, but I have just some basics from school.

I'd really appreciate someone's help, thanks a lot
Monow.
In a Theremin, both oscillators are actually radio frequency oscillators (or at least super-sonic) and it's really NOT very important that they are sinusoidal! This is because any harmonics fall OUTSIDE the hearing range...by a long shot. Let me look through some ANCIENT archives and see what I can find.
Eric

#### Monow

Joined Jan 19, 2016
10
Oh that's good, I just chose colpitts because it's one of those we learned. So it would be fine to use colpitts with distorted sinus, if I get i right?
I'll be glad to see any inspiration if you find something

#### Alec_t

Joined Sep 17, 2013
14,370
it's really NOT very important that they are sinusoidal!
Probably not, but the input waveform does determine the output waveform if the two oscillators are mixed in a ring modulator. Input sines will give an output sine, whereas input square waves will give a triangular output waveform.

#### Monow

Joined Jan 19, 2016
10
So I finally managed to get all the components a try it running. I completly threw away the feedback resistor, it worked really bad with it and what's better worked without it.
Well firstly I measured the 600pF signal which looked like this: (I don't know why)

so next was the 200pF signal - way better, but the top of the sinus doesn't seem right:
and also the base signal -

Do you, please, see what the problem is and how to get it work better?

I was told to fix the top sinus curve by reducing base resistor R3, but there shock for me came. Even though I put it away the circuit worked pretty much the same, how's that possible ...
thank you all