Simple Theramin

Thread Starter

FredM

Joined Dec 27, 2005
124
Hi folks - I have done a lot of posting here, but not really contributed much.. Hope this makes up for it a bit!

This is an incredibly simple theramin which has stunned me with its performance.. I got the idea about 36 hours ago, while looking at the TS555C data sheet, and noticing it could oscillate at 2.7Mhz, and could have extremely small timing capacitors.. I wondered if it could be used as a reliable capacitive sensor, which led to the theramin idea, which led to me staying up all night designing and building it, and all day playing with it!

The concept is simple and based on the original theramin.. there are 2 555's operating in astable mode (using feedback from the output to charge / discharge the timing capacitor - this gives equal M/S ratio).. on one 555, the timing capacitance is the 'antenna' and controlled by proximity of earthed entity... I used a 6" strip of ribon cable, and had another strip next to it (but not too close) connected to -8V.

The other 555 is set up to oscillate at a constant frequency, which can be tuned. With no 'entity' in the detection range, the 'constant' 555 output frequency should be adjusted to be the same as the frequency from the detection 555 (this process is easier said than done!)

Outputs from the 555's are taken to crude low pass filters and attenuators, and drive the X and Y inputs of an AD633J analogue multiplier (the only slightly expensive part in this design).. This produces a signal which is the sum of the two incoming frequencies, modulated by the difference of the two incoming frequencies.

Incoming frequencies should differ by (at maximum) about 15 kHz.. So the output from the multiplier is fed to a crude LPF which rolls off at about 10kHz.. This gets rid of the 'sum' component in the waveform (this component is arround 350 to 450 kHz)..

Take the output (with care) to an audio amplifier, and have fun! When set up correctly, the audio output varies from 0 (no sound) at a distance (from antenna) of about 1 meter, to about 15kHz when touching the antenna.. But these figures will vary hugely depending on construction and environment.

Down side to all this? .. Only one.. I always believed that all I needed to do was get an instrument like this, and it would be a doddle to play .. On this, and this alone, I was Deeply dissapointed!
 

Salgat

Joined Dec 23, 2006
218
If I had a ad633j I would build that thing, such a simple circuit. Can I just sum the two signals with op amps in order to replace that IC?
 

Thread Starter

FredM

Joined Dec 27, 2005
124
If I had a ad633j I would build that thing, such a simple circuit. Can I just sum the two signals with op amps in order to replace that IC?
Hi Salgat,

No. Summing wont work.. it is analogue multiplication which is required... However, I have been trying to get rid of the AD633J myself, and have an idea which I have not tested or simulated yet..

The 555C's have a CV input which can be used to modulate their frequency.. If one takes the output from the sensor 555 (via appropriate attenuation and biasing - probably ac coupled, perhaps with some waveshaping .. ie simple filter) and fed this into the CV of the reference 555, and set the reference 555 so that it was running at the maximum frequency available from the sensor 555..

The output of the reference 555 circuit should, I believe, be a modulated waveform containing sum and difference frequencies (same, or similar, to the output one gets from AD633)..

I have been distracted by real work :(, but will get back to this soon.. I have scrawled up some plans for an even simpler theremin (based on what I describe above) which has 2 'antennas' - one for pitch, and one for amplitude, and the whole circuit uses one 40106 IC, one dual opamp, a transistor, and an H11F1 opto-coupler (for gain control) + passives.

The circuit I posted sounds wonderful though - one of those times when one bashes something crude together, and it works better than something one carefully crafts! - I think it is the complexity of the waveform harmonics which does it.. The HF rejection of my 'filters' is poor, so there is foldback of someof the inaudible content into the audio output.. This is usually awful - but somehow (probably due to the levels being low) it adds a 'growl' at the bass end which is real punchy and pleasant.

I think the audio is going to be severely affected by what one plays this through, however.. Nothing with class D or containing any SMPS .. There are a LOT of nasty HF (ultrasonic) signals still coming from the output!
 

Thread Starter

FredM

Joined Dec 27, 2005
124
Nope.. Not as simple as using CV im afraid.. There may be a way to do it, but all one gets is frequency modulation which is not what the AD633 produces.. demodulating the signal from the reference 555 simply returns the sensor 555 frequency, (with applied waveshaping) which is not what one wants..
there must be a cheap way to do it (AD633J costs me about £8 in one off - I dont like paying more than £1!)
I think that four quadrant multiplication is required - I am looking at using a transconductance amplifier (LM13600 or equivalent).. The specification for the AD633 is probably better than required for this application.
 

Thread Starter

FredM

Joined Dec 27, 2005
124
Update - An LM13700 can be used to replace the AD633 .. A few extra passives required, but configured as four quadrant multiplier as per data sheet, it should work.. This part is <£1.
The LM13700 has 2 OTA's, and the second can be used to implement a good filter and/or VCA controlled from another 555 acting on a second 'gain' antenna.. this is simple to implement as all one need to do is convert its frequency output to a voltage (drive a monostable which produces a pulse every change of state.. for example, if the maximum frequency from the 555 is 500khz, the monostable time should be set to 1us.. Then integrate these pulses through an R*C with a TC of about 5ms.. this will give a smooth voltage output proportional to frequency.. In fact, this method could be used for frequency control as well - if you dont like the multiplier principle of the theremin.. Simply drive a VCO from this voltage!)

I will post more schematics soon..
 

John Luciani

Joined Apr 3, 2007
475
You can replace the multiplier chip with a diode (and a couple of passives).

By mixing (summing) the signals through a non-linear device (the diode)
you are using heterodyning to generate new frequencies. Once you filter out
the high frequencies you will be left with the difference frequency you
desire.

For an example see the Bob Moog article "Build the EM Theremin" in Electronic Musician Feb 1996. Also on the Moog music website there is an article called
"Understanding, Customizing, and Hot-Rodding Your Etherwave Theremin"
which contains the schematic.

I did a PCB layout for the pitch oscillator of the Etherwave. See the bottom picture
at http://www.luciani.org/eng-notes/ee-notes/ee-notes-index.html

(* jcl *)
 

Thread Starter

FredM

Joined Dec 27, 2005
124
I did a PCB layout for the pitch oscillator of the Etherwave. See the bottom picture
at http://www.luciani.org/eng-notes/ee-notes/ee-notes-index.html(* jcl *)
Nice site + lovely board(s)! - I have tried other methods for 4Q analogue modulation before, and not found them satisfactory.. probably because I got something wrong! .. The one which worked (well) for me in the past (with audio input signals) was the ring modulator (4 diodes + transformer)..

As input signals for this design can sum to high frequencies, I chose the 633 (I also happened to have a couple of these spare, which biases me when I am knocking something together for myself, and not designing for production) which has a BW of 1Mhz..

The LM13700 is cheap + simple (and easier to simulate than 4Q multiplication based on non-linear component charactaristics)

To be honest, this 'project' is a bit of a diversion from my primary endevour.. I am designing a range of capacitance based musical instruments using the PSoC CapSense technology.. These instruments ( www.Flewt.co.uk ) are at a reasonably advanced stage of development, and a bit "under wraps" at present, as I intend to put them into production early next year.. They require much better pitch <-> position control than is required for a theramin (and with digital processing achieve this ) .. I wanted to see how a simple analogue design compared... and there is no comparison, the digital system is orders of magnitude more stable / controlled / repeatable, and far easier to play as one can tailor the detection -> output curves, and easily provide a system which switches in semitone steps (when ETS mode is selected).

However, having played with the theramin, I realise it is a completely different instrument.. a lot more experimental and 'weird' !- so I will probably continue a parrallel deveopment on this.
 

Thread Starter

FredM

Joined Dec 27, 2005
124
Hi again John - I have just downloaded EM Theramin .pdf - Thanks for the info, extremely interesting article! The multiplier looks so simple I will give it a try!
 

Thread Starter

FredM

Joined Dec 27, 2005
124
Simple Demodulator..

Take the outputs from the reference and pitch 555's directly to inputs of an EXOR gate (4070).. Then the output from the XOR to a LPF (10k R, 10n C) .. The output you get is the difference frequency!
 
Top