Tube/Valve Tannerin/Ondes Martenot

Hi everyone,

I am thinking of making a musical instrument with a sliding controller, similar to the Tannerin or Ondes Martenot. I want to use a sine wave oscillator for the sound if possible, or a triangle wave.

I have been looking at different types of oscillators, and I think the best way to make it would be with a beat frequency oscillator like the theremin. Would it be possible to modify a theremin circuit to use a sliding controller (variable capacitor)?

I don't want to use computers or anything digital, or even ICs if possible. The original Tannerin, Ondes Martenot etc. didn't use those, and I like that.

Any help would be much appreciated. 

Lucia

I had hoped that there was something simple that would work, but it looks like there isn't.

OMG  

Technically, any LC oscillator will work just fine with a variable condensator instead of, or in addition of the fixed cap. The issue is how you will control that capacitor in order to get an usable musical instrument.

"The original Tannerin, Ondes Martenot etc. didn't use those, and I like that."  - StaticBuildup

No offense intended re. the following.  

Hard constraints on a project can often help channel creativity.  But if there is the option, these limitations should be selected up-front and very carefully.  If Martenot, Theremin, Tanner, etc. were alive and active in the field today they wouldn't be using tubes, discrete transistors - or even much analog for that matter - for the core of their designs. Those elements are now largely relegated to the retro movement (which I generally associate with a lack of creativity, ironically enough).

I know that it's a job in itself to come up to speed on digital approaches, and everyone has to start somewhere.  But tubes don't even have a complementary driver (NPN, PNP) so you have to introduce reflective power elements like transformers to do push-pull, and to lower their output impedance.  There are good reasons they were almost completely abandoned the second some mediocre, affordable transistors hit the scene.  Maybe start with transistors or simple ICs.  Though producing a "perfect enough" (no audible harmonics) sine wave in digital (via polynomial) takes like 16 instructions, which can be quite alluring.  And separating the controller from the sound generation helps to compartmentalize the problems on both ends, which can be quite freeing (sensor & conditioning => number tossed over the wall => sound generator) - an awful lot of conquering is dividing.

But I do understand how interesting it can be to follow in the footsteps of the founders.  If that is indeed your main goal with this project.  For me (and I don't mean to make this all about me), I just want to make the best possible instrument, by whatever means necessary.

Hi Christopher, Dominique and Dewster, thank you for replying. 

I guess I will have to start with something modern. I had hoped that there was something simple that would work, but it looks like there isn't.

I have got a sawtooth wave oscillator that works perfectly over a good range and can be easily controlled with a variable resistor, but the sawtooth wave tone isn't what I'm after. If the output could be modified to something more interesting, that would be ideal.

The reason for the constraints is that I do electronics projects which are mostly unusual or retro, and my experience has been that people are much more interested in those than in computers, microcontrollers and similar things. There is much more to see, and it is different from the technology we are already surrounded by.

I would try to learn to play the theremin if I could, but I have a disability and I don't think I would get very far.

"StaticBuildup is a name that would send most theremin designer running."

Some of my projects are about sparks and static, but I keep them away from the others!

Lucia

There are function generator ICs that can give you sine / square / triangle / sawtooth waves.  I once designed and constructed a function generator from op-amps and the sine was produced by non-linearly distorting a triangle wave via a transistor pair.

There's wein-bridge too for sine.

The digital approach has a pretty steep learning curve, but once you're past it and have some good constructs to draw on, almost everything is easier than doing it analog.  I'd much rather slave over code than a hot iron.

dewster, I don't agree with you about tubes, transistors and digital. Each technology have its advantages and disadvantages. Even tubes have advantages, otherwise guitar players and the armies would not use them anymore. 

As a musician, I like the hazard we get with the imperfections of the analog technologies. You can simulate tube amplifiers with the digital technology, but you will never get the same feeling and experience than with a real tube amplifier. For me, that's the end of the story. I can use my PC to play at home in the night, I plug the guitar, run a software effect rack and put headphones, but most of the time, I use my classical guitar at home or even better, in the garden or in the wild - I like to play with the birds, they are our best teachers. I also use the PC for recording and mixing, digital is very good for that and for a lot of other tasks, but to make the sound, I really prefer the analog technologies. It was a long time ago, I played with a DSP (Digital Signal Processor) and A/D and D/A converters. It was time consuming to make assembler code for it, but it was very fun at the same time. I was able to retrieve in real time the notes I was playing on the guitar, but this was not as fun than playing the same guitar directly plugged to a good amplifier. I just get bored with that DSP after a few weeks playing with it.

Dominique, I prefer the sound of my classical guitar too, but that's an acoustic controller / sound source.  And a guitar tube amplifier + speaker is a very complex effect that's difficult to capture / model.  That's not really what I'm talking about in terms of analog vs. digital regarding instruments that are fundamentally electronic, such as the Tannerin, Theremin, Ondes Martenot, etc.  The original models were made with tubes because that was the best and only solution they had at the time.  

One can certainly learn a lot in the process of making a replica, or while expanding on the original in the original medium, but one should be aware that that is what's going on, that there are vanishingly small advantages (ESD, certain circuit topologies) and many substantial downsides (availability, expense, size, heat, drift, voltage, power, limited lifetime, chassis & cabinetry, etc.) associated with the use of tubes in these types of musical instruments.  Some of the knowledge obtained will be broadly useful, though much will be arcane / academic / moot in this more modern world.

I had to step out of this thread because it started to be derailed, robots can only think like robots.

We are visited by a creative lady trying her best to do something with sound as an Artist. She has a vision and also knows secrets.

Christopher

This analog vs. digital stuff is going off-topic a bit and I probably shouldn't stick my nose in, but it sounds as if Lucia has good reasons for wanting to avoid the items he mentioned right up front, and it may be because tradition and some semblance of authenticity might carry more importance to him than pure functionality and performance.  Sometimes vintage technology is just cool and worth replicating, despite all the the quirks and reliability issues.  Sometimes people prefer a volume knob instead of up/down buttons and presets with menu screens (ask Ford).

People are still building tube and discrete analog devices not necessarily because they are superior, but because that's what they can do or want to do, and it is entirely suitable for the project at hand.  I think we all appreciate how much performance and functionality you can pack into a digital design, but along with that low cost, feature-laden capability comes the baggage of software and impending component obsolescence that can doom the entire product to the landfills in a few short years. I'm not sure that anyone will be coveting and lovingly restoring a Moog Theremini in 75 years in the same way that they would today treat an RCA theremin, for example.

And as far as analog skills being arcane (!), remember that these high density, high-speed digital Lego blocks that are proclaimed to be superior wouldn't exist without the fundamental analog elements and engineering inside.  And as you know if you've ever been involved with very high-speed digital designs that fail to function even though they are wired, timed, and programmed correctly, the problems often turn out to be caused by ignorance of or even total disregard for proper analog design techniques.  Digital technology has done amazing things at lower cost and will do more, but in some applications it can be a step backward in an analog world.

Back on the original topic, I too have been interested in the Ondes Martenot since I first heard about it. I think I ran across someone that was building modern replicas, but I couldn't find any contemporary circuits. 

After my little rant dissing digital above I hesitate to admit that my first thought was to take the simplest possible approach and hack into a cheap MIDI keyboard just to use the pitch wheel interface electronics.  Make a sliding wire arrangement that turns a pot over the full range of rotation for the full span of wire motion, and then just wire that pot in place of the pitch wheel pot.  My keyboard pitch wheel can do a four octave span, but that would vary with the synth.  But this approach is quite off-topic as well.

The sliding wire interface could be fun to use, and possibly more agile than a ribbon controller, except vibrato might be a little harder to keep stable.  It would be nice to try something like that just to see what it felt like.