[sdiy] Building an Ondes Martinot or French Connectionstylecontroller?

[Mike]

> Thanks ken, i was thinking this was a possible solution. Very cool..
> Is this pulley something that goes around once (so it's circumference
> is slightly larger than the length of the travel?) or multiple times?
> If the pulley goes around multiple times that means the string goes
> around the pulley multiple times as well, and would there be anything
> there to stop it from getting tangled up if 10 loops of string go
> around it for example?
>
You really need to look at a few old radios. This a very common solution to
the tuning dial problem, the difference here being in the gearing. A tuning
dial needed several turns of a knob to move a variable capacitor through
roughly 180 degrees. here you are 'driving' the device by effectively moving
the 'pointer'. Bang and Olufsen used a scheme like this for the 'slide
rule' tuning system on some of their portable radios and tuners/receivers.

By the way, if I remember correctly from working on one briefly, before a
performance of Messians 'Turangalila Symphony', the Ondes Martenot ring
slider did indeed operate a variable capacitor as part of a valve
oscillator. This may have been a BFO type. Thinking of the technology of the
day, and the radio expertise around at the time, this does seem likely. Has
anyone got a pointer to an Ondes Martenot circuit diagram? By the way - the
diagram alone is nowhere near enough to build one: the strange louspeakers
had as much, if not more, to do with the tone. The only way I can think of
to come up with a reasonable substitute is a PC based 'model' or Scott with
his FPGA modelling. How to do the analysis presents a whole other challenge.

There are some very good, very linear, light action, 5 turn pots out there
as well as 10 turn ones. I think that's where I'd start too.  If wear is a
problem, then a high quality, absolute shaft encoder, a processor and an
A-->D might be another option. There must  be a capacitance based solution
too, though I think that might take you into the exotic instrumentation
systems
area.

HTH

Mike


> On Dec 2, 2007 3:02 AM, Ken Stone <sasami at hotkey.net.au> wrote:
> > The string slip problem is solved by anchoring the string at the ends to
the
> > pulley. It is not a continuous loop, but rather a long string (3 times
the
> > length of the travel + spare to cover the remaining distances between
the
> > pully and the playing area), with a one length wound on the pully, one
> > length along the keyboard, and one length returning from the other end
of
> > the keyboard to the pulley again. Usually one end was terminated in a
spring
> > housed within the pulley itself.
> >
> > Ken
> >
> >
> > >I've built a few simple portamento controllers, and I want to build a
> > >sort of Ondes Martinot controller (Sort of like the Analogue systems
> > >"French connection" controller -
> > >http://www.analoguesystems.co.uk/Reviews/fconnection_review.htm). I
> > >mean JUST the portamento string-ring interface, not the keyboard
> > >interface. If nobody's seen one before, here's a nice video explaining
> > >one.. http://youtube.com/watch?v=ybYIhomm5KM
> > >
> > >I've never actually seen the inside of any of these controllers and
> > >I'm sort of curious as to how they might work. I'm wondering if anyone
> > >has seen inside, or know exactly how they might work mechanically?
> > >
> > >My first guess would be to have the string wrap a few times around a
> > >pulley wheel attached to a 10-turn potentiometer (or a geared down
> > >single turn pot, to avoid having a pulley wheel with a circumference
> > >as long as the actual playing field area). I'm wondering if maybe
> > >either of these controllers came up with a way of avoiding the string
> > >from "slipping" on the pulley over the course of time, or if just
> > >wrapping a string around a pulley multiple times would cause enough
> > >friction to avoid this (I'm reminded of old FM tuners, which I THINK
> > >slip over time). Also, if there's something to deal with tension on
> > >the string.. and what kind of string it might be?
> > >
> > >The second thing (and this one is a lot more complicated in my mind)
> > >is how the "button" style volume controller works. The best solution I
> > >came up with brainstorming last night would be to have the button a
> > >miniature version of a cry-baby wah-wah pedal, but with a spring so it
> > >returns to it's top position. The lever would be attached to a geared
> > >shaft which would ride against the gear of a potentiometer (just like
> > >in the wah wah pedal).
> > >
> > >Mike