Serge Waveshaper was Re: Harmonics question

[Magnus Danielson]

>>>>> "SC" == Sean Costello <costello at costello.seanet.com> writes:

 SC> At 09:59 PM 5/19/98 EDT, you wrote:
 >> 
 >> In a message dated 5/19/98 3:52:56 PM, magnus at analogue.org wrote:
 >> 
 >>> I recall seeing diagrams of such unlinear waveshapers. I even think
 >>> that we have discussed this several times. For instance do I recall
 >>> that the Serge waveshapers would do odd and even harmonics and that
 >>> the schematic is very simple indeed.
 >> 
 >> 
 >> Hi Magnus
 >> 
 >> I wonder if you're confusing the Serge Triple Waveshaper with the Wave
 >> Multiplier module. One waveshaper is a very simple circuit based around one
 >> quarter of an LM 3900 and does non-linear "Harmonic Reduction" (unlike
 >> anything done in music theory). It is similar to a CA 3080 tri-to-sine
 >> converter I believe.

 SC> The Triple Waveshaper does two things:  it converts a sawtooth wave to a
 SC> triangle wave (via full wave rectification), and converts that into a sine
 SC> via, um, breakpoint...diode...thingies...(man, I wish I knew the lingo).
 SC> The amount of rectification is voltage controlled, as described in Serge
 SC> Tcherepnin's patent.

 >> The Wave Multipliers do some nutty half and full wave rectification (I
 >> believe) and have a whole lot of chips. The sound to me is very much what I
 >> would expect from the timbre modulator circuits in the oft mentioned
 >> Electronotes and in Barry Klein's book. I would never attempt to build one,
 >> too complicated for a timid sole such as me. Now if some one offered some kits
 >> of these circuits...

 SC> As far as I can guess, the Wave Multiplier works as follows.  (This subject
 SC> has come up a lot, but the Wave Multiplier is one of my big obsessions,
 SC> along with Moog filter distortion, the Barberpole phaser, and washing the
 SC> germs off of my hands several hundred times a day so they stop sending my
 SC> brain radio waves).

 SC> The first section is a simple clipping amplifier, that "rounds" the signal.
 SC> Possibly a CA3080 that is overdriven.  The amount of overdrive is voltage
 SC> controllable.

A LM13600 I think from the back of my head from an earlier take on the
Serge Wave Multiplier.

 SC> The second section is...well, no one really seems to know (prove me wrong,
 SC> folks!). It seems to work as follows: When the signal voltage reaches
 SC> certain thresholds (positive or negative), the signal "wraps around."  Easy
 SC> to do in digital (the Nord Modular implements it, and it sounds GREAT), but
 SC> I don't know how to do it in analog. 

This sounds exactly to what you can do with the Formant Waveform
Processor (described in the second Formant book, check Anders
Sponton's nice site).
The Formant version will only fold over on one side, but it is simple
to do for both sides. The schematics consist of 3 op-amps where one is
just a standard summing amp of two input amps. One is the input buffer
for the modulation CV/Clipping Niveau. The output of the modulation
opamp hit a standard diod in reverse operation. The other end of the
diod is connected to a resistor from the signal input buffer. The
diod/resistor point signal is then feeded to the negative input of a
summing amp. The positive input of the same summing amp is being
feeded with a pot-dampend version of the output of the signal summing
amp. The output of the last summing amp is the actual output.

There. Add another op-amp for negative wrap and have fun!

Was this what you where looking for?

 SC> The third section is four voltage-controlled full-wave rectifiers in series
 SC> (see the Serge patent for details).  Probably uses all four sections of a
 SC> LM3900.

 SC> Any more details, folks?

I hope I could add some...

Cheers,
Magnus