Freq Shifter / Ring Mod etc.

[Haible_Juergen#Tel2743]

 > If you can use AD533s to make a freq shifter, why can't you use 1496s?

You can, but one of the inputs of the 1496 is quite nonlinear (like
an OTA input). Ok, you could use this one for the Oscillator input
(rather constant amplitude), so SNR might be not that bad.
Or you might even use the nonlinearity to shape a sine wave from
a triangle oscillator, as Don once suggested. But then you could
do no more filtering after the sine shaper ...
If memory serves, the Bode / Moog frequency shifter uses two
1496's in the BFO section, but two linearized ones ('95 or '94)
in the shifter section. But this is just vaguely from memory,
not sure at all.
Thinking of it again: Using the 1496 with the BFO at the nonlinear
input (fixed amplitude) might not be a bad solution (SNR-wise),
even without overdriving/waveshaping.
I haven't made a layout for this part yet, so I might still switch to
the 1496's ... any general opinions about RC4200 vs. 1496 ?

JH.


PS.: I started with the layout for the dome filter (hilbert transformator).
        I *hope* that the electronotes-solution will work (no trimmers, but
       measuring of C's and composing the R's from two measured ones).
       The oscillators will be Bode-style, i.e. similar topology for fixed
        and variable oscillator to have some temperature compensation
        (In the electronotes shifter, there is a fixed wien-bridge 
oscillator
         and an OTA-VCO.)