Through-Zero FM

[Donald Johnson]

 "jhaible" <jhaible at primus-online.de> wrote:

>Hi Dan, you may be right in theory, and I have long considered this method
>when I designed my Frequency Shifter.
>
>The main problem for VCO applications is that you don't have V/Oct tracking
>anymore ! In your example you would need a variable VCO that goes from
>5100 Hz to 5200 Hz with a 1V CV step, inorder to have the output changing
>from 100 Hz to 200 Hz. 
>Our ordinary exponential VCOs would do something very different; they would
>go from 5100 to 10200 Hz with a 1V CV step.
>
>This would not be a problem with V/Hz VCOs of course. And one could 
>construct a "hybrid" VCO that could do the trick, too. (Offset current ...)
>But you'd have problems with ordinary V/Oct VCOs.
>
>JH.

 I haven't finished building my frequency shifter yet, but I have
 toyed with a linear (V/Hz) VCO, but adding a modulation control
 voltage input with a sinh() or modified cosh() response:

 By using a pair of exponential voltage converters I made:
   Vout = sinh( Vin ) = k * ( 2^+Vin - 2^-Vin )/2
 or
   Vout = cosh( Vin ) = k * ( 2^+Vin + 2^-Vin )/2

 This yields a shifting frequency with a sort of exponential
 response on either side of the centre frequency.

 The cosh() function needed to be shifted and inverted for
 Vin less than 0. (Have a look at one of your old math texts.)

 Considering the large number of components in the shifter, I
 don't think the extra converters are too much.

 don.