DC Transfer function problem...

[Magnus Danielson]

> Hello all,
> 	over the last couple of days I've been trying to build a
> compressor/waveshaper using a VCA block (based on an LM13700) and some
> control circuitry. I'm having difficulty in generating an appropriate DC
> transfer function for the control circuit: what I'm after is a transfer
> characteristic something like this:
> 
> Vout
>    |
>    |
>    |--------------\
>    |               \
>    |                \
>    -------------------------------
>         Vin
> 
> where both the `knee' point (threshold) and the slope (compression ratio)
> can be changed independently. Note that this *isn't* a frequency domain
> graph :)

Well, what if you let it be just that... a simple linear VCO (say a 4046) and
a filter with the "rigth" properties and then a full-wave rectifier... it may
not allways be a bad solution.

Just a simple circuit idea came to mind:

The offset Vout value is set by a pot. To this is a non-linear current source
connected, such that it will at a certain point of Vin will draw such a high
current that it will pull the voltage down. The non-linear current source is
being built around an op-amp and a diod (or possibly a transistor) to get a
exponential (antilog) function. This is just rougth edges, but the general idea
is more important here then "that pot should be over there instead" or where 
the
signal should be inverted...

You should find enougth material in the books you have on how to make this.

A small current wount't change the resistor, an increesing one will. Maybe 
should a fixed resistor (to +Vcc) be used and then let the put change the 
offset
current for the conversion.

You may change the knee with a resistor over the diod.

Hope it gives some inspitation...

Magnus