simple way to add harmonics to a fundamental signal

[Martin Czech]

> > The question is how to get the x**3, x**2 functions needed.
> > Multipliers are expensive, that's why I thought about a nonlinear diode network.
> 
> Did you consider x**n=exp(n*log(x)) ?
> 
> log(x) might be obtained once and used to all x**n terms required.
> I don't think very precise log/exp converters are needed here, so it's only
> one tranny and one opamp per x**n term + one pair for log(x).
> 
True. And even better, n could be a real.
So n is adjustable.

But wait a minute...
What about x=0 and x<0.
If we aggree that a real value is returned, we have problems
with both, x**n amd log(x).

The actual implementation would decide what to do in these cases.
Perhaps x=0 could be handled by some special circuitry to avoid
extreme circuit reactions. If this is possible we could limit x to positive
values via rectifier. 

This solution is atractive, because Ut of the expo and log circuit will cancel.
So no temp compensation is needed (e.g. 4 trannys of a CA3046).

m.c.