[sdiy] tanh distortion in a filter

[David G. Dixon]

> in some ways you're treading on ground already trod many 
> times in the guitar stompbox DIY arena.
> The largest number of 'distortion' effects use some form of 
> limiting just like you are proposing, although they usually 
> desire to really clip the signal, as opposed to a subtle 
> 'rounding' effect.

I'm looking for "warmth" and "presence", not "grunge" or "crunch".

> Often these circuits use diodes as the waveshaping elements, 
> but almost all semiconductors could be pressed into service as well.
> 
> Guitar distortion also explores asymmetrical clipping, as 
> well as just plain waveform mangling (which is outside our 
> current discussion).

Adding a DC bias is pretty easy.

> The Moog Etherwave theremin used the overload characteristics 
> of the OTA as a tone shaping element.
> Many other examples are out there...

I've got tonnes of LM13700 sitting unused...

> The Moog ladder is curious, at low input levels its capable 
> of very high Q values, but overdrive it and the resonance 
> falls off rapidly, making it more like a waveshaper than 
> filter. Supposedly this was an accident in the Minimoog 
> caused by allowing too-high signals in the filter stage...

Yes, the Minimoog sounds pretty unique, but I think if I wanted a filter to
sound like that, I'd probably just build a clone.  I'm more interested in
warming up my 4P just a little bit.  I suppose I could have done it by using
LM13700 instead of SSM2164 for the interstage gain cells (I have one on each
filter stage, as this makes the filter much "hotter" with increasing
resonance than the typical COTA-style 4P, which can sound quite wimpy at
high resonance IM(NS)HO.  It might be worth building a version like that to
see.  Plus, this would simplify the resonance control, since the OTA is
linear and 2164 is exponential (in the wrong direction for a COTA 4P,
although I've fixed it in mine).  One might even be able to have "warmth"
control by sending current from a single source into each of the four diode
linearization inputs of the OTAs to linearize or non-linearize them.
However, this would probably be very touchy.  It looks like some more
simulation is in order...

> Saying you are walking familiar ground does NOT imply that 
> its not good ground to walk on, its fascinating territory !!!

It's new to me!