Speaking of Sine Shapers

[Haible Juergen]

Speaking of sine shapers.

I bet this is not new either, but I discovered this just some weeks ago.
I was looking for a simple but fairly good triangle to sine shaper
(for the PWM of my PolyModular VCO).

You know that a single overdriven OTA does the trick - somehow.
Problem: When the input amplitude is trimmed for minimum THD,
the waveform has a rather pointed peak. So you have to choose 
your compromise, different overdrive for VCO or LFO applications.

There is the 2-stage solution used by Moog. Nice, and quite
involved. (Two stages ...)

And there is the degenerate emitter method, where a discrete 
differential pair is used instead of the OTA, and small emitter
resistors are added. I've mostly used this method so far.
(See FS-1)

But then it struck me that these emitter resistors are just a method
of (local) feedback, i.e. slightly linearizing the differential pair's
input function.
The same thing must work with a standard OTA and an external
feedback loop ! And it does. 3080 driving inverting opamp at
virtual GND summing node. Output of opamp *slightly* fed back
to the OTA's noninverting input. Works like a charm. 
I'll post a circuit of this with the next PolyModular update.

JH.