AW: Re: AW,2:Re: AW: Another VCF Finished

Haible_Juergen#Tel2743 HJ2743 at denbgm3xm.scnn1.msmgate.m30x.nbg.scn.de
Wed Apr 17 21:50:00 CEST 1996


>     without the 100K/220 linear-mode input resistor dividers on the 3080
>     inputs? Is this to utilize the negative feedback to keep the
>     differential input voltage on the 3080 to zero, thereby eliminating
>     any non-linearities? Is there any danger to the 3080, during
>     slew-limited clipping for example?

Oh. I saw that I had accidentially quoted the ota-without-resistors stuff as 
well.
Well, ota circuits with many otas and few resistor dividers do exist in some
integrated circuits (later CEM or SSM filters amongst them!), but that
wasn't what I had in mind.
Remember that the nonlinearities are eliminated only at *low* frequencies,
i.e. way below the pole frequency of this stage. What I had in mind
(and what I have actually built in some all pass stages), is a design that 
uses
a far smaller attenuation ratio than 100k/220. (It was something like 22k
or 10k / 220, if memory serves). With the 100k/220 divider you make sure 
that
there is no serious distortion from any signal between +/-10V at any
frequency. With a 22k / 220 divider *some* frequencies may be able
to overdrive the ota inputs if they are larger than a few volts. This
may be tolerable to some extend because
(1) "some frequencies" would be those equal or larger than the pole
frequencies in an allpas design. This is interesting if you use a phaser
on string sounds, with the pole frequencies rather high (a "slight" 
phasing).
The spectrum of the string sound isn't strong enough at the high end
to overdrive the phaser; so you gain a lot of SNR. (just like with
preemphasis in good old FM radio).
(2) "some frequencies" might be those frequencies close to the cutoff
frequency in a LP filter. Distortion on higher frequency is filtered by the
following stages and sounds rather pleasant.
(3) You don't have a hard clipping point in an OTA input (without
linearisation diodes). So the effects mentioned in (1) and (2) come
in very smooth and gradually.

If you like this effect or not is just a matter of taste. But building a 
filter with
100k/220 dividers gives you no choice: you're always on the save side.
OTAH, with  22k/220 dividers, you have the choice to have a clean
filter or an overdriven filter, depending on input attenuation.

JH.



More information about the Synth-diy mailing list