BassBalls - Envelope/Variable filter interface doubt

[lepra85]

Hi All,
i am going to build a BassBalls and before start to build i am trying to understand the circuit.
the schematic of the Circuit that i am trying to understand is the following:
http://experimentalistsanonymous.com/diy/Schematics/Filters%20Wahs%20and%20VCFs/Electro%20Harmonix%20Bassballs.pdf
but, for this specific doubt, the Doctor Q schematic is also representative:
http://experimentalistsanonymous.com/diy/Schematics/Filters%20Wahs%20and%20VCFs/Electro%20Harmonix%20Doctor%20Q.gif

The BassBalls (as all the autowhas that I saw) can be summarized as an "envelope detector circuit stage" that modify the center frequency of a "filter stage" by modifying a "control element" (an excellent explanation of the autowhas is on:http://www.geofex.com/article_folders/ecftech/ecftech.htm).
In the BassBalls the "control element" is a transistor, used as a variable resistor (varying with the envelope). To have a good correlation between the envelope and the resistance of the BJT, I understand that this one must be biased to work in a linear way.
here the doubt:
Is the BJTs of the bassballs biased to work in  a linear way?
seems like the transistors are in cut-off mode if there is no input signal, is that ok?

as always, thanks for your comments.

[PRR]

> must be biased to work in a linear way.

Why?

Here we trying to change a pitch in some relation to loudness. What should that relation be? Linear? Exponential? I utterly do not know. Pitch and amplitude are two very different things.

> Is the BJTs of the bassballs biased to work in  a linear way?

The whole thing is clearly non-linear. The crude peak detector is non-linear. The buffer is non-linear. Together they do nothing for small signals, and don't get "pretty linear" before the detector driver runs out of headroom.

The filter BJTs would be near-linear if driven with linear current, near-exponential if driven with linear voltage. But here they are driven from trim pots of significant resistance-- they are neither linear nor exponential.

Together they may work elegantly. The peak detector is "lazy" for low amplitude but the near-exponential response of the BJT at low drive may cause a large (obvious) shift of frequency. At the upper end the near-linear response of resistor-driven BJTs make keep the frequency from soaring to the sky too fast.

Much more precise detectors and filters are readily designable (by a clever experienced signal processor geek). Would a more "math perfect" response really sound better in music? After the 3rd beer? Enough to justify many more hours with slide-rule and breadboard?

I don't mean to discourage you. Only pick your battles. You have a "winning design" (I assume somebody likes it), well documented, no obscure parts.

> transistors are in cut-off mode if there is no input signal, is that ok?

If there is NO input, there should be no output, so who cares what it sounds like? The transistors won't be harmed by being cut-off. The filter action is limited by the 47K in parallel with the BJT, so the pitch won't soar down to zero (bottoms at 50Hz?).