v-mid: starting to understand the problem...

Before undertaking the substantial effort of changing my v-mid circuit back to "v1" as I discussed previously, I took some time to try to understand the Baxandall tone control design a little better, and I think I made some headway.  In short, I don't think there's a practical way I can use it for my v-mid circuit: i.e., there's no way to save the "v2" circuit, it simply has to go back to "v1".

This shows the bass & treble active tone control from Baxandall's original paper.  The fact that the treble pot is connected directly between the input and output lines, was what misled me in my attempt to adapt this circuit to the variable-frequency midrange control.  I basically have my level pot directly across input and output in the same way, and then I substituted my variable-frequency bandpass circuit for the capacitor C3 in the circuit above.  But this is not really valid (as the failure of my circuit to operate properly, demonstrates!).

In my seat-of-the-pants (or skirt) approach to engineering, I now look at the schematic above, and understand it to be more or less a duplicate pair of R-C filters for each of bass and treble, one filter on each side of the level control pots.  Thus, when a pot is set for full boost, it is selecting the R-C on the input side; when it's on full cut, it's selecting the other R-C, driven by the inverted feedback signal.  The pots mix between the two.  Not sure if it's quite that simplistic, but it seems roughly reasonable, and this starts to explain why my v-mid modification didn't work: I didn't have two identical copies of the bandpass R-C circuit -- I would've needed a four-gang frequency pot for this!  I just had the one circuit, in the "middle", on the output leg of the level pot.  When the pot was turned to full cut position, the input of the R-C was tied directly to the output, which might as well be a power rail or ground.  No chance of *any* input signal affecting the circuit, coming through the full 500k resistance of the pot.  So the simplest solution to the differential equation was zero: no signal out.  So in my "hand-wavy" way, it makes sense that it behaved like a volume control.

So unfortunately, I don't immediately see any simple way to rectify this problem, within the bounds of the Baxandall topology.  This serves to solidify my resolve, that the only way to make the v-mid function work, is to revert the circuit back to my "v1" design.  It seemed like kind of a kluge, but hey, at least it worked!