Rhythm Wolf Bass Synth VCA
The VCA has compression on it. The main gain cell is a standard diffpair, but the tail current generator is unique. it has a current mirror that tracks the tail current, and then brings down the CV input as the gain goes up. My main issue with this, is that it completely defeats the purpose of having exp envelopes for the CV, as this compression completely undoes the exp behavior. This compression starts when Q12 turns on, which happens when the current through R73,76 cause a voltage greater than a volt or so (Vbe + R56 drop). C16 puts a time constant on this compression, which is slightly backwards from what you might expect: fast moving increases in CV increase compression, and fast moving decreases in CV decrease compression.
The real oddities in this circuit are the inputs to U1B. U1B sets the lower rail that serves the tail current. U20B is a just a fixed voltage generator, but with its signal taken from the inverting pin, rather than the output. C97 is not populated, as this would make the opamp unstable by delaying the feedback signal. ultimately, this opamp isn't required at all, as a resistor from Vcc to the summing junction of U1B would have done the same thing. U9A is even more curious. This takes the peak of the VCF signal, and tries to adjust the lower rail accordingly. This could act as a sort of compression on the audio signal, as opposed to just the CV. This could be pretty useful for dealing with resonance related volume changes, or if there was a distortion circuit before this. But, since there is no resistor across C58, it just charges up and stays charged forever, so it can't respond to changing ampitudes. its just the peak value it has ever seen. i'm not sure what C55 is about, as it would only allow for a temporary compression. The oddest thing about this, is that the feedback is backwards, such that larger signals actually increase the gain of the VCA. Reversing D8 would fix this. But, utlimately the CV compression circuit mostly over-rides any effect this has.