Voltage Controlled Glide

This circuit is by Jim Patchell

V.C. Glide Schematic

Glide circuit

The circuit built around U14:A and U13:B is really not much more than a single pole low pass VCF. Assuming at first that the input CV is 0v and that everything has settled so that U13:B's output is also at 0v, the differential voltage on the U14:A OTA will be 0v, so there will be no output current, and so the steady state is preserved.

The configuration with the internal OTA input diodes biased on by R23, will cause the OTA inputs act somewhat like "virtual grounds", that amplify the differential currents at the inputs. We convert the input voltages to currents via the resistors R22 and R24.

Now let's apply 1v to the input. This will make the - input more positive than the + input, so the OTA output current will be negative (flowing into the OTA). Since this current drives the integrator built around U13:B, the output of U13:B will ramp positive at a rate determined by the control current coming from Q8. As the U13:B output goes positive, this is fed back to the OTA + input, which will decrease the differential input voltage to the OTA.

Since the output current of the OTA depends on both the input differntial voltage and the control current (See Eq 1 below), that means the OTA output current will decrease, which will in turn cause the output of U13:B to ramp up at a slower rate, which will continue to slow until it reaches the input voltage of 1v. This condition produces the classic exponential shape of a one pole RC filter. Varying the control current has the effect of using a potentiometer for the "R".

    Eq 1)  Iout = 19.2 * Vdiff * Iabc

Exponential Current Source

The exponential current source built around Q8 and Q9 and U13:A is the familiar configuration we've used in other circuits, described in more detail elsewhere. Since this circuit only controls the "virtual potentiomenter" of our simulated RC circuit and thus the rate of the glide, it is not necessary to temperature compensate it.

The most critical component in this circuit is the OTA. Half a CA3280 is used here rather than a LM13600 or LM13700 (can't use a CA3080 because it doesn't have the input bias diodes) because it has superior offset characteristics than the other OTAs (it's also harder to find and more expensive!). Offset voltages in the OTA result in control voltage feedthrough, and even a mV or two would result in noticable control voltage errors. In a pinch an offset trimmed LM13700 could be used, but the CA3280 will work better.

Pitch Bend Summer

Rather than use a separate patch cord to each oscillator to sum in the pitch bend voltage, I chose to sum it with the keyboard CV. This is the purpose of U15:B. Since the summation also produces an inversion, we have to invert the signal again with U15:A. In retrospect I could have eliminated the need for U15:A just by swapping the inputs of the OTA. Then U13:B would have already been inverted, and U15:B would invert again to restore the to proper polarity.