Moog ladder filter revised

After all the work I did on improving the range and controls of the Juno filter, I decided to have a closer look at the Moog filter. I have improved the following:

• VCA CV is now similar to the Juno - it has an adjustable deadband at the start and a summing point for multiple CVs

• Resonance CV is similar to the Juno in the same way

• Cutoff CV summer has been altered to give a 10Hz cutoff from a 0V CV, with +/- 5.6 octaves trimability, fairly similar to the Juno.

• I have added a tempco resistor to the exponential converter.

• I have added an option to tap the audio output at 12dB/oct instead of 24. Resonance is still tapped at 24dB to get a nice high resonance. No switch has been drawn but either an electronic or mechanical DPDT switch will do.

• I replaced the output buffer with a transinductance circuit and instead flipped the inputs to the VCA OTA to keep phase. This fixed the maximum cutoff frequency similarly to the Juno.

• I added a linear FM input to the cutoff frequency.

In addition to all this, I've completely changed the signal attenuation/gain to enable overdriving.

Overdrive 

The Memorymoog schematics indicates that a 40mV p.p. at the transistor base of the bottom ladder transistor is the 'norm'. My simulations show severe distortion at this level. I could see distortion all the way down to 10mV p.p. I have suggested three component values that will work for 5, 10 and 20V p.p. input (with a 15mV p.p. base voltage at those values). This gives a slight distortion at the highest inputs, and further distortion should the input exceed the specification. For example, a 10V p.p wave maxes out at about 16V p.p even with much higher input.

How the overdrive sounds in practice is still unknown.

Input (blue) vs output (green), input is 8V to 30V p.p

Computer calibration

It would probably be possible to tune most parameters of the filter using lookup tables in the computer control. Here are some thoughts on that:

- If no deadband is needed (i.e. 0V input to the VCA CV gives an 'off' output), the U17 trim pot and R41 may removed. Keeping R41 but connecting it to -15, and changing the value to 1.5MOhm will give a 100mV deadband.

- The U13 VCA gain potentiometer can be replaced with a fixed resistor that gives a more-than-unity gain at max CV. By doing this, the VCA CV can be calculated in software to give a correct curve with a end point that results in unity gain.

- Similarly, U20 can be chosen in such way that it results in a resonance higher than what is wanted, and CV then corrected in software for max resonance.

- I would probably still keep the tracking potentiometer, but R36 can be replaced with a lower value, giving a lower frequency as a starting point. Then you could replace R32 with a 25k resistor to get 20 octaves of CV control, enabling proper calibration of lowest and highest cutoff.

- It IS also possible to replace the tracking pot U9 and resistor R23 with a 100k resistor. Tracking could then be adjusted in software.