Minimoog clone VCF

The core of this filter is almost exactly the same circuit as in the Minimoog. Only a couple of minor changes had to be made, to be able to run it on +-15 volts, instead of +-10 volts. These changes were developed by René Schmitz, and you can find the schematics on his web site.
To get the best performance, I matched all transistors and capacitors in the transistor ladder. Of course, the transistors for the differential amplifier were also matched.

I added an opamp to get the output amplitude up to my standard of 10 volts peak to peak. Another opamp vas used as an input adder. The input impedance of the original circuit (10k) was a little low for my taste. I also added opamps to make the modulation amount knobs bipolar. So by turning the knob to the left you get negative modulation and to the right positive.
Finally I added my signal indicator circuit.

This filter really sounds sweet. It has the classic Minomoog sound, and noise is quite low. I don't have genuine Minimoog to compare, but maybe my filter is a little quieter thanks to the BC560 low noise transistors.

Circuit board layout (PDF-file)

Component placement (PDF-file)

This is how the Minimoog VCF module looks from the back. Note the written numbers on the transistors. This is measurements for the matching.

The resonance control (labeled "emphasis") uses a reverse taper potentiometer on the Minimoog. These are quite difficult to obtain. I tried with a linear pot, but the function was not at all satisfactory. The solution was to use a normal logarithmic taper pot, with gears to change the direction of rotation.

The assembly containing the reversed emphasis pot and the TM pot. The gears are 40 teeth, module 0,3.

This picture shows the pot assebmly installed. Note the compact construction.

Korg-35 clone VCF

This filter is believed to be functionally identical, or at least very similar, to filters built around the Korg-35 IC. It was used in the first series of Korg MS-20 and in the MS-10. Later MS-20s used an OTA-based filter, that sound different.
The Korg-35 is a very simple design, built with only three transistors and one op-amp as active components. This is not the design to use if you want low distortion and low control voltage bleedthrough. But the filter has a rather unique sound, which makes it a worthwile complement to other filter designs. The core circuit for this filter was designed by Jürgen Haible, based on information from K. Tanaka.
Like the MS-20 (but unlike the MS-10) my filter has two cascaded filter sections. On the MS-20, the first section is highpass and the second is lowpass. On my filter, both sections are switchable to either highpass or lowpass. Each section has a slope of 12 dB per octave. With a toggle switch, I can gang the cutoff and envelope controls. This way, I can get a 24 dB/oct response if I set both filters to either lowpass or highpass. The peak and FM controls are still separate, so you could achieve some special responses and modulate the filters separately. The latter might be interesting in conjunction with a high resonance.
Unlike the Korg synths, my filter tracks the keyboard (1 V/oct). The tracking acurracy isn't very high but it is possible to play melodies while the filter oscillates. There is no temperature compensation though. The tracking can also be set to half or a fourth (2 or 4 volts per octave). This is useful if you want higher notes to be a little less bright, which is often the case with acoustic instruments. To prevent the filter from thumping when you play the keyboard, I have added a lag circuit to the keyboard tracking input.
I have added separate knobs and inputs for envelope and general FM. Both knobs are bipolar, so the modulation can be positive or inverted, depending on which direction the knob is turned. The envelope input has one thumbwheel per filter to select source.
Another addition is a signal indicator. It drives a bicolor LED, that lights green when there is enough signal and red when there is too much signal.
This filter has high distortion when the input signal is strong. This can be used to good effect. Therefore I have added a drive switch. In the normal mode there is only slight distortion when the filter is fed with a 10 V p-p signal and the input signal knob is set to "8" (unity gain).
But with the drive switch in "high" position the input gain is increased ten times. Now you can use the input level knobs to control the amount of distortion. The output gain is reduced in this mode, to prevent overloading of the following circuits.

This is how the Korg-35 VCF module looks from the back.