Posted on November 8th, 2011 No comments
Ever since Samuel Freeman produced some great sounds playing his Vibrati Punk Console through a Korg Monotron at Dorkcamp I’ve been mildly obsessed with the Monotron circuit. The heart of the Monotron is the Voltage Controlled Filter (VCF) which is based on Korg’s classic design for the MS20 synth. Korg did DIYers a great favour by publishing the Monotron schematic back in 2010.
To try and understand the filter better I built a VCF “clone” on breadboard and then moved it over on to a perf-board. It has been an experience that is at times fascinating and at times frustrating. Internally the VCF uses some very low signal levels which are then boosted with a monster-gain amp. Net effect of this is to make the circuit very sensitive to component changes and environmental factors. I had to substitute for the transistors and the FET in the original design and this threw out a lot of other things. I found that swapping the three transistors around could make the difference between the circuit self-resonating or not. It seems unlikely that Korg will be selecting individual components on their Monotron production line so I don’t know how they manage to get good enough consistency in their builds.
The final working model is interesting. Perhaps rather noisier and not as musical as the original, but the sound is certainly striking.
For those that like the details here are the mods compared to the drawn Monotron schematic (component designations are Korg’s):
- Q12-Q14 are 2N3904. As noted above the order of selection of these can make a big difference.
- F3 is J112. As this has a different Vgs from the original I then had to add an AC coupling via a 1uF capacitor to the op-amp with the input DC linked to the voltage reference via a 100k resistor.
- D1 and D2 are red LEDs to increase the clipping voltage on the output.
- R60 is 680k to make sure the circuit will self-resonate
- R73 is 680R to reduce cross-over distortion in the Op-amps output. I think the way that Korg bias the op-amp the output doesn’t go in to the cross-over zone so this wasn’t an issue for them. Because I changed the biasing it becomes an issue for me.