Shruthi-1 IR3109 filter board

About this board

The IR3109 is Roland’s VCF building block, used in the SH-101, the MC-202, the Jupiter-8, the Juno-6 and others. This chip contains an exponential converter and 4 OTA-buffer blocks ; just like the SSM2040. Through different configurations of external resistors and capacitors, various voltage-controlled transfer functions can be achieved.

This filter board uses the IR3109 in a traditional 4-pole low-pass configuration. The choice of resistors and capacitors values is the same as on most Roland designs. Does the OTA input gain put the signal in a sweet spot in terms of distortion, SNR, and CV feedthrough? I don’t know, we have to trust the Roland engineers on this one since there’s no IR3109 datasheet around! The CV scaler circuit was adapted to match the Shruthi-1’s 0.46875 V/octave scale. A LM13700 is used to implement two linear VCAs: one of them controls the amount of filter output fed back to the filter input (resonance), the other one is the final stage VCA. There’s nothing much going in this board – you can think of it as a variant of the SMR-4 in which the bottom LM13700s, TL074, and the exponential converters are all rolled into a single chip. The only bit in which it significantly differs from the SMR-4 is in the way resonance loudness compensation is achieved. On the SMR-4, this is done by boosting the filter output as resonance increases. On this board, this is done by boosting the filter input as resonance increases (resonance feeds back the filter output but also the filter input itself), which causes a smoother resonance. Finally, tiny caps have been sprinkled in gain stages here and there to tame the high-ends above 20kHz, to remove some of the digital harshness. Sort of a SMR-4 done a bit better…

Schematics and PCB

You can find the Eagle files for this board in the shruthi/hardware_design/pcb directory of the source code hosted on github.

The schematics in PDF format are here.

Most of the SMR-4 analysis applies here. The expo current source lives somewhere in the chip , and it is not clear what the value of the resistor setting the output current is. The OTA-C stages are similar, with the exception that a (non-inverting) buffer internal to the chip is used instead of a bunch of op-amps in “transimpedance” configuration on the outside.

Bill of materials

In bold, the editors’ picks

Some notes:

  1. The lower the resistor, the more obvious it will be to notice the effect of the trimmer (but the harder it will be to do the adjustment).
  2. Op-amps with good offset performance are recommended for IC4. Op-amps with good audio characteristics are recommended for IC2 and IC6.

Releases

v0.1

  • Release candidate.