CV controlled Filter Response

I had some fun trying to get the Filter Response column on the matrix working correctly. It is a modification to the original circuit, and according to various sources it should be as simple as connecting up the matrix column with a resistor to the relevant part of the circuit. (Hinton, Synthi Modifications, Phutney (item 6 of Board Y).

Filter Response schematic

Filter Response schematic

Despite having tried the various suggestions, I couldn’t get it working correctly, as no matter what value resistors I tried, it affected the response control – either preventing the filter from resonating at all or too much. Derek Revell’s suggestion of using a 10K resistor and a diode didn’t work for me either.

The resonance control sends a 0V to -9V voltage to board B – where it is halved by a voltage divider made by R89 and R90 so the gate of Q26 gets a voltage between 0 and -4.5V.  When I measured it using the DMM, when the voltage got above about 2.9V the filter started oscillating relating to, I think, the pinch-off voltage of Q26. I am using a selected BF245B JFET here.

On photos of later revision EMS B boards, there is a 5V1 Zener diode in this area – though it doesn’t appear on the available schematics for the filter. I think it is there to clamp the voltage to about -5V to protect the Q26 FET.

In the end I decided to move the problem off the board, and to use an opamp summing mixer to mix the voltage from the pot and the CV from the matrix. This then presents the correct voltage, a sum of half the voltage of the pot, and the full voltage from the matrix.  R89 and R90 are removed, as the voltage divider isn’t required – the output of the opamp is applied to the circuit as pointed to in red above.

The circuit I used is this – modelled in LTSpice:

LTSpice drawing of the response control mixer (click to see it larger)

LTSpice drawing of the response control mixer (click to see it larger)

This is a pretty standard summing mixer – with one input connected to the Filter Response input on the matrix, and the other to the Response control.  Half the latter’s voltage contributes to the sum due to the 200K resistor. On the output two diodes clamp the voltage. A 5.1V (or 4.7V in the LTSpice sim) Zener diode prevents the combined voltage going over -5V or so, and a Schottky diode prevents it going much above 0V (0.2V or 0.4V depending on the diode).

It’s implemented on a small square of stripboard, using a TL072 opamp – tapping power from some of the link wires on the board, and the CV input wires soldered where the voltages enter the circuit. The output is soldered in the hole for R89 which connects to the gate of the JFET.

An unsightly hack - but it works. A TL072 summer circuit bodged onto board B.

An unsightly hack – but it works. A TL072 summer circuit bodged onto board B.

All a bit hacky at the moment, but it will eventually be added to the ‘interface’ board where I already have a TL074 opamp for the other mods. In the photo above the wire colours are: Red: +12V; Blue: -9V; Black: GND; Orange: Response Control; Purple: Response CV from Matrix; White: Summed output. There are a couple of other resistors on the control input – as I originally put in the voltage divider here before realising I could simply use a larger resistor on the opamp input.

It works – allowing you to use the Response control as a bias, then a +/- control voltage like the Trapezoid, Oscillator, or even Joystick can be used to modulate the filter’s  resonance.

An example patch

Here’s a patch using the mod. Oscillators 2 and 3 are set on low frequency. Osc 2 controls the filter response and frequency of osc 3. Osc 3 controls the filter frequency, and output channels. Noise is patched to the filter.

A Patch using the response mod

A Patch using the response mod