GS01 the CGS sub-oscillator/harmonic sequencer module was one of the first synthesizer modules Ken Stone ever built, back around 1980. The first version was designed with the ETI 4600 in mind, and was published as a circuit idea in the June 1982 Australian edition of ETI.
Another circuit I used at the time to "fatten up" the sound of my single oscillator synth was a "harmony generator", achieved by running a 4017 decade counter chip wired to divide by six or divide by three.
This design combines both of these circuits, giving a two channel sub oscillator, which allows each channel to be used independently, or driven from the same oscillator, but set to different intervals. As a bonus, both channels can be multiplied or "digitally ring modulated" giving even more effects.
Some ideas on how to use this module
Fed by two VCOs it can operate as two independent sub-oscillators, with optional ring modulated outputs.
Running both sub-oscillators from the same VCO, tuned to a fifth over the base note you require, and with the prescaler set to divide by 3, it is possible to create harmonies. An unusual effect here is that the ring modulated outputs give a fatter chord-like sound that remains indepenant of minor or major scales, allowing "one finger chords" which can be handy when used with a related sequencer driven bass line.
Careful mixing of the ring modulated output with one channel's divided output results in some interesting sounds, especially if one of the channels is being driven from a low frequency oscillator. Running like this, it could be considered to be a "harmonic sequencer"
Running both inputs from the same LFO/clock signal, and feeding the mixed output into the 1V/octave input of a VCO gives you a pattern based sequencer, where the prescaler, ring modulated outputs, etc., all have a bearing on the pattern. Quite interesting patterns can be created this way. Each knob of course varies part of the overall structure, instead of a single note as per a regular step sequencer.
- When using the sub oscillator, try feeding an audio signal into the "a" input, and a low frequency signal into the "b" input, then mix the "a" channel and multiplied signals together. Due to the action of the EXOR gates, the output of the "a" channel can be gated partially to fully on and off, giving some interesting tonal sequences.
- Asymmetrical input waveshapes will not drive the input multiplier properly. Nothing says you can't exploit this either.
- If you wish to mix the frequency doubled outputs (x2 and x2i) into your output, you will need to increase the values of the 330pf capacitors to perhaps 1500 pF, or even higher so the pulse can be heard. At 330 pF, it's just too narrow for anything but internal use. Note that enlarging this capacitor also lowers the maximum frequency at which the sub-oscillator will work. From memory, I'm using 1500 pF capacitors.
- Don't tie all the wires going to the pots together. This causes some cross-talk. Either keep the wires apart, or use shielded cable.
Having had the prototype in my synthesizer for some time now, a second layout became obvious. Instead of having pots for both the B channel and the multiplier outputs, I used a 4 pole 2 position switch to select between the B channel and the multiplier outputs, saving me a lot of panel space. The x2 output was sacrificed, and the rotary switch to select the prescale was replaced by an SPDT switch for the bypass function, and a SPDT center off switch was used to select between divide by 3, 7 and 5. 7 is hard wired as the default on the PCB, so no connection is needed to the switch.
A little on how it works
In a way, this design can be considered as a sub oscillator construction kit, as there are many ways it can be configured.]]
The circuit can best be understood if viewed in smaller sections, which is why I've broken the circuit diagram up into parts.
The first part simply takes the input signal (preferably symmetrical, such as a square, triangle or sine wave), squares it up, and multiplies it by two using a fairly conventional square wave frequency doubler made from a few NAND gates. Both normal and inverted signals are brought out to PCB pads, and the normal signal is fed on to the prescaler or dividers. There are two of these input stages on the PCB, one per channel.
The second diagram covers the prescaler, giving effective frequency divisions of 1, 1.5, 2.5 and 3.5 when the frequency doubler in the input stage is taken into account. If a double pole switch is used as shown, it is possible to route the frequency doubled signal straight through to the divider as well. Of note is that the signal is being fed into the clock enable line, while the clock is held high. The clock enable line is simply inverted then AND gated with the signal from the clock pin within the 4017. Grounding the clock enable line and feeding the clock signal into the clock pin would have worked equally well.
A link is shown on the PCB overlay between points B and C. This link is only installed if you want the prescaler active all of the time. In this case a 1 pole rotary switch can be used, and all wiring between A B and C on the diagram can be ignored, except for the installation of the link shown on the PCB.
The third diagram is the divider, of which there are two. Dividing by two, four, eight etc. gives the various octaves.
The fourth diagram simply shows how the multipliers are wired to the two dividers. The outputs are marked M1-4 and can be fed into the mixer as per the example in the mixer diagram, much like any of the signals from the dividers. These four EXOR gates do the digital "ring modulation", with each gate wired so that it is feeding from the same numbered divider stage from each channel.
The fifth diagram shows the mixer. It also gives an example of how the inputs are fed through a resister then a pot and into the mixer. My prototype used thirteen pots, one more than was provided for on the PCB. There is no hard and fast rule to what is fed into the mixer. That much is up to the individual's needs. In my case channel "a" was the divider chain without the prescaler, and the one with the prescaler was channel "b". From channel "a" I fed the squared input, /2, /4 and /8 signals, being the pads marked 1 to 4 on the divider into the mixer. From channel "b" I fed the frequency doubled input (see note below) the squared input, /2, /4 and /8 signals into the mixer. I also fed all four of the multiplier outputs (M1-4) into the mixer.
A few resistors can be omitted from the mixer, if desired. All of the upper parts of the input voltage dividers (47k marked in red) can be left out, if the resistor in the feedback loop of the first op amp is reduced to 22k (marked in red). These voltage dividers were initially included to allow voltage followers to be hooked in at this point, to bring the individual frequencies out to their own jacks.
This is a guide only. Parts needed will vary with individual constructor's needs.
|LF353 or TL072||2|
|LM358 or TL072||1|
|Tinned copper wire||10 cm|
|50k Lin pots||13|
|2 pole 5 pos switch||1|
|* See text|
Readers are permitted to construct these circuits for their own personal use only. Ken Stone retains all rights to his work.
- CGS sub-oscillator/harmonic sequencer
- The CGS modules
- CGS sub-oscillator/harmonic sequencer (UNI format)
- CGS parts FAQ
- CGS DC mixer
- Sub Oscillator for music synthesizers (archived) by Ken Stone, 1999, with permission of the author
- CGS Synth discussion group, for discussion of locating parts, modifications and corrections etc.