TUBES:
These are scans of an old book on vacuum tube music circuits
Judd pt1 (3MB)

Judd pt2 (8MB)

SERGE
This manual used be online, but the page is long gone, so here 'tis again: 1976 Serge User's Manual (no schematics)



SYNTH 12/2009
frequency shift register
This is still a draft version and needs a lot of refining, but the basic idea works well.

The Analog Shift Register is well known, where voltages are shifted over three stages. This is the same in concept except it shifts frequencies and uses 4046 PLLs to do it.
Which means you have to be very optimistic to expect to see the frequency you fed into stage 1 to appear at stage 4.
For a single audio input and a clock signal, it produces 16 audio (or CV/gate/trigger) outputs of related frequencies.

March 2010: Here is an UNTESTED draft of the full schematic. The circuit has been breadboarded with two 4046 stages and certainly put on a performance.
The clock ticks over the 4017 decade counter which enables the 4046 PLLs to shift their frequencies to the next stage.
first clock shifts PLL3 to PLL4, 2nd is PLL2 to PLL3, 3rd is PLL1 to PLL2 and 4th allows a new frequency to be introduced to PLL1.
Or you can feedback one of the outputs from PLL4 and get a loop going....oh yes:-)
But don't expect any kind of perfect tracking, these are 4046s. Your loops will drift horribly at best or just generally misbehave.
The frequencies are shifted by connecting the comparator outputs to pin 9 of the 4046s. The 4017 counter allows the 4066 switches to close in turn. The frequency from the comparator is smoothed to a voltage by the 10k resistor and 0.1u cap.
Once the 4066 switch opens, the voltage in the cap is held thanks to the hi-Z input on pin 9.
This voltage on pin 9 determines the frequency of that PLL's VCO.


Bindubba 2 pdf [vers.1](385kB)

have a listen!





1/2010 CHAOS-FILTER

This circuit is based on one published by Sprott (http://sprott.physics.wisc.edu/pubs/paper245.htm). Added some pots, output buffers and changed some component values to get it in line with typical synth standards.
Also can switch between nonlinear and linear feedback paths. The idea was to allow the circuit to double as a filter when linear feedback is selected. Actually it is a pretty poor filter, but a pretty good chaos generator. It can be interesting feeding it a CV causing the output to move in and out of chaotic regions.
Tried to get it to oscillate as slowly as possible, hence the 1uF caps. Chaos sounds like noise at audio rates, so is much better to use this module as a contol voltage generator. Ian Fritz gave a lot of help with this - thanks!! Sorry if the component values are a bit hard to read, should be okay. When I upgraded my design software, it deleted all my old files, so there was only this low-rez pdf left.
Chaos-Filter pdf (250kB)
Here's a vid of the circuit being tested:




Feb2010 BINDUBBA 1 SEQUENCER


This circuit is a crossbreed of the concept of an Analog Shift Register and a CGS Sequencential Switch.
Basically it is a 4 stage sequencer. If the pots are labelled ABCD, the outputs will emit voltages in the patterns of ABCD, BCDA, CDAB, DABC and in the reverse direction; DCBA, CBAD, BADC, ADCB.
This means you can have a 4 stage ditty driving one VCO and the exact reverse of that ditty running another, or the reverse staggered by one clock pulse...or two clock pulses....etc.
Just to suck it dry, there are inverse outputs available too, and I couldn't let the extra stages of the 4024 chip be idle, they are connected to a R2R ladder to give a staircase output (and its inverse).
Untested as yet, will get PCBs made sooner or later.
In the schematic, P1 connects to the wipers of four 100k pots wired as voltage dividers. These could also go to jacks to route CV signals, but extra components would be required to ensure no negative voltages are introduced to the 4052 chips.

Some of the function blocks, such as the clock and reset subcircuits, output buffers and R2R section, are copied straight from CGS circuits. Most of my modules are CGS so it makes sense to design this module to conform to those standards.

Function Block description:
A 4024 ripple counter is driven by a clock signal obtained from a comparator sub-circuit and can be reset by a pulse to the appropriate jack. The clock signal can be any rising edge crossing approx 1.5V.
Q1 and Q2 from the 4024 are connected to the control inputs of four 4052 dual 4 channel mux/demux chips. These all switch the four inputs to the output in the order of X1, X2, X3, X4 for each section.
Four pots wired as voltage dividers provide the set voltages to the inputs of the 4052s. The way each section is connected determines the start point and direction of each 4 step sequence.
Each 4052 output is connected to a voltage buffer sub-circuit which supplies the output sequence and its inverse.
As mentioned above, the extra outputs of the 4024 are connected to a R2R ladder and then a similar output buffer sub-circuit to provide a staircase output and (of course!) its inverse.

Bindubba 1 pdf [vers.2](1050kB)

Bindubba 1 build notes 3/2010 [vers.1](130kB)