Author 
Circuit to control pitch relationship between multiple oscs 
trip 
I've been trying to come up with am analogue way to distribute cv to multiple oscillators, where a division of the frequency of the master oscillator can be selected for each other oscillator. So, say you have a master oscillator at 400hz, with a rotary switch you can set the other oscillators to be 1/2, 1/3, 1/4 etc. of that frequency.
I've yet to come up with a way of doing this that works satisfactorily, and wondered if anyone had any ideas?
The difficulty is that you want the oscillator to be able to accept bipolar voltages to control the pitch, which means that you can't perform a simple division of the voltage going into the exponential converter of the oscillators. Offsetting the voltage to a positive only range and then back again will introduce too much error.
I wonder is it worth trying to offset the initial pitch of the oscillators so that at 0v they oscillate at say 40hz, and then bias the incoming cv to be positive before dividing it for the subsequent oscillators. Will this interfere with the temperature stability?
Is there a simpler way I'm missing?
Cheers. 

elektrouwe 
trip wrote:  you can't perform a simple division of the voltage going into the exponential converter of the oscillators 
expo. convertion means: adding voltages @ input side is a multiplication @ output (here frequency), subtraction of a voltage leads to freq. division.
dividing by 2 ( = 1 octave down) is just subtracting 1V ( if you have a 1V/octave expo scaling) @ CV input of this 1/2 sub VCO. For 1/4 freq subtract 2V ...
offset errors mean div. ratio errors, so this method may need syncing the oscillators. Personally I would rather use a freq. divider to get hard synced, exact div. ratios 

trip 
elektrouwe wrote: 
expo. convertion means: adding voltages @ input side is a multiplication @ output (here frequency), subtraction of a voltage leads to freq. division.

Of course, you're correct thanks. A series of offset voltages that are subtracted at the expo converter could work quite well for my purposes. It will not be perfect, but fine tune can help with that. I can hook up some switches to optionally sync the oscillators as well (they are saw core).
Frequency dividers would be nice too, but I'm trying to work around what I already have 

Idunno 
The ADDAC VC Transitions module allows offset scaling of five outputs relative to a slider pot's position. I have no idea if this is helpful for what you want to do though. 

basicbasic 
I was after something exactly like this a while back. I found a PCB by Diablojoy that fits the bill ( http://www.electromusic.com/forum/viewtopic.php?t=37689 ) but i've ended up using Ornament & Crime in Harrington mode which kind of does what I need and has lots of CV options as well. 

EATyourGUITAR 
the clock divider outputs can be fed into individual 24dB low pass filters with 1v/oct tracking to produce almost sine waves at consistent amplitudes maybe. it depends on the accuracy of your equipment. you feed the oscillator and the low pass filters the same 1v/oct pitch cv. you need to adjust the offset (cutoff) at each filter to get the wave to be mostly sine without loosing all your amplitude. you will probably need a recovery stage after the filters to bring back the amplitude of your sines to something that is more ideal for the headroom of your synthesizer. this is the only way to eliminate wandering phase of each division. the good news is that you can control phase 180 by phase flipping the square outputs of the clock divider. precise arbitrary individual phase adjustments requires special filters, special clock dividers or both. research scientists have been doing this in software for years because it is easy to program and you have total control over everything. you can also log the results and use any desired precision. the downside is that it is not analog therefor it suffers from aliasing, DAC error, and nyquist band limiting. 
