Exponential converters / CV question

[lorencz]

I'm really new to synth DIY, so I'm sorry if these are really basic silly questions, but anyway:

As far as I understand, basically all oscillators work by charging and discharging caps. The oscillator I've been playing around with is a basic square/tri generator based on a comparator/integrator pair. I've looked at a number of expo designs and want to experiment with some basic V/Oct control. I'm confused by how CV input affects cap charging. I thought that the amps in the CV signal would affect the charging rate of the caps, but then it occurred to me that one of the functions of the op amp CV summers that are in most OSC circuits is not only mixing inputs, but also acting as a buffer, isolating the current of the CV signal from the OSC circuit, so that the power of the CV input doesn't matter, only the voltage amplitude of the signal. Is this correct?

Secondly, if I'm working with a basic square/tri OSC based on an integrator/comparator pair, would the CV output from the expo be fed into the inverting input of the integrator?

For reference, this is the basic circuit I'm working with: http://www.circuitstoday.com/wp-content/uploads/2009/09/triangular-waveform-using-schmitt-trigger-and-integrator.jpg

[blue hell]

Welcome lorencz

I am not an expert, but could give some hints as to look into what direction.

To make the frequency change you would need to change the rate at which the capacitor of the integrator is charged / discharged. This would be the same as altering the input voltage for the integrator stage. so you would need to change the amplification of the circuit driving he integrator. Just inputting a voltage to the comparator stage would alter the DC set point but not its amplification, so that will not really work.

The first stage usually does a voltage to current conversion, where the current out depends exponentially on the voltage in.

The current is then used to charge the capacitor until the comparator finds it enough after which the current will either be reversed (for a tri-core oscillator) or the capacitor will be shorted (for a saw-core oscillator).

This way you will get a frequency depending on the input voltage, as a higher current will charg (or discharge) the capacitor faster.

The reason for doing this is that a transistor basically makes the conversion needed (except there are a couple of drawbacks that need be ironed out which complicates the circuit a bit).

Anyway ... hope this may help a bit, and maybe some expert here could chime in :-)

[lorencz]

Thank you for your reply, it was helpful, but i think i should clarify. Let's assume there are two CV controllers I'm working with. I'm putting out a 4v signal from both, but one puts out 20mA and the other puts out 15mA. Wouldn't the one with the higher current charge the cap faster, AND: doesn't an op amp CV summer serve as a means to make the current of the CV signal irrelevant, basically making the voltage the only variable signal passed on due to the fact that the op amp is running off the same power rails as the OSC?

Secondly, today I put together a few expo circuits in LTspice, connected them to an integrator/comparator pair, hoping i could figure out how this works. I had a bipolar 12v supply running the OSC, and the CV input was a unipolar positive voltage from a separate power supply. I fed the output from the expo into the inverting input of the integrator, but when i changed the voltage of the CV input signal, the Hz of the saw wave was not affected, so i feel like maybe I'm applying the signal to the wrong part of the circuit.

Anyway, hopefully that makes sense?

[elmegil]

OP amp buffers shouldn't care what the source can push, unless the source can't push whatever input impedance you've set up (most oscillators are around 100K). Op amps have very very low (ideal is 0) input current. I would agree that part of the function of the summing op amps is buffering, and part of it is simplicity in mixing disparate signals together.

I don't know much about software simulations, other than I've noted that the mac program "iCircuit" is flat wrong about some basic op amp circuits I've been trying to simulate lately.

[L´Andratté]

Hi lorencz!
In the circuit you posted you would theoretically make r3 a variable resistor and the control voltage change it´s resistance (what blue hell meant with changing the integrator drive). So that circuit has to be hugely modified for voltage control unless your using a vactrol in which case there is no exact v/oct.
Check out the ota thread, phobos posted a document that contains a nice beginner friendly triangle/square oscillator, made of an lm13700.

[gdavis]