[sdiy] more on which VCO

[Peter Grenader]

>> the VCO at midwest analog looks promising, but it lacks oscillator sync
>> and a sawtooth waveshaper (the basic waveform is triangle)

Sync can be added in a snap to just about any VCO - find the input to the
reset comparator and configure it as a summer, with a buffered external
signal as a second input.

<<like that oscillator sync is the same thing as FM>>

not twue!  

Think of a VCO for a second as an envelope generator. If you were to take
the output of that EG through a comparator that had both Q and QNOT outputs,
set the reference voltage to just a smidge above ground - what would happen?

The Q output would go high the moment the EG started it's cycle and the QNOT
out would go high the moment it completed it's cycle.

What would happen if you connected the QNOT output back into the EG's clock
input?

Then it would cycle, like a VCO.  That's basically what a VCO does.

In layman terms written by a true layman:

Simplified, inside a VCO there's a current loop with a gate in it.  If the
gate is closed, current flows.  If it's open, it's open - the loop has been
broken.  No current.  That loop is used to charge a capacitor and based on
it's design,  the cap will either charge up once (saw core) or up and down
again (triangle core) - once per cycle. .If you take the output of that
voltage from the cap and stick into a comparator, it can be configured to go
true (high) when the cycle is complete.  If you feed that output of that
comparator into the gate that's opening and closing the loop and set the
threshold of that comparator so that this happens at the right time in the
charge cycle, the whole mess then oscillates.

So, the output of the VCO core is used to trip a comparator. A sync pulse
acts as a second source into the same comparator, so it will start the loop
cycle again even if it hasn't completed the one it was on.

A VCO circuit also had the ability to vary the amount of current in the
loop,  which in turn determines the frequency of a single charge cycle.  You
can change that by forcing current into that loop. - and is where frequency
modulation FM occurs - through either a manually set voltage from a pot or
external VC input.

The circuit required to force that current is called an expo converter.  An
expo converter is such that variations in it's temperature will effect the
amount of current coming out and this is where stability comes into play.
The term 'stable' is general.  I do not know of a single analog VCO which
can be called truly stable.  What people usually mean when they say this its
instability is tolerable.   To make things as good as possible, as tolerable
as possible, designers use temp compensation (not elimination) to alter the
expo converter  n one of three  ways that I am aware of:  1) They seal off
critical components in the expo converter circuit from outside air flow so
it runs off its own heat only 2) they heat the expo converter intentionally
above room temp so that atmospheric changes have much less effect 3) They
'use' the outside air fluctuations to their benefit, using what's called a
'tempco' resistor that actually senses external temp changes and they stick
that tempco in a  critical point in either circuit that drives the expo
converter or the expo converter itself so it compensates for the changes
happening from the output of the expo converter.

Let's say your puppy is howling.  If you yell at him to be quite, he quiets
If you stop yelling, he howls louder again.  This is how a tempco works.  It
senses the same changes in air temp that's effecting the expo converter, and
adjust the drive accordingly. So, fluctuation isn't eliminated, it's just
sort of yelled at. 

Once that's all been worked out, the output of the VCO core, the output of
the cap basically is sent into a series of waveshapers which are used to
create the final output waveforms.  Usually each waveform requires it's own
waveshaper.  Don Buchla used one to create two.  The net effect was it made
a real good waveform for the shape the circuit was designed to produce
(sine) and a rather crappy waveform of the one it was morphed into by
circuit bending:  either a saw or square. The morphs sounded OK, but a
purist would freak as on a scope they look like crap when compared against
what they a supposed to replicate.  The saw looks like a upwardly wiggly
pigtail and the square is a bit rounded as the corners.

I'm OK with the way he did this.  A not so good saw or square is sonically
much more acceptable than a dirty sine, which need to be pure.

- hope this didn't confuse you all as much as it confused me!

- P