[sdiy] VCO tuning philosophy re-visited

[Mike Pepper]

> >As an aside, there's another kind of design called a "charge pump" that I
> >think avoids the droop issue; instead of discharging the integrator cap
> >through a "resistor" (i.e. a FET or some other transistor acting as a
> >switch), it actually shoves current in the opposite direction. The Rhodes
> >Chroma oscillator does this. I think the ultrasonic VCO Ian Fritz uses in
> >his Double Decka (huh, it came up again) does something like that too
>
> Right. The DD uses an LM331 V-F converter in its core. This works on the
> charge balance principle, which does indeed involve a charge
> injector.  This chip has phenomenal 0.01% tracking, which is a consequence
> of the integrating current flowing over the whole period of a cycle (no
> error due to discharge time).
>
There was a design published in the UK Journal "Wireless World", (at one
time the most repected electronics publication in the UK), which showed a
function generator based on this principle too. I've not been able to find a
copy on-line, and I lost my originals in a house move 8 years ago.

The article was sometime in the 1970's, 1972 onwards.

It was a discrete transistor design, quite complex for the era. It has a
grounded capacitor fed by a current source, buffered by a FET. The FET
connected to two comparators, which in turn set and reset a bistable. The
bistable controlled some kind of current mirror in the charging circuit,
(this is the bit I can't recall properly). The core wave produced was a
triangle. It climbed  to the trigger voltage of one comparator, which then
set the bistable, switching on the current mirror, and discharging the
capacitor to the second comparator trigger voltage, where the bistable
reset, and the sequence continued.

The same designer began with a voltage controlled astable using bipolar
transistor cuurent sources, in a FET astable, claiming a 10^7:1 frequency
range. Tim Orr used that VCO as the driver for his VCO3, a distant ancestor
of Ian's 'Double Deka'; originally a 6 step sequencer, but with the note
that the range extended up to audio..

http://www.timewind.se/schema/Wwess/Fig17.png

The designer was probably D T Smith, the oscillator Tim used was certanily
his design, it's referenced in the synth article. If anyone of the academics
on here has access to a paper library with the bound volumes of 'wireless
world', then the orginal article is:

D. T. Smith, 'Multivibrators with Sevendecade  Range  in  Period",
Wireless World,  February l972, pp. 85-86.

Sadly, I can't find the more interesting function generator article.It may
be that the current mirror compromises the temperature stability, but I did
like the idea, since it removed the 'instant reset'
problem of integrator ramp type oscillators.

||\/||ike