2 VCO's - 1 Log convertor (polysynth design)

[jh]

>The Trident had two VCO's per voice, and the PolySix had one.
>
>I'm pretty sure the PolySix schematics are somewhere out there on the web, 
>but I don't know where.  You might try Tony Clark's DIY schematic archive.
>
>Juergen Haible could probably explain this circuit better than I - maybe 
>he'll add some comments to clarify.
>
>Michael Bacich

Here I am - and still learning from these brilliant Korg circuits myself.

I suggest a combination of the Trident/Polysix mux'ed expo converter,
and the MS-10/20/50 linear VC current source with additional expo input.

1st step: Get the 0 ...5V of yout typical keyboard voltages converted to
V/Hz with one precise multiplexed expo converter. The Trident had
a 726 (oh well), and the Polysix used a vactrol instead of the usual
tempco resistor, and a 7th mux channel with reference voltages to
change the LDR resistance. Permanenent autotuning in the background,
but still the slight imperfections of free running linear VCOs - imagine
that ! I'd say we go for a LM394 + tempco or a heated array for the
PolyModular.

2nd step: Get the V/Hz CV to control a linear VCO. As all (?) VCOs
have a current controlled oscillator core, you need a precise linear
voltage controlled current source. The Polysix has a rather simple one,
and you loose the wonderful expo modulation input of V/Oct VCOs.
Not good. But the MS-20 shows the way. The VC current source 
uses a (cheap) differential pair, which allows for individual tuning of
each VCO, and any exponential (!) modulation you want. You just
don't invest in the precision that would be required for multi octave
tracking here.
The footage selection is done by switching resistors, i.e. without
interfering with expo converters, nor opamp offset voltages.

At the moment I try to do a slightly differnt thing: build a Korg 700
or Korg 800DV - style VCO. Only run the saw at the highest footage
and use a CD4024 to make the lower octaves. (Adding a portion
of the top octave saw to the staircase produces a nice saw wave for
all octaves. There *are* some glitches, but I wantto find out if these
contribute to the "Early Korg" sound.)

I suspect that these 3-transistor (thyristor) VCOs are excellent
for operating on the high end of the range. Very fast discharge
of the cap, and no delay from opamp buffers, comparators, monoflops
or similar things.

JH.