[sdiy] Linear Analog Synth Portamento

[David G Dixon]

> What I hear from my Fatman sounds natural enough to me.
> 
> So what am I missing?  Is "linear portamento" bad?  What 
> won't "linear portamento" do that I need or want?

A pot-and-cap gives exponential portamento, which is actually constant-time
portamento (because the cap-charging current which defines the glide time is
the integral of the voltage-time curve, and this integral is proportional to
the initial voltage difference for an exponential decay function), which is
the normal kind of portamento (because it's so damn easy to do).  In this
case, the initial glide rate is very fast, but the final approach to the
destination is asymptotic (i.e., much slower).

With linear portamento, the glide rate is constant (because it is determined
by the rail voltage of a comparator), but the glide time is proportional to
the initial voltage difference.  It gives a different sound, and is perhaps
less useful for keyboards and sequencers, because the time of arrival is not
uniform -- i.e., of the glide time is to large, some notes will not reach
their destination voltages before the next note is played.

It is possible to have constant-time linear portamento, but (AFAIK) this
requires storing the voltage difference and using it to program a
variable-current source to the Miller integrator.  This would be a very cool
effect, and it is something I'd like to try to build in the near future
(probably with linearized 2164s).  (Consider: three VCOs in tune playing
random 3-note chords, and each time the chord changes, all three notes snap
(not in an asymptotic manner, but abruptly) to their destinations at exactly
the same time, no matter how big their individual intervallic jumps are.
That would be awesome!)  If someone here is sitting on such a circuit, I'd
love to see the schematic!