From obxwindsurf@yahoo.com Mon Oct 06 06:28:02 2003
Subject:Re: Next Generation Vibrato/Chorus Clone / Geeking again...

Steffan,

I'd thought of this and in fact if I was going down this road I'd
probably just buy a good phase shifter and create a summing amp to
add the dry to the phase shifted signal. I don't think, however that
the continuously variable change would achieve the same effect.

The digital controlled phase shifter will literally jump to 1 of 16
possible phase shifts over the range determined by the resistors in a
discrete manner similar to the discrete phase shift jumps of the
scanner.

The phase shift stages in my proposed design are in fact all-pass
phase shifters. The delay line of the scanner, while comprised of
cascaded second-order low pass filters I assume that the cutoff
frequencies are selected far enough out so that considering the upper
register frequencies and harmonics of the TWG that it behaves as an
all pass to those, otherwise the sound would just be muffled.

What all LPFs exhibit, however, regardless of cutoff frequency is a
frequency dependent phase shift. Since they did not have voltage
control in the days of vacuum tube designs a phase shift network with
fixed delays and a means of scanning up and down the line was
needed. The drawback was that the means for doing this (the scanner)
was not a smooth transition between one phase angle and another, but
rather a choppier change in phase angle when heard by the ear.

This is what gives the Hammond that "excited, nervous, animated" tone
using chorus.

While the same effect can be achieved using cascaded 2nd order LPFs
and a pseudo scanner (or even the real thing), analyzing functionally
what is happening and then selecting a design that performs a similar
function can probably realize the design in less components.

My set of 3 cascaded all pass phase shifters, is almost what you
describe. Although the digital control gives it a choppiness that
will probably produce the desired result. It's promising that Dave
Amels has already achieved this with his design, although his
resistors to achieve the phase shift had more steps (256 vs 16).

I'm thinking that by keeping the 4 binary-weighted resistors external
to the switching source (DG201 quad analog switches), and decoupling
the scan control, and switching chips from the analog supply using LC
filters, I should be able to keep the feedthrough digital noise to a
minimum.

Regards,
Kevin