[sdiy] SAD1024 different modes

[Mike I.]

JH,
The way Eventide does it in the plugin delay card for the Instant Flanger
seems like a good solution. Each 512-stage section is differentially-driven,
the two sections are operated in parallel-multiplex, and the n, n-1 outputs
within each section  are resistively summed and fed to the inputs of a
single diff amp at the output. Two of these blocks are put in series to
provide the main and aux outputs.

I think this is the main advantage of parallel multiplex -  for a given
clock rate the resulting sample rate is 2 x clock.  If many sections are
operated in a "multiphase multiplexed" arrangement the ratio of S.R. to
clock becomes even larger. What this does is allow you to extend the the max
delay time without changing the minimum delay time... increasing the max/min
ratio. Shortest delay being limited by how fast the clock can go and how
much insertion loss variation, bias change etc,  one is willing to tolerate,
the long  delays are limited by increase in noise, aliasing and how well the
output filter can attenuate the clock signal if it drops into the audio
range. With several sections in multiplex, a relatively low clock rate can
be used to obtain long delay times while  keeping the sample rate
reasonably high to maintain good performance.

The Marshall Time Modulator probably  took full advantage of  differential
operation and n x parallel multiplexing.
Regards, Mike