[sdiy] BBD Chips

[Jaroslaw Ziembicki]

----- Original Message -----
From: "Scott Stites" <scottnoanh at peoplepc.com>
To: <synth-diy at dropmix.xs4all.nl>
Sent: Tuesday, July 22, 2003 5:51 PM
Subject: Re: Re: [sdiy] BBD Chips
> But wait - don't ensemble choruses by nature have the BBD clocks
'crossing'
> themselves in the frequency domain?  Now I'm confused.....

Yes, the clocks are independent, but there's no danger that any new
frequencies in the
audio range would arise.
It's due to a fact that the output signals of all BBD's are mixed together
in a *linear*
summing circuit. That means, no additional sums/differences of output
frequencies can
arise. The signal after mixing is a sum of "useful" low frequencies, and
high frequencies
which result from the sampling. After low-pass filtering, only the "good"
low frequencies
will be left at the output.
Of course the input signal or signals must be well low-pass filtered before
passing in the
BBD (or BBDs) in order to cut off any frequency that is above the half of
the BBD
clock frequency. It's the well-known Shannon theorem - also called the
Kotielnikov
theorem in some countries ;o)

>
> Subject: Re: Re: [sdiy] BBD Chips
> From: Roman <modular at go2.pl>
> Date: Tue, 22 Jul 2003 17:03:06 +0100
> To: synth-diy at dropmix.xs4all.nl
>
> I know at least one person who'd hate me for saying this, but...
> What if we use 2 BBDs and 2 clocks. When one frequency rises, second
drops. When
> frequencies are equal, there's no delay between BBD outputs, so flanging
can
> begin at zero delay, or go back and forth into positive and negative
delay. As
> long as there's no original signal added, and everything passes thru BBDs,
> there's no minimum time limit.
> Only twice the noise...
>
> Roman
>

Roman, you are a man of genius. In this manner, you can really get very,
very short
delays without "overloading" the BBD (the danger at too high clock
frequencies is that
the electrical charges that represent the samples may be lost).
And there is another advantage of having two BBDs in one device: you can get
a typical "string ensemble" chorus effect, or you can reconfigure the
circuit and connect
both BBDs in a series (tap delay?).
In the second case, if the BBD clocks are independent: a very good low-pass
filtering
between the BBDs is necessary (Shannon/Kotielnikov)!

Regards to all
Jarek