[sdiy] 20ms delay? analog allpass filter?

[David G Dixon]

Fair enough, but it's fairly straightforward to achieve scanning crossfade
with analog.  I've got a very nice circuit for a noiseless 8-channel
interpolating scanner that uses 5 TL074, 1 2164, 1 LM393, 1 LM339, 2 4077,
and 2 DG444.  (It can also be built for 4 channels, with about half the
parts, or 16 channels with about twice the parts.)  Most of the opamps, the
comparators, and the XNOR gates are configured into a "folding ADC" (look it
up on Wikipedia) which converts a -5V to +5V control voltage into 8 digital
bits which drive the 444 SPDT switches.  The folding ADC frequency
multiplier drives the control of a linearized 2164 crossfader.  One
crossfader suffices for all of the channels because only two channels are
ever being expressed at the output.  Based on a clever switching
arrangement, there is absolutely zero switching noise in the scanner.  It
seems like a lot of parts, but it's probably a helluva lot easier than
working a physical rotating thing into an electronic circuit.

The only issue is that the opamps in the ADC circuit can scream like
banshees if they aren't properly stabilized (use 10k resistors instead of
100k, and provide for 100pF caps across all the feedback resistors).


-----Original Message-----
From: René Schmitz [mailto:synth at schmitzbits.de] 
Sent: Wednesday, February 08, 2023 4:58 AM
To: David G Dixon; synth-diy at synth-diy.org
Subject: Re: [sdiy] 20ms delay? analog allpass filter?

[CAUTION: Non-UBC Email]

HI David,

On 08.02.2023 00:12, David G Dixon wrote:
> That's cool, but why not use a 4-bit counter and a 16-channel digital
> switch?
>
Because that would be hard switching, while the scanner does a crossfade

(plus some high pass action) as the rotor plates move over the stator 
plates.

The whole thing is a complex variable capacitor.


Best,

  René


--
synth at schmitzbits.de
http://schmitzbits.de