AW: Don's interpolating scanner

[Magnus Danielson]

>>>>> "DT" == Don Tillman <don at till.com> writes:

Hi!

 DT>    From: Magnus Danielson <magnus at analogue.org>
 DT>    Date: Sun, 03 Jan 1999 04:21:57 +0100

 DT>    I agree, certainly an interesting piece of work!

 DT> Thanks!

Anytime!

 DT>    I have a little diffrent solution that does not require as many
 DT>    VCAs. In it's basic layout it only requires 2 VCAs. Let's consider a
 DT>    8-input scanner.

 DT>    First I would have two 8-to-1 analog muxes (4051's) to select two
 DT>    adjacent inputs, each being feed into the 2 VCAs. 

 DT> Yeah, this will work fine.  The big issue here is converting from an
 DT> input control voltage to the digital select lines for the mux and the
 DT> analog control voltages for the VCAs, and the circuitry to do this
 DT> accurately and without glitches is not easy.

Sure. But once you have reduced the audio path you can do this.

 DT> But that's for a *linear* scanner, like the circuits JH and I did.

Right.

 DT> A *circular* scanner is where your approach becomes powerful.  A
 DT> circular scanner doesn't care about a scan control voltage so there's
 DT> no need to deal with that; a triangle wave oscillator and a counter
 DT> are all you need.  

Right again.

 DT> A speed control voltage would change the frequency of the triangle
 DT> wave oscillator, and, yeah, through-zero is just begging, isn't it?

Sure is.

 DT> The original Hammond Chorus uses, clearly, a circular scanner.

It does. Hooked to the synchronous motor shaft, you switch depth (and
those shift amount) by selecting how the scanner is hooked to the
AP-delay line.

Now, what I have been thinking lately is how one could combine the
linear and rotating scanners, so you have a linear (phi) and rotation
speed (omega) modulation input to a common scanning core.

To do this you would have to first transform the rotation speed
(omega) into a wrapped ramp (covering the full range), then add the
linear modulation (phi) and put this through a wrapper again. I
suppose that one could do this in the analog domain, but I really
think it is better done in the digital domain (like a PIC or so).
Could be done in analog as well, but it becomes a bit trickier.

Cheers,
Magnus