AW: Don's interpolating scanner

[Magnus Danielson]

>>>>> "j" == jhaible  <jhaible at primus-online.de> writes:

 >> I've *finally* finished the article I was writing about my
 >> interpolating scanner design.  Enjoy...
 >> 
 >> http://www.till.com/articles/scanner/index.html
 >> 
 >> (It was heavily inspired by Juergen's work and benefits from
 >> at least a couple of his suggestions.)
 >> 
 >> The basic idea is to build an N-way diff amp circuit to drive the
 >> current-controlled amps, taking advantage of the symmetrical
 >> exponential transfer characteristic for a smoother transition between
 >> the inputs.  It also has a pretty low parts count.
 >> 
 >> -- Don

 j> Congrats to Don for further improoving the Scanner idea !
 j> Now I hope you will soon come up with a ready circuit for the full
 j> Hammond Chorus emulation ! (for the list: Don gave me valuable
 j> hints to simplify a general gyrator circuit for this application, but
 j> I never came to really try it.)

I agree, certainly an interesting piece of work!

Now, this rings a bell. I have also been pondering over the Hammond
A-100 schematics and the scanner chorus.

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

First I would have two 8-to-1 analog muxes (4051's) to select two
adjacent inputs, each being feed into the 2 VCAs. The VCAs are being
feed with linear CVs being directly invertion of the other (other
relations may be selected, I take this for simplification). When one
of the VCAs is in the bottom end, that VCAs channel is switched to the
other adjacent channel (say, if we just faded from 1 to 2, the
outfaded ch. 1 will now be replaced with a outfaded ch. 3). The
outputs of these VCAs is summed just as you would expect. You would
need to care some for the exact handling of scanner CV input to VCA
control as well as the machinery that selects channels. This becomes
apparent if you consider that a CV may change direction (the channel
selection must not be upset) or that you make a infinitely
increasing scan (for which a rising CV is eventually not healty and
you got a few kV arcs happening every now and then).

This is a budget cursuit, replacing many VCAs with pretty cheap CMOS
switches and extra control. However, the benefit with this approach is
that it scales better, it does not become very much more expensive as
you go up to 16 or 32 channels.

To make a infinitely increasing scan one would not be interested in
having the scan being linear to the CV, but the rise (or falltime) of
the scan to be linear with the CV, thus, it represents the scanning
frequency.

Basically this thing would remind one of a barber pole thing.

Anyway, with such a scanner you could redo one of the tricks that the
Hammond scanner does, frequency shifting the tone. This, shifted tone
is then added to create the beating effect. Designing a AP-filter
network that acheives the delays should be manegable. A gyrator or
leapfrog network is not a bad idea.

 j> Don's circuit is very elegant - can't await the full Hammond Vibrato 
 j> version !

That would be interesting indeed.

Cheers,
Magnus