That commutator idea of Andy LaTorre's sounds like fun.  Its just my
kind of mix of mechanical and electronic wizardry.  It eliminates the
need to match photodetectors and eliminates the need to have a lot of
channels or electronic multiplexing.  I'm adding it to the top of my
list of possible reader\scanner designs. I can think of a couple of
additional improvements.

Instead of optic fibers being placed at hole centers, use a continuous
line of them going from the tracker bar to the ring.  This way all formats
could be covered with one setup.  The read positions could be selected by
stepper motor control.  Tracking could also be done with the same unit by
re-defining the hole positions based on the detected paper edges.  The
note sheet could just run off the spool without regard to tracking it
mechanically. The interface and the software should be quite easy.

Idea #2:  Forget the L-shaped fiber optic and mount the detector on the
spinning shaft.  Use slip rings and maybe mount the preamplifier on the
shaft too, to increase signal-to-noise ratio.

Idea #3:  Use a fixed speed motor and an encoder or some markers and a
clock.  Does anyone have any ideas about how slow this method would be?

Spencer Chase

 [ Idea #3 would be _very_ fast, probably 10,000 rpm is possible.
 [ No special markers are needed, just have a pattern of alternating
 [ on/off to identify the margin at one side.  (A little paint on the
 [ fibers at the "tracker bar"!)
 [ The remaining electronics needed closely resembles the hardware
 [ which supports a CCD array and sends the data into the computer.
 [ -- Robbie