Spencer Chase started a thread on "Music Roll Perforator Control
Program" followed by comments from John Farmer.  Many of the items
that were discussed by Chase and Farmer are present in my computer
controlled punching system that I developed in 1978 and am still
using today.

My punch consists of a punch head that traverses the paper in .001"
increments using a stepper motor.  The total travel is 17 inches.
The paper is advanced in .001" increments by another stepper motor.
Rotary cutters are attached to the paper dive shafts so that any width
paper can be cut after the punching is finished.  The punch traverses
one row of punches and then the paper is advanced so that holes can
be punched on the reverse travel.  The stepping motors are ramped in
stepping rate to maximize movement speed.

Since I punch for many different instruments, the punch head is easily
changed to different sizes and shapes.  The one item that I don't have
on the machine is multiple punch heads, as Spencer would like to employ.

I employ pneumatics to drive the punch through the paper.  Initially
I employed a single direction cylinder that incorporated an internal
spring for punch return.  I soon found that the punch could fail to
return due a hanging chad.  If the program continued, the paper was
destroyed.  I then added sensors to indicate when the punch had
returned before continuing to move the punch or advance the paper.
The computer would signal a paper jam and wait for operator
intervention.  This required considerable operator intervention.
I decided the best approach was to dive the punch in both directions.
I no longer had to sense when the punch had returned, because it
always did.

When I started the project, the PC did not exist.  I started with
a Cromemco computer that I built from a kit.  Because of the limited
memory, I devised a system for storing punch files in bit patterns.
Thus a 14-note Organette would require two bytes to store the data
for one row of perforations.  A 20-note machine required 3 bytes.
The initial programs were written in machine language, but later
converted to Pascal when a PC replaced the Cromemco.

Roll scanners were years away and the appropriate technology did not
exist.  I built photoelectric sensors arrays for the different note
spacing (2/in., 3/in., 4/in., 5/in., 6/in. and 7/in.).  The sensors
were mounted on the punch table and the paper advance was used to move
the paper over the sensors.  A DC-powered Lumiline lamp was used as
the light source.  Computer programs were developed to allow editing
of the scanned rolls, and my son developed programs to convert my file
format to MIDI and MIDI files to my punch format.

The only Organette strips that I cannot punch easily are for the
Autophone that require sprocket holes that are of a different size than
the notes.  This could be accomplished if I added two fixed punches for
the sprocket holes.  I don't know if I will get around to that project.

Ed Schmidt - Schmidt's Music Rolls