Niederviehbach, 7. October 1996

                            AMADEUS

       Computer Assisted Music Perforating Machine

Part 1 : Mechanical Function and History

This AMADEUS was born in summer 1992 after 6 long years of searching
and building.  Between the idea and the certitude "that it is possible"
and the first music without distortions or "computer-sound" was a long
way of hunting little devils of details. I also needed to meet the
right person:  Doctor Patrick Sanchez, of the CNRS (Centre National de
Recherches Scientifiques) of Marseille in the south of France. This
genius man was the first one to understand my dream of making
perforated music without suffering (and who had the scientific package
to help me in this adventure) !

So we seat down on a table, looking on the Mediterranean Sea under the
wonderful light of the French "Provence" and put together out ideas in
front of a glass of "rose".

In the beginning I thought I knew what I wanted but in getting further
on with working on the system I had to understand that the problem was
much more complicated than I thought ! "Easy" things now seemed
difficult and difficult things did not take time at all. But between me
and Patrick was (and is) respect and comprehension of a beginning
friendship and so we went on efficiently.

After our first session I could start building all the mechanic parts
of the machine.

I choose a wooden structure because it was easy for me in my
organbuilder-workshop to make one, and,  as my father was constructor of
robots and wooden aircrafts, I had good knowledge between my childhood
in building light and easy complicated things. Another reason was, that
I wanted Amadeus to be able to travel easily in my van to be able to
make demonstrations outside.

The punching-tool itself is traveling in the y-axis on a carriage with
longitudinal ball-bearings. The paper or book is driven in x-axis by
rubber-rolls, both axis are powered by step-motors with 2/100 mm
steps.  That is the best resolution, it makes no difference if you make
it higher, the paper would not "feel" it. In "human" punching-systems I
observed best results (in the morning, no head-ache, no lover's grief)
of 0.5 mm, normally 1 mm. That makes 1/60 of a second calculated with
normal speed ! I researched tolerance-limits of "normal" people in
making tests with music and found out that 2/100 of a second are
perceived. Trained musicians or instrument-builders can hear better
than 5/1000 sec. That is about 0.3 or 0.2 mm on a roll or book !

We have here the difference between "music" and "mechanical" music. And
we find here the reason why one can make very nice music with a MIDI
system : it punches exactly what you can hear before and even the copy
number 37 is as good as the first !  The quality of interpretation is
more important than understanding how it goes, that means that
musicians having no idea how to sharpen a punching-tool or folding
cardboard-books or adjust a complicated machinery can concentrate on
their real work:  making music. This opens a lot of possibilities for
many people.  One of the reasons that made me choose a MIDI compatible
system and not one more "internal standard", was the opening on the
world of thousands of musicians in the whole world able to write or
even "record" music for mechanical instruments as long as they have a
computer and know how to generate a MIDIfile after reading a precise
scale. It is one more time possible to make records of famous musicians
as in the beginning of the century with reproducing pianos.

But I come back to the mechanical parts of Amadeus: the punching-tool
is changeable in a few seconds, with one screw, so it is possible to
make holes of any size or form: round or square. The matrix is
interchangeable too. So I have sets of tools in any shapes and sizes.
It is possible to make sets for metal music-box disks and to copy or
create any shape of pins in the matrix, but I never did it.  It is also
possible to punch disks of metal sheet or cardboard. My request was to
make a universal system with Amadeus.

The paper (or cardboard) comes between adjustable guides with a maximum
width of 500 mm. It is possible to punch up to 20 paper-rolls at once
or 2 thick cardboard books for fairground-organs. There is a system to
deliver correctly the papers with individual brakes on each roll. On
the reception side are universal rolls to wind up ready punched paper
stripes. The folded books fall in boxes.

According to the side where the guide-side is, left or right (or up or
down) I  punch from the upside or turn it around. So I am sure that the
paper or books are always adjusted on the trackerbar.  It takes about 2
hours to finish a "normal" music, but the machine works without any
human control, day and night, so it is possible in theory to make 240
rolls a day or 24 music-books. I made a control system that detects
lack of paper or any other problem, it stops the system and calls me,
even in my bedroom. Fortunately this is rare. But as I make music for
theater, television, video clips and other productions, it is sometimes
necessary to make and deliver music very quickly and it is not possible
to waste a night.

There was a choice in the beginning: to make a rather slow machine
with a single tool or a much faster and easier system with a whole
rank of tools in the right division. But it was clear for me, that a
universal system could have only one punching-tool moving in x-y
coordinates. It is also a question of hardware how fast it can be. A
simple paper roll machine could be faster because the tool is much
lighter, but it takes about 500 N/m  to perforate a thick cardboard
instead of 100 of 200 N/m even with many layers of paper. So the weight
my moving part is about 20 kg. The other reason of this weight is the
width of almost 450 mm. But many big organs, orchestrions and pianos
have this size.  The PC is connected with serial bus (as a mouse) to an
interface were the TTL signals are interpreted in powerful commands for
the motors.  The step-motors have to be strong and the interface must
supply many amperes. All that needs cooling in the summer, but small
integrated ventilators do the job.

The U formed part where the punching tool and the matrix are fixed, is
driven by a screw and a nut with ball-bearings, no play and easy to
turn.  It is out of special hard steel, I had to make 2 other of them
from ordinary construction-iron before I understood that any elasticity
in that part was really a problem. It had to be very stiff and in the
same time as light as possible.  In the meantime I made 2 other
machines and the problem is resolved to my full satisfaction.

The punching-tool is gliding on adjustable longitudinal ball-bearings
too and has no play. That is very important because it is impossible to
cut thin paper properly without a absolutely precise tool.  It is
driven by a eccentric and an electrical clutch that makes only one turn
at a signal. I tried out magnets, pneumatic and hydraulic devices, but
only the eccentric is fast enough and one can be sure that it went
through even at the end of the week, when the tools needs
sharpening...

In a few days I will talk about software and MIDI problems.

•--
---------------------------------------------------------------------
| Quercus Edition - Robert Hopp ----------              -------
| Erlenweg 16                   ---             ---
| D-84183 Niederviehbach                ---             ---
| Germany               -----------     --      ---     -------  ----
| Tel : +49 (0) 8702 91183                --       ---
| Fax : +49 (0) 8702 91184                  --        ---        --
| e-mail : Robert.Hopp@landshut.netsurf.de               ---  --
---------------------------------------------------------------------