Marc Wrote regarding the Welte recording piano:

> I think I heard that the Welte recording mechanism attached carbon rods
> to the keys.  The end of the rods were dipped in mercury, so moving the
> key changed the resistance.  As the key was pressed down, an inked,
> wedge shaped, rubber roller contacted the master roll.  The further the
> key was pressed, the harder the roller pressed, and the wider the
> line.
>
> I don't know how playback was accomplished.

    I would like to know, myself, Marc.  It has only been speculated as
to how such a thing could have been accomplished.  Technical writers of
all ages whose job it is to second-guess the designer usually get
things about half right, so whether or not we even have an accurate
verbal model is another question.

    A carbon rod attached to the bottom of each key firmly would have
its velocity measured perhaps, but unless they had valve springs for
front rail punchings which would respond by depth to the impact, I
don't see how that would measure striking force by dipping further into
the mercury.

    The deformation change of an inking roller however might be
designed to be at about the same rate of increase as the force measured
at the key, when using either conductance increase or as a variable of
the degree of saturation of a transformer.  If the carbon rod were
sprung from the key and small enough so that the buoyant of the
mercury didn't affect the touch, and if the rod itself were tapered,
with a non-conductive fiber end on it as both a guide and a "constant,"
force-wise which could be weighted out so that key touch was
compensated for, then we might have something, recording-wise.

    Suppose The registering circuit were designed to control core flux,
which in turn operated the marking roller and determined its force with
a light duty control winding(?) Not much current required there when
placed into a magnetic circuit, and they were loving the mysteries of
inductance about then.

    Dr. Clarence Hickman gave a talk in Philadelphia about 1979 at the
AMICA Convention to which I was attending, and mentioned that the
Aberdeen Spark Chronograph he redesigned to record Ampico B rolls on
was able to use ordinary paper, instead of the "expensive stuff" that
was available.  I suspect he was referring to a recording paper used by
others in the business which had an opaque white wax coat over a carbon
black base sheet.

    There was other roll paper too, some waterproof, which could be
electrodeposited over conductive ink almost as quickly as that ink was
laid down and dried.(Michael Faraday and others were doing that sort of
thing shortly after the turn of the 19th century).  So I wonder if the
technique had something to do with a conductive base sheet, an inking,
a coating, a wash, and then an electrodeposition that connected the
marks to the carbon paper and used in exactly the same way the
information was taken to begin with-- to control windings in a magnet
structure or solenoid under each key?  Individual styli as wide as the
widest copper mark would then sense the resistance of each mark in a
precise, fine  line across the master note sheet and measure its
conductance as a function of its width-- something like that, perhaps.
Any further ideas?

    I have found that too often, modern skeptics are wrong more often
than they are right.  When we judge our forefathers' motives by the
level of our own, we sometimes are  embarrassed.  and rightly so.  I
believe in giving everyone the benefit of the doubt first, asking
questions instead of always having the answer, wondering how it could
have been done, and then continuing the discussion now and then until
somebody pops up with a definitive answer.  There is one out there.  I
happen to believe that Edmund Welte was an honorable man, and would not
stoop to fraud or deceit in any form just to sell more pianos.  Can
anybody vouch for him?

Craig B.