I've been reading the thread about the Welte ink-based piano recording
system with considerable interest.

The recording technique seems straightforward enough.  I might add that
while rods dipped into a trough of mercury will indeed make surface
waves therein, these waves have rather low amplitude and move with
great speed.  You can see this if you've ever fooled with a jar of
mercury.

(Aside: troughs of mercury were once used as dynamic computer memories
back in Eniac days.  A piezo or electromagnetic transducer placed at
one end of the trough would vibrate a few bytes of data into surface
waves in the mercury.  These would be reflected back, read, and
re-transmitted back down the trough.  Eventually, the data could be
read when needed.  I don't know how this form of memory was erased.)

In any event, it's not at all clear to me how the inked data would be
read.  If we were doing it now, we'd use 88 small lenses and photocells.
But while various photocells existed in 1903, I don't think they were
a practical proposition.

I seem to recall from a previous discussion that the ink was conductive.
This leads to some speculation as to how the playback system might have
worked.

- There would have to be two probes for each key.  These would have
to be placed such that a line drawn between them was parallel to the
movement of the paper.  Otherwise, the thickness of the conductive
trace would have no effect on the resistance seen between the two
probes.  However, the required distance between the two probes would
reduce the resolution of the system on short-duration notes.

- The trace of conductive ink would have a resistance that was far
too high to actuate anything but the most sensitive of relay coils,
and its current capacity would be very low in any case.  Thus we'd
need 88 sensitive relays to amplify the signal.  Now, telegraphy was
sufficiently advanced at that time so that the required relays should
have been readily available, so this amplification shouldn't have been
a problem.  These relays should have been fast enough to keep up with
keystrokes, given that fast multiplex telegraphy was in use by that
time.

- This hole punch would have had either big electromagnets that punched
holes directly, or a motor-assisted arrangement somewhat like the
electric typewriter.  I like the idea of this last arrangement, because
sustained notes would have produced an evenly spaced line of holes.

Does anyone know if this bit of reverse engineering (done out of utter
ignorance of the real system) is correct?

Mark Kinsler
Lancaster, Ohio USA
http://home.earthlink.net/~mkinsler1

 [ Richard Simonton, Jr., related his father's statements about the
 [ T-100 recording system, carbon rods, the trough of mercury and
 [ conductive ink in 000129 MMDigest, "Welte-Mignon T-100 Recording
 [ System Technology".  -- Robbie