Dick Merchant has some good points.  First, ever since my own positive
posting, I've had misgivings about the idea that notes struck hard
would plunge the carbon rods any deeper into the mercury -- such a
"feel" to a piano key would not spur good performances!

So Dick's idea of using a series inductor as a differentiator (rate of
change detector) is a good one.  Expensive and bulky for 88 inductors,
but we've already agreed that Welte would have liked it big and
complex.  And there's no extra maintenance with inductors, whose
properties remain stable over years.

Some fine points: Dick is right, the recording wheels would have to
be on the verge of touching the paper, so soft notes would pick up.
Even more important, a key held down would have to drop enough voltage
across the inductor (even though the current is steady now) to hold the
wheel down and show the note's duration on the paper.

Fortunately, the inductor's DC resistance (it's wound with many feet
of wire) would solve that problem, *and* make sure that soft notes had
the beginnings recorded, since even a very slow key depression would
increase the current and develop the DC voltage drop through the
inductors resistance.

Looking at a roll using Dick's scheme, only the heads of the lines
would show the notes' loudnesses, with uniform lines for the durations.
Just like the piano, which couldn't care less how you press the key
after the hammer strikes.

One glitch is what happens when the key is released: you get a voltage
spike across the inductor, but of the opposite polarity.  Good diodes
weren't available then to "rectify" the situation, but each recording
pen magnet could have had a second winding with a fixed DC current to
polarize it, such that the voltage spikes from key presses would add to
the magnetism and mark the note's onset, but the opposite surge from
key release would subtract from the magnetism -- and help the pen lift
off the paper even quicker at the end!

Nice going, Dick -- would you like a job in Freiburg?  Lovely town!

Mike Knudsen

 [ I hope that, among the music rolls in the Simonton collection at USC,
 [ there is a roll of paper with ink markings, not holes.  Then we might
 [ confirm that the trace is variable width, and measure the electrical
 [ resistance of the inked trace.  -- Robbie