I wrote in Digest 97.01.25, "... by decoupling the keyboard you remove
mass that the striking pneumatic has to accelerate when it plays ..."

This is wrong, at least on my piano.  I didn't look at my action before I
sent this out .  Rob DeLand's comments on this subject made my light bulb
go on.

Upon inspection of my piano (Autopiano), the striking pneumatic only
accelerates the upper part of the action, from the wippen on up.  An
unlocked key resting in the keybed is balanced such that when the upper
action is moved, the weighted key rocks on the balance rail key pin to
follow the moving wippen; this key part is all done by gravity.

By not locking the keybed, the resistance part comes when the note stops
playing, and the returning action moves back to it's resting position. If
the key has moved with the playing note (unlocked keybed), the entire
action (key included) has to return to home base resting position.  More
mass = slower action, thus Rob Deland's assumption that a faster
repeating marimba roll (barf!) will play correctly.

My player never "sees" the mass of the key, in the locked or unlocked
mode.

Karl Ellison - Ashland, Massachusetts U.S.A.¶
KBEllison@aol.com  - http://members.aol.com/kbellison/kbe.html

 [ In most player grand pianos the pneumatic pushes upon the bottom
 [ side of the key, at the rear, and so the key must move when the
 [ player is operating.  In this case, yes, key friction and key mass
 [ greatly affect repetition, especially at low pressures.  -- Robbie