Time out!  I invite everyone making comments and suggesting procedures
for regulating the Ampico A to step back, take a deep breath and review
the literature to make sure they have basic understanding of how the
mechanism works so as to avoid putting out inaccurate information.

The first place to start is the "Inspector's Reference Book", Copyright
1923.  The functioning of the crescendo pneumatic(s) begins on page 7.
The last sentence of the 7th paragraph is instructive and _must_ be
understood:

  "This adjustment [in reference to screw #31 which adjusts the
   force or tension of the large coil spring, #28] is for the
   purpose of setting the No. 1 intensity to the proper loudness."

Notably, it is _not_ used for, and it has no effect, in adjusting
the speed with which the crescendo either closes or opens.

Now reference Illustration #2 on page 6.  Those speeds, both slow
and fast, are essentially non-adjustable (at least in post-Stoddard
designs) and are determined almost entirely by the fixed orifice in
nipple #15 (connecting to nipple #3) for slow crescendo/decrescendo,
and the straight-through channel of nipple #14 (connecting to nipple
#2) for fast crescendo/decrescendo.

Because of their small sizes, these two tubes will move air at a
constant velocity even though the supply vacuum for the crescendos
is unregulated pump vacuum, which of course varies over a wide range.
This insures that the crescendo/decrescendo speeds will remain constant
despite other changing parameters.

Changing the position of the operating lever of the so-called "spring
pneumatic" (call-out #23, Illustration 3A, page 8) higher or lower on
the expression regulator valve stem (#21) by adjusting the leather nuts
on either side, will, and should, _have no effect_ on how much force it
develops on the regulating valve stem.  That force is a product of the
vacuum level inside the crescendo pneumatic, communicated to the spring
pneumatic via tube #13, times the surface area of the pneumatic.

Since that surface area doesn't change, any given vacuum level in the
crescendo will develop a corresponding force in the spring pneumatic,
which _will not change_ simply by increasing or decreasing its span
opening by adjusting the nuts.  Otherwise, the spring pneumatic remains
in a state of equilibrium between the force developed as a result of
the crescendo's internal vacuum pulling it up and the unselected
intensity pneumatics (2/4/6//B/T) pulling it down.

Given that the spring(s) on the crescendo pneumatic(s) are for the
purpose of setting the _1st_ (not zero) intensity on each side of
the stack, I would have to take exception to an instruction to set it
to any specific vacuum value, especially if it is disconnected from
the spring pneumatic.  Since this setting is highly dependent on the
friction forces present in the piano's key and hammer action, it will
of necessity, vary from instrument to instrument, and is most easily
set "by ear" to the point that notes are not "dropped" at minimum
coding levels.

Certainly it is clear, at least to me, that no amount of adjusting 
the nuts on the spring pneumatic will have any effect on the "at rest"
1st intensity determined by the crescendo pneumatic.  If such effects
are being observed, I suggest that it is due to some other phenomenon.

I defer further discussion to Bob Taylor.

John Grant