In order to get a "feel" for the minimum amount of vacuum that one is
able to play a piano at, we should first consider the pneumatic itself.

The treble pneumatic in most Duo-Arts measures 1-3/8" wide and 4" long.
This equates to 5.5" square.  When 2.5" of vacuum is directly applied
to this pneumatic, it will lift with an initial force of 141 grams on
open cloth (at the finger) and drop to 105 grams fully closed (cosine
effect) if I did my trig right, no friction or air losses included.
Trouble is, no Duo-Art uses open cloth.  There is a rest rail under
each finger which closes the cloth to about 60 degrees before it
initiates closure, so for all practical purposes, our initial lift is
about 109 grams with zero loss, zero friction.  (Theoretical.  You
won't ever see it)

The threshold gram force required to overcome jack letoff in a grand
action has always been around the mark of 150 grams and up.  (We are
not referring to the static balance gram weight).  This is the minimum
average force on the center notes of grand actions required to play
that note.  Many grands require somewhere around 200 grams to overcome
jack letoff, and ironically, even more for Duo-Arts since many
rebuilders often add a turn back on the jack rest screw to make sure
they don't have any slipping jacks after they've got the player
installed above the keys.

Now uprights are considerably less, some actuating at about 70 grams or
so, minimum threshold.  Of course all pianos vary, sometimes widely
within themselves from bass to treble, but these are reasonable figures
to put us in the ballpark.  (A gram is tiny -- 1/28th of an ounce)

This means that Paddy Handscombe's player whose first intensity, he
claims, is 2.5," second is 3," and third intensity  3.5," can only play
a grand piano in outer space.  It might lift the hammer a little, after
we consider an extra 10-20% frictional loss and a bit of traveling
first (1/32") by the striker, and a weightless and frictionless valve.
In addition to these loss factors, there are wide natural variances
between note actuation weights, sometimes varying by a factor of 2 on
the same piano.  Just normal.  The weight I cite refers to the average
ideal minimum key force for white center notes only.

But to put this in perspective, those same 3 intensities are also used
with the soft pedal on, which requires the #2 intensity just to
overcome the lost motion of an unloaded pneumatic.  Measure it and see
if they knew what they were talking about or not.  That is why Duo-Art
always turned on the #2 intensity in addition to anything else played,
just to operate the soft pedal.  The #2 overcame lost motion.  That
should tell us something here, and remember, Aeolian made it plain for
years that their zero accompaniment intensity was usually around 5" on
a water gauge.  Not for all pianos, but as an average.

The only player piano that I am aware of that can play a valve at 3.5"
of vacuum and actually seat it reliably without leaking and wobble is a
few Ampico model Bs, but that valve was specially designed late in the
game for this purpose.  It does this through a precision-machined
active "ball bleed" that allows its pouch to initialize on a tiny #70
bleed hole, and then hold itself open through the combined two bleed
holes.  It is also optimized in other ways, and the original pouch
leather was usually between .005 and .007" thick.  I've never seen a
Duo-Art valve which will  initiate at under 2.5" of vacuum, so that
2.5" of vacuum would be easy for it to handle.  This is because of the
normal #60 metal bleeds (and those which vary considerably among the
celluloid varieties), then you have normal pouch seepage, nipple leaks,
note gaskets and such, all adding up to another (approximately) #65
size bleed, overall seepage.  Remember, the Ampico model B valve itself
is not really designed to play reliably at 3.5" of vacuum, and with new
pouches at the very thinnest averaging .010 and sealed with something
modern, do good to actuate at 4." Luckily, as softly as the Ampico can
play, controllably and reliably, it doesn't need to operate at 3.5,"
but is about the only valve I know of that can.

I realize that some rebuilders are certain their pianos operate at
3.5".  I have no objections.  But I do question 2-1/2" and 35".  That is
30% less than the minimum that an Ampico model B valve is able to
operate on reliably, and it isn't going to happen.  Add to this the
normal leakage of a lightly actuated valve and also the losses of each
pneumatic actuated, and the pneumatic should for the most part barely
even budge, even if a valve or two has been able to actuated by itself
at that pressure.

Too often, rebuilders trust a poor gauge or one that they have had a
long time, and water gauges likewise are not all the same.  For a
manometer water gauge to be accurate at very low pressures, they need
to have a little surfactant added to the water and should be at least
1/2" ID. and preferably tipped at a 45 degree angle and then
recalibrated accordingly at, say the first 6" of very low levels.
I am not doubting Paddy's veracity, only his gauge.  What he has given
us so far however, is a physical impossibility.

Craig Brougher