Pete Knobloch presented a set of vacuum readings that I would say I
have never seen the likes of.  I don't even know how to get a Duo-Art to
do that if I wanted to, given that everything else is correct.

   If in fact his pump speed is actually running 7% slower than
optimum, all the while putting out these awesome slopes of expression
gradient, then I would have to agree with Pete-- I wouldn't like it,
either.  I also would not allow the machine to beat my piano to death.

   Without actually having the expression box in front of me, I cannot
offer a positive solution.  I cannot explain how a Duo-Art pump could be
capable of doing what he claims for it at that speed with fixed leaks
required to do this test.  It has a fixed capacity, and if 100%
efficient, cannot propel a piano into orbit like this one apparently
is.  The concert piano pump speeds were doubled, and perform about
similarly, except their Theme curves were still straight.  This one is
just nuts! The Theme chart is really wild.

   My first question is, Were these figures made with the suggested
fixed bleeds, or dead-headed straight into a stack? The linearity of
the accompaniment side of the box however is pretty good up until the
13th step, at which time it goes off "into the Boondocks".  That
suggests some problem like a warped knife valve that cannot close (or a
leaky gasket under the knife slide), at the same time the stack valves
have been sucked down cartridge-tight, so up goes the pressure.

   If we are in the ballpark, here, and the gauge used to chart with
was accurate, then I suggest that the only way these figures could be
had is if we are playing into a (believe-it or not) leaky stack! What
might happen is this:  If we do not have fixed (linear) bleeds of at
least 5/32" to represent valve playing losses, then these figures are
irrelevant.  As the pressure builds in the stack, the valves get tighter
and tighter, and at the accompaniment pressure of 26", suddenly jumps
to 34" (as it shoots past the 29.5") and tops out to whatever the
spring allows.  [Warped knife valves will still close tight, usually,
but not exactly where they should. ]

   However, we don't see the same effect during the theme portion of
the curve.  It actually rises dramatically between 6 and 9, then loses
slope between 9 and 12, becoming less slope than the accompaniment
(although higher in intensities).  So that might tell us that the half
of the stack tested under theme air simply did not have the same leaky
valves which could have caused the effect we see when reading
accompaniment.

   Let me explain that.  You do not have to even move your gauge.  When
you test accompaniment, you are actually testing the entire stack
characteristic.  There is no stack separation.  When you test the theme
air, only then is your stack divided.  So if the box is linear for a
ways and then jumps up, that suggests a combination problem of leaky
knife valve plus tightening stack valves.  In other words, perhaps not
enough clearance above the pouches, or a bunch of cross valves whose
time has (ahem) come.  (Characteristic of rotated cross valves, too).
And often, you'll see a combination of those problems.

   The nonlinearity of the Theme chart is a puzzle, frankly.  However, I
would not gray-scale the box to suit your fancy.  Every thing is
dependent upon everything else, so it just can't be done.  The piano
will never play realistically that way.  When you misadjust one thing,
you cause something else dependent upon it to be really fouled up.

   With the appropriate bleeds, My box tests begin at 5" and 5.5"
usually, and end up at somewhere between 24-26" for the accompaniment
and 34-38 " for the theme.  It depends on the springs your box comes
with, but the important thing is the linearity, not the top end.  If you
get good linearity, then you can hop up your pump and still maintain
it.  My bleeds are .161" and I clamp my pump at something like 40" so
that I purposely cannot get above that.  (It is a Duo-Art pump for a
test pump, but has its own regulator.)

   Before wondering what to do next, I would use the leakage tester
device pictured in the Orchestrion Builder's Manual to test both sides
of the stack, first.  Disconnect the expression box, cover the
trackerbar, and test the leakage with a steady vacuum source set at
about 15".  You can just bleed off the excess, as long as the load
doesn't change the pump pressure.  Test both ends of the stack at
7-1/2".  Then raise the pump pressure to 40" or 50" and test each end by
setting the tester to half of that.  If it tests considerably different,
you have leaky valves which are pressure sensitive.  Probably cross
valves, but sometimes valves set too close to the pouches, too.  A tight
(but not too tight) Duo-Art stack should read a leakage equivalent of
about 2-3/32" dia. holes, or a little less.  just one 3/32" hole may be
sufficient leakage to use with a knife valve, but I prefer a leakage
equivalent to about 1-1/2 3/32" holes.  The Duo-Art must have valve
leakage to work correctly, and that leakage should be fairly constant,
although not perfect.

   Any Duo-Art having patent leather valves for instance, will never
adjust correctly.  Neither can you add a fixed leak to it, because that
leak is too constant.  Don't get me wrong, now.  The valve leakage will
never be "linear," but  it is necessary at all times except at the very
last step intensity.  So it's normal for valve leakage to decrease as
pressure rises, but not very much or very fast.  When you see an abrupt
change as the pressure is slowly increased, that is a problem valve(s).
Let me know, and I'll tell you what to do next.  With luck you will be
able to fix this problem right in the piano.  And if your vacuum
numbers were taken without bleeds, then before we can know anything
else, you will have to provide them.

Craig B.