In Bob Taylor's letter, "Duo-Art Regulator Leakage', it would seem that
perfectly airtight valves are ideal valves, and that the problem in
Molly Yeckley's player was not the tight valves, after all.  Bob said:

"If the stack is extremely tight, the vacuum leaking around the knife
valve now starts to build on the back side of the knife valve.  Now the
effect of the diminished pump vacuum is exacerbated by a drop in the
pressure behind the knife valve.  The result of all this is divergence.
In a no-demand situation, a small leak will lead to near full vacuum
at the stack, when it should be at minimum.  The next note played will
be loud regardless of roll coding."

This effect happens in all Duo-Arts, including the old ones with their
original valve leather -- unless, of course, they are so weak that
nothing much happens at all.  The stack doesn't have to be tight.  That
proves the effect isn't only caused by over-tight valves or leaking
knife valves, but by something else.  However, I want to stress that we
are speaking here of degrees.

That something else is really regulator inertia and overshoot into an
overly tight stack.  When Duo-Art valves are too tight for it, it's
going to build up vacuum unless something can drain it.  So it relies
on valve travel losses and constant motion, vacuum levels changes, etc.
When everything is moving, it's going to be fine.

"If the stack valves leak a moderate amount, the divergence does not
occur.  But leaking stack valves are really not what anyone wants.
The best valve is one that does not leak."

First of all, the knife valve is never completely covering the pump
supply hole when the regulator is set to zero intensity (unless the
knife valve is warped).  So it could not be possible that any stack
is ever 100% airtight, or the stack would instantly rise to the pump
pressure.

Yet, with its maple construction and taped joints, the overall valve
leakage of a Duo-Art is necessary, warped knife valve or not.  That
leakage is created by its stack valves.  But their leakage is somewhat
inversely proportional to the vacuum supplied.  The greater the vacuum,
the tighter the valves become, and the less in proportion the system
leaks.

Now this has its caveat, too.  There are other leaks, all summing
together, which are not valve leaks.  Those leaks are not pressure
sensitive.  But it is the sum total of leakage in the stack at zero
intensity that the box is regulated to.  From that point upward, the
percentage of valve leaks gets a little smaller.

Perfectly airtight valves create a problem during play (Bob has alluded
to that already), in that pressures develop at a different slew rate
than expected and provided for by the roll coding.  In each instant
that the regulator overshoots, as happens all the time, the supply hole
can completely close.

Normally, this is anticipated by the roll.  But depending on the time
it remains closed, a synthetic covered valve will always build up
higher pressures unexpectedly and almost instantaneously, whereas suede
leather can be timed, and is always the same regardless.  Remember,
this buildup became a bit more noticeable with the later Duo-Art valves
which were tighter, hence the equalizer hole.  But with patent leather,
they were MUCH tighter.  We are talking about degrees, here.

Also, suede returns softly, whereas patent leather and sponge hits the
seat and stops the air instantly.  So when large chords are released,
there is a "shock effect" in the stack as they close their holes
simultaneously, which you can see is much harder, by watching the
effect on the regulator.  That shock drops the regulator so hard that
its overshoot might be double that which it was designed to compensate
for.  I don't know, but you can see the effect on it, and it is
different.

While the equalizer hole in the accompaniment regulator of a _very
precious few_ late model Duo-Arts would prevent stack vacuum build-up
through the regulator on the accompaniment side, it would have
absolutely no effect at all during themeing, since at that time, the
accompaniment is closed off to the themed side(s) of the stack.  The
dual bleeds would affect both sides of the stack at all times, but
then, they also affected the expression curve above MF as previously
mentioned.

Also, I believe that one will discover that the Duo-Arts with the
equalizer hole have the large pump, the small valves, and possibly
identical regulator springs.  In any wise, that model has been
redesigned to accommodate it.  The effect of the constant flow of air
through the expression box is to prevent a buildup of pressure due to
the smaller and considerably tighter Duo-Art valves of the later
variety.  Bob agrees.  That's why I do not understand his comment here:

"The patent leather valves in that piano, as cited by Craig, seemingly
created the problem of the stack being too tight.  But was it really?"

I interpreted it that way at the time, and Bob has already given us
Duo-Art's supposed solution to the problem.  So if he says it can be
caused by overly tight valves, if your knife valve is warped, he is
right, but then, the knife valve is not intended to completely cover
the supply hole, even at zero intensity.  And if your knife valve is
warped, that's the easiest problem there is to diagnose in the Duo-Art.

"To counteract that problem (of warped knife valves), later Duo-Art
expression boxes have an atmosphere bleed introduced directly into the
chamber of the expression regulator.

The reason that the equalizer hole was a good idea with the late
stacks, is because the accompaniment supply hole was so large in
comparison to the new valves.  Yet it needed to be for the top end
power.  So when you spread the contact area of a knife so widely around
a large hole and then bottle it up twice as tight, you are going to get
leakage past it, despite how flat it may be.  Knife valves leak through
a closed system, but so will any valve.

To show this effect, place your thumb tightly over the outside valve
hole of a player valve and watch the pneumatic crawl upward.  The
leakage is coming directly through what was otherwise a "perfectly
tight" valve.

"The Duo-Art knife valve can be very tight, almost 100%.  Testing and
achieving the 100% tight condition is elusive.  The most revealing
tests are the ones at very low vacuum levels, as the higher levels tend
to aid in sealing the knife valve, thus masking a problem."

Actually, no amount of vacuum is able to flatten down a Duo-Art knife
valve, if that is what Bob meant.  But when higher levels are called
for, that he refers to, that means the knife valve is opening anyway,
so its warpage is no longer the major factor (unless it's really
unbelievably awful).

Warped knife valves are easily diagnosed.  If you have a warped knife
valve, you will not be able to adjust your zero intensity down low
enough.  It will always be leaking a little, and no amount of turning
the regulator screw will seem to help.  Molly's player didn't have that
problem.

"The bottom line is, tight stack valves are only a problem if the knife
valves leak.  Leaking valves degrade performance.  Tight valves enhance
performance.  If both valves, stack and knife, are tight, the
performance is breathless, in several ways."

As we can see, this is not the case.  Bob has already said that overly
tight valves will cause the Duo-Art reproducer to build up too much.  He
is right.  However, even moderately tight valves will do it.  We are
speaking here about degrees.

The main problem however, isn't the switch from the quiescent state to
the dynamic one, it's the problem that overly tight D/A valves create
during play.  That's the part that you cannot control as well, and why
no one could ever get Molly's piano to play correctly.

Also, adding the equalizer hole to an older Duo-Art is not the solution
either, since other things, geometry and the accompaniment spring have
to be changed.  There are also other factors that have an effect too.
What you could do with the new valves, you cannot do with the older
ones.

Craig Brougher