It has been interesting to read the varied viewpoints of recent
contributors on the subject of modern material used in place of leather
for valve facings, as well as the past discussions of the topic in the
MMD Archives, as referenced by Robbie.

I've recently dealt with at least two players previously rebuilt with
synthetic valve facings.  One of these came to me described by its
previous owner as "a recently rebuilt and functioning 1922 Steinway OR
Duo-Art that just needs minor adjustment."  I suspected otherwise even
before I saw it in person.

The late 1980s materials looked great, and both cloth and the "thin
leather over foam rubber" valve facings seemed supple, clean, generally
airtight when tested at the individual valve level, and more-or-less
properly installed.  The materials seemed to have aged well with little
sign of hardening or degradation.

However, the overall stack tested as somewhat leaky, and no amount of
troubleshooting or adjustment would make the Duo-Art play properly.
With the "Level 0" adjustment set normally, it played the very softest
and very loudest passages, but failed to play reliably at intermediate
levels.  I gathered from the previous owner that this had been the case
ever since the expensive 1980s rebuild.

I completed the full rebuilding of the piano and player just last week.
It took me about a month longer than usual, mostly spent on the stack
removing the 1980s materials and synthetic glues.  With new brushed
leather valve facings, the rebuilt stack is very airtight, and it plays
to my ear as the Duo-Art should, seen in these YouTube postings:

  http://www.youtube.com/watch?v=DdL8k_1q4TE 
  http://www.youtube.com/watch?v=8XldZ6IFkLw 

Sponge rubber or similar modern materials may not always be inferior
for all applications but I much prefer time-tested natural materials
like quality pneumatic leather, particularly in instruments designed
with those materials.  The focus on how natural or synthetic materials
behave over years of use and aging is important.  As a polymer chemist
and materials scientist in my previous life, with a focus on how
materials age, I feel I understand many of the issues, and I certainly
don't reject using all synthetic materials.  But accelerated testing
for long-term material behavior is tricky -- I am dubious of the
assumption that newer synthetic materials are always a step in the
right direction.

Equally important is whether the properties and performance of any
new material match what was required by the original design.  And with
that, it is important to fully understand the function intended by the
original design, especially for reproducing players.  I believe there
are several common misconceptions regarding both optimal player function
and replacement materials, leading to recurring problems with rebuilding
and regulation, especially for the Duo-Art.  Several of these have
surfaced in the recent discussion.

To begin with, I do not accept the widespread assertion that the
original Duo-Art design benefits from additional air leaks into the
stack, either through the lower valve seats or through bleed holes
drilled into the stack as has been recommended by some.  While the
zero-level knife-valve settings can compensate for a little leakage
airflow from the stack without totally degrading performance, the
original Duo-Art design operates best when the stack is very nearly
perfectly airtight when no notes are playing.

The spill-valve in the expression box (open during quiet playing)
is adequate to provide airflow to keep the pump from dragging and to
properly shape the expression response at lower level settings.  I see
no basis for the idea that the Duo-Art expression regulation requires
steady airflow from the stack to function.

Significant leaks into the Duo-Art stack will always degrade
performance.  Less harmful are simple leaks through an opening, where
the flow from the leak increases as the stack vacuum increases.  More
harmful are leaks most pronounced during quiet playing a low vacuum but
which "self-seal" at higher vacuum during louder playing.  And this is
exactly what often happens with 1980's era synthetic foam valve facings.
The 1922 Steinway had both problems after its 1980s rebuild.

The 1980's "thin pouch leather bonded to foam rubber" valve facing,
even when new, was mechanically the opposite of the original Aeolian
brushed pneumatic leather installed with the brushed side out, sealing
on the metal valve seat.  The synthetic foam is a "flat, smooth surface
over a compliant core", rather than the "very soft, compliant surface
over a tough, resilient core" of normal valve leather.

The soft brushed leather surface is ideal for sealing the lower valve
seats completely with just the pressure of gravity on the valve button
and the 4-5 water-inches of vacuum of the stack at the lowest vacuum.
The smooth-surface foam rubber requires higher pressure, meaning higher
vacuum levels, to conform the flat valve surface to the rigid seat to
fully seal the opening.

While the low-vacuum leak from each individual valve may be small, the
combined leak from all 88 foam rubber valves was enough to make proper
regulation impossible for this particular Steinway Duo-Art.

In the case of a simple leak (like an added bleed) the knife valve
adjustment can be opened to compensate for the leak to get the correct
"Level 0" power for playing single quiet notes.  But then the knife
valve is also open further than originally designed at all higher level
settings.  And this leads to several common adjustment problems for the
Duo-Art.

First, chords at intermediate levels will play too weakly, often
missing notes at Levels 2-4 or so, because the roll coding assumes that
Levels 2-4 have significantly more power than Level 0.  But with the
leaky stack the knife valve opening size isn't changing enough between
Level 0 and Levels 2-4 to make the difference needed to play multiple
notes.  The original design assumed that the knife-valve opening is
very tiny (almost pinhole-sized) at Level 0 with an airtight stack, and
that Levels 1-4 open a proportionally much larger aperture difference
in the air channel.

To compensate for this problem, the "Level 0" setting for the leaky
stack is often adjusted even further open so that mid-level chords
don't miss.  But then the quietest playing is way too loud.  And
usually this also causes the expression response to plateau at the
highest levels, where the knife-valve aperture is open so wide that
it is no longer restricts the flow, and Levels 12-15 or so all play at
about the same maximum power.  The net effect is very poor expression
performance, with the softest levels too loud, intermediate too quiet,
and no differentiation of the loudest levels.

Even worse is the effect of valve facings that seal better at higher
vacuum than at lower vacuum.  In that case, the "Level 0" knife valve
setting is usually also adjusted too far open to compensate for the
leak.  On top of that, the stack vacuum becomes unstable at
intermediate pressures, as the foam valve facings start to deform and
seal better than at low vacuum.

The net effect is that the player plays generally either too quietly
(at low vacuum levels where the valves leak) or too loudly, (as the
valves deform and seal better but the knife valve is mostly open).
And there is almost no controlled playing at intermediate volume.
While I've encountered a few people who consider such "playing at
extremes" as good expression control, I do not believe it is what was
originally intended.

The common recommendation that valve leather be installed to seat
"smooth side on metal seats and brushed side on wood seats" also seems
to cause this problem, as well as to be opposite of what I've seen in
original Aeolian installations with the brushed leather surface on the
metal seat.

Of course, not all of this applies to other systems like the Ampico,
which operate on very different regulation principles compared to the
Duo-Art, and where the original designs often included spill valves or
fixed bleeds into the stack.  All original instrument designs may have
some minor aspects that might be improved upon.  But with the Duo-Art
especially, I tend to be dubious of those who propose wholesale changes
in how it should be rebuilt or regulated with the assumption that the
original design was somehow inferior.

After all, the design was manufactured and marketed with few changes
over almost 20 years.  And a 1922 Steinway Duo-Art OR was considered
the pinnacle of the form, at a new cost of about seven times the
average annual US family income.  And to my ear the best-sounding
Duo-Arts I have heard have generally been rebuilt closely to original
specifications.

Ralph Nielsen