I now think that I understand for the first time what is going on
with the accordion adjustment procedure and the test roll for the
Duo-Art player.  This is not a step by step example but only a general
description around specific portions of the procedure.

I haven't tested it out but would like some comments.

In MMD 031203, Ray Fairfield says he is having problems with the vacuum
levels of his Duo-Art regulator.

> I set up the accordion pneumatics at 1/16" increments, as prescribed,
> on the bench using a dial indicator. But when installed in the piano,
> the vacuum levels ramp up much too rapidly, and playing the test roll
> reveals that each succeeding level is too loud.

By regulating the accordions by the book, it seems that the vacuum
levels can go up drastically.  It all depends on the strength of the
two springs found on the regulator.  If you were to reduce your spring
tension, then it would cause your accordion distances to increase,
which would put your accordion distances more into spec and the vacuum
levels can still be the same.

Rule #1: You want the total accordion valves to move at their maximum
travel (15/16") so that when the roll calls for a small change of 1/8
inch, the linkage can move that knife valve the appropriate distance.
The felt bushings and linkage does have physical limitations and
anything smaller than 1/8 inch movement for level 1 can give you
problems from one intensity to the next.

Setting the 0 adjust is fairly easy using the test roll and must
be done accurately.  There shouldn't be any problems here.

Something that has bothered me about the test roll, is the way it
is worded when it comes to the Dynamic Chord Test.  When it starts to
call for "Chords to play at level 4 dynamic" and, "These chords should
not play at 4 or play very softly", it gets kind of fuzzy.  We hear
it over and over again, "It didn't pass the no play test" but...

I have finally realized what they are doing in these tests.  This
wording doesn't mean that because it is level 4 that it should play
at level 4, it means:  The notes should just play with the level 4
accompaniment accordion closed.  The next test should read "These
chords should _not_ play with the level 4 accordion closed or the
notes may play very softly".

The test roll will run through accordion level 1, 2, and 4 and you
should adjust each test case to barely play and not play.

The table below shows the number of holes used on the test roll for
accordion level 0, 1, 2, and 4.

 Accordion  | Number of notes     | Number of notes
 Level      | used to barely play | used to not play
 _________  |_____________________|______________
 0          | Not tested here     | 8
 1          | 3                   | 5
 2          | 4                   | 8
 4          | 7                   | 14

This test is done this way to establish a known playing pattern between
the piano and the player without assigning a physical vacuum level to
each step.

The different types of Duo-Art players have different pneumatic sizes.
Different actions have a heavier touch and need more vacuum to strike
the strings properly.  The other problems also come about because it
takes more vacuum to get 4 keys to play than 1.  Other variables are,
how leaky is the leather on the bottom seal, and how does it change
with different vacuum levels?  How leaky is the leather on the upper
seal?  How leaky is the key pneumatic cloth when the valve is ON?
How well sealed is the stack?  Adjusting the #1, #2, and #4 accordions
using the test roll will compensate for all of these differences.

Rule #2: Adjust the 1,2, and 4 accordions using the test roll.  If the
Play and No Play tests seem to pass but it looks like you are having
to change the distances that the pneumatics travel by too much, change
the spring tension and adjust the accordion distances again (Rule #1).
Don't expect any two pianos to be the same.  Don't be concerned with
specific vacuum levels at this time.

What rule #2 does is customize the lowest playing vacuum levels for the
specific piano.  If you plot a graph of intensity level versus vacuum
applied to the stack, this gives a specific slope to the line.  If you
continue this line up to an imaginary level 8, which has not been
tested yet, this should give you an approximation as to where this
vacuum level should be set.

If the real vacuum levels on intensity 0, 1, 2, and 4 don't line up
perfectly with the drawn graph, go ahead and readjust the accordion
adjustments so the vacuum level values do line up.  A few tweaks here
and there shouldn't change things to drastically.  Of course you want
to tweak them in such a way that they still pass the roll tests.
(I don't know if this will work on a non-linear curve?)

Now that you have accordion levels 1, 2, and 4 adjusted, you need to
set level 8.  You get this by setting level 8 to just over the level 7
setting.  Turn ON accordions #2 and #4 and measure the vacuum at level
#6 and write it down.  Now turn on #1, #2, and #4 accordions and
measure the vacuum at level 7 and write this number down.  Now subtract
these two numbers which gives you the next incremental step to level 8.
You get this by adding the difference to the level 7 number.  Now turn
off the #1, #2, and #4 accordions and turn on #8.  Now adjust the #8
accordion to read your calculated value.

Now go through each intensity value (0-15) and measure the vacuum levels.
Plot a graph on paper as before and look for irregularities.  Adjust out
any of these irregularities or jumps by doing fine adjustments to the
accordion distances.

 - - -

Like I said at the top, I don't know if this procedure works but it
is something to shoot holes in and discuss.  Other things that I don't
know how to deal with are: linkage that is adjusted to give a
non-linear curve and the Solo #15 level spill valve that needs its
vacuum level reduced.

I am hoping for constructive criticism and not the "you can't do it
that way" answer.

Pete Knobloch (Tempe Arizona, USA)