Hi everybody,  I was just now reinstated in MMD, and thought although I
cannot be quite as active as I once was, I will start off by answering
a few questions of Sam Harris about the Duo-Art box regulation.

The "ideal box" will chart perfectly linear with the step numbers
across the chart and the pressure measured with a fixed bleed leak at
either output elbow.  Straight lines are not possible in a practical
sense but, as Sam Harris has already shown, one can get it fairly
linear.  The Theme slope should be greater than the Accompaniment
slope, because the rate of the theme spring measures double that of the
accompaniment spring, due to its thicker wire (at least, this is the
conclusion I have come to from having been measuring spring rates on
these boxes for many years).

Most of my measurements and charts have shown that when you begin the
Theme at 5.5" and the Accompaniment at 5" you will usually end up at
40" and 28", respectively, in a reasonably straight line with little
or none of that hysteresis 'S' curve so common with mechanical
contrivances that are a bit slow to start and then tend to top out
toward the last few steps.

For all those rebuilders who believe that by merely adjusting the
accordions perfectly they will have a perfect box, I think they have
a rather large surprise in store for them.  The Duo-Art expression box
is a complete system in itself, dependent upon many, many interacting
parts, all of which (except the spill felt) move.  Change one, and you
have to change them all.  Just simple things like spill spring tension
can change what you are getting when you chart the box, so when you
replace the box and add expression lever tension, that too enters in
to the accordion speed and hence depth of some expression codes.

The most troublesome of all the box's problems can be shaft runout
or wobble and/or tight or misaligned knife valve hook wires inside the
"regulators," as the felt bushings in the box and the holes they are
glued into can allow the shaft to deflect.  Also there is the problem
of the accordion brackets deflecting slightly as they operate, and
especially the two accordion shafts are quite often too tight for a
variety of reasons, making them have too much static friction and it
causes a return error, preventing the accordions from returning quickly
enough or at exactly the same place, given the same expression code.
When a difference of just 1/32" in the step 1 accordion can make a 1.5"
vacuum difference at Zero intensity, and when a total system like this
box can only be held to a 1/16" overall tolerance at best, then I think
you get a more accurate picture of what you're up against, and how
difficult it may be to do what you would like to do.

What most rebuilders do to improve repeatability is to increase the
spill spring tension.  This is exactly wrong!  They exchange poor
repeatability with a slow box (another characteristic of a good box
that I'm not even covering, here).  But then, they worry that if they
ease the snugness of the top accordion shafts to prevent rotational
friction, they will cause the shafts to deflect laterally and create
more non-linearity (which is correct).  That's why doing a Duo-Art box
is an artistic tradeoff that may require a full day of my time just to
get it absolutely right again (and I've been doing these for years).
That's why it's really not possible to explain all of this in a letter.
This box is a Rube Goldberg trapeze act.  You either figure it all out
and then take the time required to chart the box and keep doing it
until it's right, or you say, "That's good 'nuf." How critical are you?

By the way, all this testing begins with some good gauges and specially
built testing equipment for the purpose, plus a very accurate bench
pump and regulator, and a box fixture that allows you full access.
Centrifugal suction boxes will NOT work.  You must have a positive
displacement Duo-Art pump if you want it absolutely right and if your
figures are going to mean anything.  This all takes a lot of preparation
and time.  Then your first box or two might take 3-4 days of work and
you might find yourself saying, 'This is impossible."  One thing for
sure: it isn't cost-effective, and few can tell the difference between
mediocre and great.

I mentioned the accordion settings earlier.  The accordion settings
should be set approximately right to begin with.  That gives you a
good base to start with.  But as anybody can see, the degree of closure
depends on 3 things: a screw, a cushiony, soft felt pad, and the actual
pressures being generated at the time.  The third item, the pump
pressure, varies, varying the depth that the screws can bury themselves
into the felt cushions, anyway.  Then after the box's many other
parameters and adjustments and easings have been optimized, the
perfectionist will return to the accordions and rescale them for a
perfectly linear static adjustment of the box in both Accompaniment
and Theme slopes.

The regulation requirement of the Duo-Art box is that the settings of
Accompaniment and Theme at their Zero intensities are such that the
Theme will always take precedence over  Accompaniment.  That can only
be done if the difference in their preset level is great enough to
raise the one way flap valve for that side.  For all practical
purposes, and for the characteristic of the average piano, that level
is about 1/2" difference (not always, but usually).  So you always
begin with the box zero adjustments begun at the ideal level with an
approximate fixed bleed of about 3/32" or so.  Too small a fixed bleed,
and you will get possibly more ideal charts, but not very good
linearity in actual practice, and you will have to do it all over
again.  But then again, the sizes of those bleeds depend in large part
on the pump capacity, because the Duo-Art system is also scaled to gas
out the pump and it's own capacity drops at a predetermined rate as
demand and pressures increase.

Craig Brougher