Hi,  I am getting a bit worried about the correctness of information
supplied by some experts on this group as well as from other sources.
Let us start again from scratch:

Duo-Art has 16 overall levels, from #0 to #15.

The original manual gives these as #1 to # 16. This has definitely
confused everybody as they then speak of the zero setting which would
then apply to #1. Also, the spill must close at power 10 according to
the manual. This would be pneumatic 1&8 according to the table in the
1927 Service Manual p16.  However, if one looks on the illustrated test
roll no. 3, level ten is given on the roll as 2&8.  Aeolian even
confused themselves.

I repeat, the mathematical correct way is level 0 to 15 and this is
what I will use in this article unless otherwise mentioned. The test
roll is then also correct as level ten being 2&8.

'Big deal'. you might say, but not so far I have not seen one pressure
measurement chart which is correct or understandable to me to the point
of full trustworthiness.

Lets start at Craig Brougher's values of MMD yesterday:  zero level
= 5".  Then I assume he uses the confusing system of Aeolian 1927
Service Manual p16 which states that zero level = level no. 1.  Now let
us redo his table the way I think it should be:

  Acc Pressure levels:

  #0 = 5"            #8 = 16.3
  #1 = 7.5           #9 = 17.6
  #2 = 8.8          #10 = 19
  #3 = 10           #11 = 20.2
  #4 = 11.3         #12 = Not given
  #5 = 12.5         #13 = Not given
  #6 = 13.8         #14 = 24
  #7 = 15           #15 = Not given

I hope this is what Craig meant.

Now, let us look at a table given by Mike Kitner in Vestal Press
Technical Series No. 5 (C)1975, "Rebuilding the Duo-Art Reproducing
Piano Mechanism", page 19.  Now, he starts with #0 but strangely gives
levels up to #16 here is the table:

  Acc Pressure levels:

  #0 = 5"
  #1 = 6.5
  #2 = 8
  #4 = 10
  #8 = 18
  #12 = 30
  #15 = 34
  #16 = 38

So where did he get confused?  Should Vestal Press not have this
corrected after some 20 years?

The mathematically correct way is to number the levels from 0 to 15.
This will eliminate all confusion in future.

Now lets look at the contents of these tables:

Craig developed his chart from a straight line he drew on a piece of
paper with a zero of 5" and a #10 = 17.5 (I assume he took #10 as level
9). He states that the pressure should increase linear for each level,
giving him a #15 (level 14) of 24".

All I can say is that Craig did not make these measurements but
created them on paper.  This adds to the confusion, could they
be wrong!

Let us now have a look at Mike Kitner's chart and let us assume that
from his #12 to #16, he means level 11 to 15.

Have a look at the vast pressure differences of 10" at level 14 (24"
versus 34") between Mike's and Craig's tables

So who is right here?

I have measured the pressures of 2 Duo- Art pianos and graphed them
with a spreadsheet along with Mike and now also Craig's table.

Craig's table cannot be met at all with whatever adjustments I tried to
make to my piano, The table of Kitner comes reasonably close to the
behaviour of my piano and I assume that he actually measured them
before making a table.

Here are some values I have measured on an English Weber Grand pedal/
electric and an upright Steck Pedal/Electric Export Model:

Table A. - Grand Piano, Acc Pressure levels:

  #0 = 8"            #8 = 23
  #1 = 9             #9 = 26.5
  #2 = 10           #10 = 29
  #3 = 12           #11 = 32
  #4 = 13           #12 = 35
  #5 = 16           #13 = 37
  #6 = 19.5         #14 = 38
  #7 = 21           #15 = 39.5

Table B - Upright Acc Pressure levels:

  #0 = 4.5"          #8 = 19.5
  #1 = 6             #9 = 23.5
  #2 = 7.5          #10 = 26
  #3 = 9.5          #11 = 30
  #4 = 10.5         #12 = 31.5
  #5 = 13.5         #13 = 34
  #6 = 15.5         #14 = 36.5
  #7 = 19.0         #15 = 38

All measurements done with the spill pressure at +/-30".  The spill
valve muffler felt density does influence these readings.  This is why
I asked my original question: What is the correct spill pressure at
level 0?

If we now graph all these measurements, we find that both my tables as
well as Mike Kitner's table show the pressure not  incrementing in a
linear line as mentioned by Craig but in a slightly exponential manner
between level 2 and 12.

I read somewhere that this is due to the positive feedback into the
system by the spill valve as it closes to increase the spill pressure
from level to level.

As we can see, my measurements come reasonable close to Mike Kitner's
table and we must not forget that no two pianos can be exactly the same
so I am satisfied that my adjustments are corresponding to Mike
Kitner's table.

The problem now is, what if Mr. Kitner's table is wrong to begin with?
He also mentions a spill pressure at zero of 30-35". This spill
pressure changes the  response curve and so does the spring tension on
the regulators although not very much.

Can any third party please shed some light on this subject, preferably
with some pressure measurements so that we can come to a reliable
standard guideline which we can put into our service manuals for future
use.

I always believed that the experts in the States have a great advantage
over people in distant locations like me but the longer this group
exists, the more I come to the conclusion that a lot of experts from
the States do not put too much thought into things anymore. They have
become sloppy.

It was always stressed that one should never  adjust a reproducing
piano from scratch without a precision vacuum gauge. This is common
knowledge or what?  However, what does a fancy gauge help if nobody has
the correct pressure levels or at least the precise shape of the
pressure curve required?  As I mentioned in an earlier bulletin, these
pianos are open-loop systems and for them to produce accurate musical
variations, the adjustments must be as accurate as possible.

Furthermore, why is Mr. Kitner's table wrong?  Has nobody ever read it in
20 years?  Is there any accurate information out there?  Craig stresses
the fact of testing and measuring everything, why does he not do what
he preaches?

Lets have some answers please

Regards,

Bernt W. Damm
Cape Town, South Africa


 [ Editor's note:
 [
 [ What do you name the elements of a set of 16?  "A,B,C.." is only
 [ one method for a set of 'ordered' elements.  Both 1..16 and 0..15
 [ have been used for decades in computers, and programmers still
 [ become confused!
 [
 [ Bernt's issue is not with "mathematically correct" -- but he
 [ wants a standard nomenclature for referring to the "16 ordered
 [ levels of Duo-Art intensity."  We 'modern kids' are more-or-less
 [ comfortable with expressing values as weighted binary (8-4-2-1),
 [ which leads to the number set 0..15.  On the other hand, I can
 [ imagine that a Duo-Art test bench at the factory had marks by
 [ the water-tube manometer going from #1 to #16, and that worked
 [ fine for the factory technicians.
 [
 [ What do our engineers and player techs recommend, and, most important
 [ for Bernt, how shall the discrepancies in nomenclature be resolved?
 [ This must be laid to rest before more discussion of graphs and
 [ empirical readings continue, else the confusion will grow to
 [ utter chaos and disgust!
 [
 [ -- Robbie