I am surprised that I missed the thread in MMD about air flow
measurement as this is one of great interest to me.

I use various flow measurements extensively in the player work.  In the
design of my electric valve system I was also interested in fine-tuning
the valves to the actual demands of player pianos throughout their
extensive range.  Real figures regarding flow were necessary, as opposed
to guesses.

I have used three methods successfully and can obtain both relative and
absolute measurements with good repeatability.  I have concentrated on
relative measurements since I was largely interested in matching the
existing components used in actual player systems.  Quantities such as
liters of cubic feet per minute do not indicate much by themselves.

In order to determine absolute measurements, I can calibrate any of the
systems with one of the natural gas meters that I have that indicates
cubic feet of gas used.  I have only done this once, a long time ago as
a check on the linearity of the float type meters that I have.

Flow meters in glass tubes (and plastic tubes as well) are rather easy
to find surplus.  They use a float which is usually a precision sphere
of a material of known density (stainless steel, nylon, glass, sapphire,
pith, etc.) floating on a moving column of air (or other fluid) in a
tapered tube.  The tubes internal diameter is larger at the top of the
tube, therefore the float is moved to a higher position in the tube
(smaller ratio of float surface to open area in the larger diameter)
as the flow increases.  By choosing the correct tube and float, a wide
range of flows can be measured.

Many of my tubes are marked as linear so measurements can be made and
compared in magnitude easily.  The disadvantage of this method is that
only a limited range is covered by any tube/float combination; however,
by changing the air pressure (vacuum or positive pressure) you can
extend the range considerably if you only need to compare one flow
against another.

The two other methods use the principle of pressure drop across a
restriction, as does Phil Dayson's method.  The best restriction to use
is a sharp edged orifice that is carefully made, as gas flow formulas
that describe the flow through such orifices are the simplest and most
widely available.  If you are merely interested in relative comparisons,
a needle valve will do and is easier to set to a variety of flow
measurements.

In order to measure the pressure drop across an orifice you need a
differential pressure gauge or manometer.  A simple water column is the
cheapest and not likely to go out of calibration unless the laws of
physics change.  A low pressure gauge such as a Dwyer "Magnehelic" is
another possibility.  They can be had with ranges as low as 0.5" full
scale, and possibly even lower.  I have one of the 0.5" F.S. gauges and
it can be used to measure the tiniest flow variations.

Another way to measure very low pressure differences is with a
manometer whose tube is inclined.  This produces a much greater travel
of the water (or other fluid) column per unit of pressure difference.
I have such a gauge made by Dwyer that is readable to less than .01"
wci [inches water column].  It is just a piece of acrylic with a
straight hole bored into it.  It has a built-in level which is used
to calibrate it and it has check valves to prevent loss of fluid,
which is a red oil of specified density.

A similar device could be made for very little money.  These gauges
were used to detect restrictions in filter systems and can be found as
surplus items, since this is usually done with electronic methods now.

This brings me to the last and most impressive method, which is just
an electronic variation on the orifice method.  Instead of using a
manometer or mechanical differential pressure gauge, I have also used
an electronic type strain gauge.  You can build your own bridge
amplifier and compensation circuitry, or you can buy one, ready to use,
as an evaluation board for a pressure sensor such as those made by
Sensym.

I bought an evaluation board for the Sensym 1 psi gauge (a little over
$100) which is very easy to use.  I connected it to an inexpensive data
logger that connects to the serial port of a PC.  I can watch a display
of the pressure changes that reflect flow variations and can log these
to the computer for analysis later.  These gauges respond very quickly
and can be used for all sorts of other fun measurements.  A 5 psi gauge
covers the player piano vacuum range nicely but the 1 psi is better for
flow measurement.

One fun use of flow measurement that I used successfully was the setting
of stack valves.  I used a regulated positive pressure, applied to the
top of the outside valve with a sponge gasketed adapter.  A flow meter
was inserted in the supply line.  I could adjust each valve for the
same amount of flow instead of measuring the gap.

Valve gap is not a very reliable method of setting flow since seat and
poppet variations can produce very different average gaps with the same
vertical travel.  In the round valve Duo-Art stack that I rebuilt,
individual valves needed to have their travel set as different as .010"
to produce the same flow at low pressure.

Using a regulated pressure also assures that leather compression is
uniform in the setting process.  This is very difficult to assure if
mechanical methods of lifting the valve are used.  This method could
probably be used easily for adjusting primary valves and I intend to
try it with the Themodist stack that I will be rebuilding in the near
future.

One more very fun flow measurement method is the measurement of bleed
rates.  It can even be done on an intact player, right from the tracker
bar.  I use a test roll to uncover a single tracker hole at a time and
use a sponge-gasketed tube connected to the high side of a flow meter.
A vacuum gauge is used to assure that the stack vacuum is constant for
each measurement.  This is a simple way to determine if the total
effective bleed rate (bleed plus pouch leakage) is even across the
stack.  It also shows up defective pouches very easily.  It is also the
way I initially set the bleeds to the same rates on my Vorsetzer which
is fitted with externally adjustable bleeds.

I have not done so yet but I plan to build a frame for testing
pneumatic cloth for leaks.  Even the most expensive cloth available
today has leaky portions and it would be best to know this in advance.
Leaky cloth is very difficult to find once it is installed.

Spencer Chase
http://www.SpencersErolls.com/