In response to Arthur Nichols' interesting posting regarding measuring
airflow through primary valves, here's a method I have used.  I needed
to do this when I was setting the throw of the electrically operated
valves I designed for my "system" that allows an Ampico to play from a
computer signal.  I needed a simple arrangement that would give me good
resolution, thereby allowing me to adjust each valve to a predetermined
standard.

In my case I had 16 valves per block, so my measuring device needed to
give a reading for 16 valves simultaneously.  Each measuring device is
simply a glass tube, around 6" to 8" long, with one end in a container
of water, the other end connected to a vacuum supply of around 3" to 4"
water gage.  A tee piece is then fitted between the vacuum supply and
the top of each tube, with the third tube connected to a valve under
test.  My electric valves comprise a disc covering a hole venting to
atmosphere, simulating a tracker bar where holes are opened or closed
by the roll perforations.

In this arrangement, if the tee-piece is blocked, the water level in
the tube shows supply vacuum, if open, the water level drops to virtually
zero.  To obtain a reliable reading when setting the electric valves,
I supplied them with an electrical signal that would cause them to
pulse at around 5-10 times per second.  That is, they would open and
close at this rate, causing the water level in the glass tubes to read
say 1.5" WC (with a tiny bit of wobble).  By adjusting the throw of
each valve to give a particular reading (determined by experiment),
I was able to accurately set each valve to give the best repetition.

A static test (no pulsing) was not as reliable, as the amount of valve
opening is not consistent.  By pulsing the valves, inconsistencies like
this were overcome.  Of course it's not possible to use an electrical
signal to pulse pneumatic primary valves, but there are other ways,
such as using a test roll with repetition tests, or with some sort of
arrangement to cause the primaries to pulse.

I haven't thought about it enough to be sure if this method can be
implemented in a totally pneumatic system, but I think it should be
possible.  My point is, what ever method is used, it must be dynamic
(read "pulsing"), rather than static.

I used this arrangement during the whole time my system was in
production, and no one ever came up with an alternative or better idea.
I'm sure there are other methods, but I can say with confidence that
the method I've described worked reliably, allowing me to accurately
set all valves in less than a few minutes per set.

Peter Phillips
Australia