For several years I operated and maintained 43-key and 52-key air
calliopes, made by Tangley and National.  The ubiquitous Rootes blower
had poor oil seals, and so SAE-30 lube oil was sucked into the air
supply and combined with ambient dust.  I always had with me a short
piece of brass shim stock to clean the wind-wheet aperture, which would
be fouled with oil and dirt after and hour or so of playing.

The Tangley and National whistles were voiced for operation above 2.5
psi, I believe, and sounded best at higher pressures.  When I operated
the calliope at Disneyland, in the mid 60s, the blower was fitted with
a puny battery-powered electric motor.  I thought it sounded weak and
wheezy.  The battery charge lasted only about 1/2-hour, enough for one
parade through the park.

I loved the sound the best when the blower was powered by a 5 or 6
horsepower gasoline engine -- then I could crank down the blow-off
valve and run about 3.5 psi (pounds per square inch) pressure.  This
pressure was the maximum which could be applied without overblowing
whistles into harmonics.  I wore aircraft-engine-rated ear protectors,
too, and on a quiet afternoon in the suburbs it could easily be heard
over a quarter-mile away!

High pressure on the valves also helped playing while bouncing around
in a parade, because of the key-force hysterisis.  With high pressure
the keys felt like a harpsichord, but at low pressure they felt like an
old reed organ, and it was difficult to control.  I liked the high
pressure best.

The calliope whistle (pipe) is a simple cylindrical stopped flute
constructed entirely of brass.  The wind-sheet is continuous about the
circumference of the resonator, and the mouth of the whistle has a long
cut-up; this large wind-sheet area is what produces the great sound
intensity.  By careful adjustment of the cut-up distance, the
wind-sheet aperture and wind-sheet angle, the whistle will oscillate
initially at only the third harmonic, and then a moment later the
fundamental tone will slowly get going.  That's what makes the nice
"bark".  Sometimes the aperture plate in the larger whistles is allowed
to move slightly under pressure, thus altering the initial wind-sheet
angle. The steady-state oscillation also has a very high 3rd-harmonic
content.  I always suspected that the "white noise" of the air was
somehow amplitude-modulated by the harmonic structure, too, but I
couldn't prove this in experiments with a Yamaha DX7 synth.

The instruments I worked with were of quite simple construction:
a threaded nipple on the base of the brass whistle passed through a
hole in the sheet iron horizontal deck, and was secured underneath
with a big nut.  A neoprene hose, about 1/2-inch inside diameter,
connected the whistle to its valve in the valve chest under the
keybed.

But the long, large hose required a bit of time to charge up to
operating pressure after pushing the key, and I felt that this reduced
the "bark", or "chiff", sound.  Therefore I always wanted to make a
"tracker calliope", in which the valves would be directly under the
whistles.

One of the Tangley calliopes I operated had a player mechanism for
A-rolls.  Only one or two 10-tune rolls sounded good, though, because
piano arrangements aren't well suited for organs.  The best roll as I
recollect was entitled, "Circus Echoes".  It was tricky, adjusting the
suction regulator for the player, because when the music ended between
songs and the pipes weren't operating there wasn't enough airflow for
the wind motor!

Digest 960208 has a note about a calliope for sale.  This contemporary
replica Tangley has a separate vacuum pump for the player, and an
electric motor for the roll drive.  Other notes concerning calliopes
are in the Digests for 950821, 951030 and 951115.