I use an electronic differential pressure gauge which gives an output
of 5 volts for a differential pressure of 500 mm of water. This gauge
uses the sensor MPX10DP from Motorola (pressure range: 1.5 psi or
1 meter of water, overpressure to 15 psi).
The main advantage of this kind of gauge is its fast response (about
one or two milliseconds) -- it's possible to measure working of a model
of valve, etc., with an oscilloscope, a graphic recorder, or (preferably)
a PC card and adequate software. I used this gauge and an old
oscilloscope to check tiny electro-valves to MIDIfy my player piano.
It's not difficult to make (a student in electronics made it for me
and gave me the schema, so I cannot guarantee it), nor to adjust (with
a voltmeter and a water gauge). Other schemas are available in
manufacturers documentation (Motorola, Honeywell) or in electronic
magazines. They don't differ notably from barometers or gauges for
tyres (except the sensor).
The schema of the gauge assumes a temperature compensation (according
to Motorola documentation) and an output of zero volts when pressure
is zero. (The zero point and sensitivity are adjustable by
This schema is improvable: the two outputs connections are "floating"
in relation to the alimentation [earth potential], therefore I _must_
use a small sector adapter [power supply module (or a battery)] with
output insulated from ground, but it works.
I'm sure that water gauges are the most accurate -- we can reasonably
think that the density of water won't fail us. :-)
[ The transducer element is a small diaphragm of quartz or ceramic upon
[ which a strain-sensitive conductor is deposited. When the diaphragm
[ is flexed due to a pressure differential, the strain gage resistance
[ changes; this resistance change is converted to a DC voltage and
[ amplified. The first "drum head" resonance of the diaphragm is in the
[ kilohertz region, but the amplifier typically limits the frequency
[ response to around 1 kHz. It's a useful tool. -- Robbie