3.

CALIBRATION OF A PRESSIJRE GATJGE

Introduction
Many types of gar-rge ale available for measurement of pressure. The most simple
folm is a manometer tube, in which the rise of'level of a liquid indicates the static
head, this being converted to pressule by multiplying by the liquid density. An
exatnple of a much m,r're sophisticated instrument is a pressure transducer, n which
the pressure is used to deflect a diaphragm. The deflection causes an electrical signal
to be genelated by some rleans such as a electric resistance strain gauge, and this
signal is displayed, typically in digital forrn, as the corresponding pressllre. The
response is lapid, being typically 1 ms, and the display can be rernote fi'om the point
of measulement. The Bourdon gauge (named after its inventor Eugene Bourdon) uses
the deflection of a tube of oval cross-sectiou to cattse a pointer to move over a scale.
Its response time is therefore long, being of the order of 1 second. Moleover, the
distance between the measur"ing point and the gallge is limited by the practicable
length of the capillary line connecting the gallge to the sensing point. Nevertheless,
because of its simplicity and low cost, and the large selection of pressure ranges
which are available, the Bourdon gallge is widely used ir-r engineering practice.
All pressule gauges, of whatever type, need to be calibrated. If the required accuracy
is low, then a standald calibration obtair-red fi'om a sample of the particular model will
suffice. For higher accuracy, a manufacturer will take special caLe, and will supply a
calibration certificate for an individual gauge. As the calibration rnay change over a
pedod, repeat calibrations may weil be needed flom tin're to time. For the highest
accllracy, transducels and gallges are sometimes calibrated before each use. The
normal calibration procedure is to load the gauge witl-r known pressllres, using a dead
weight tester using oil. The present experiment, however, works satisf'actorily with
water instead of oil.
Description of Apparutus
The Bourdon presslll'e gallge shown in Fig 3.1 has a transparent dial through which
tl-ie construction may be viewed. It consists essentially of a thin-walled tube of oval
closs-section, which is bent to a circular arc erlcolrlpassing approximalely 270'. It is
rigidly held at one end, whele the pressure is aclnitted. The other end is free to rnorre
l3
and is sealed. When
pfessllre is applied, the tube tends to straighten, so that the free
end moves slightly.
This movement
operates a mechanism
which drives a pointer
round the
graduated dial, the movement
of the pointer being proportior-ral
to the
applied
pressure. The construction
of the dead weiglit tester is also shown in Fig 3'1'
A cylindrical
piston, free to move vertically in a closely-fitting
cylinder,
is loaded
with known weights. The space below the piston is filled with water', and the plessure
is transmitted
by the water to the
gauge under test through a transparent
hose' The
presslrre genelated by the piston is easily found in terms of the total weight supported
ancl the closs-sectional
area of the piston'
Weight
Bourdon
tube
Pressure
gauge
calibrated ir-r kN/rn'
Piston
Transparent
Gauge turned
througit 90o to
show mechanism
Fig 3.1 Apparattts.for
calibration of pressure gauge
Procedure
The weight of the
piston, and its cross-sectional
area, shor-rld be noted'
To fiIl the
cylincler, the piston is removed,
ancl water is por'rred into the cylinder ur-rtil it is full to
tlie overf'low
level. Aly air trappecl
i1 the tube may be cleared by tiltilg and
gently
tappir-rg the apparatus.
In point of fact. a sn-iall amount of air lefl in tlie syster-n
will
not afJ-ect the experitnent,
unless thele is so tluch as to cause the piston to bottotn on
tl-re base of the cylincler.
The pistorr is then leplaced in the cylinder and allowed
to
settle. A spirit level
placed on the platforr-r-r at the top of the pistol may be used to
ensure that the cylinder stauds
quite vertically'
14
Weights are now added in convenient increments, and at each increment the pressure
gauge leading is observed. A similal set of resrilts is then taken with decreasing
weights. To guard against the piston sticking in the cylinder, it is advisable to rotate
the piston gently while the pressure gauge is being read.
Results
Weight of piston
Cross-sectional area
:i
kgf
:9.81
N
:333
mm2
:0.333
x 10-3m2
Total load including piston
weight
True
Pressure
(kN/m2)
Gause Readins
Increasing
Pressure
(kN/m2)
Decreasing
Pressure
1t<N/m2)
(ke0 (N)
1.00
1.s0
2.00
2.s0
3.00
3.40
3.80
4.80
s.80
9.8
14.1
19.6
24.5
29.4
33.4
1a 1
)/.)
47.1
569
29.s
44.2
5 8.9
73.6
88.4
100.2
111.9
t4l.4
n0.9
30
45
61
76
9I
103
1ls
145
t75
30
45
61
76
92
103
115
145
Table 3.1 Tt'ue pressLn'es and gauge readings
Table 3.1 presents a typical set of results. The weights are converted fiom units of
kiloglam-force (kgf) to newton (N) simply by mr-rltiplying by the factor 9.81, andthe
tt"ue pressure follows by dividing this figure by the piston area. For example. in the
last line of the table.
True plessllre
:
,
56'9
,
:
fl09 x103
N/m2
:
170.9 x103pa
(0.:::
,.
1o
3)
:
170.9 kN/rn2
- 170.9 kPa
l5
Fig 3.2 shows in graphical form the gauge readings in terms of true pressure, and Fig
3.3 shows the gauge error, again in terms of true pressllre.
200
Increasing pressure
Decreasing pressure
100 200
True Pressure (kN/m')
Fig 3.2 Pressm"e gaLtge calibration
+
()
Increasing pressure
Decreasing pressure
1 00 200
True Pressure
(kN/m')
Fig 3.3 Pressure gauge error
z
b0

too
e)
q)
u0
d
z
,&+
14
4)i
a0L
16
2.
Discr.rssion of Result
Two different kinds of eror may nonnally be expected in a gauge of this type. First,
there is the possibility of hysteresis due to friction and backlash, so that the gauge will
tend to read lower values when the pressule is increasing than when decreasingl'. Th.
gauge tested here has not more than 1 kN/rn2 of this kind of error. Secoridly, there is
the graduation eror due to the scale being malked off incorrectly. In this gauge, the
graduation error increases fairly steadily fi'om zero to approximately 4 kNim2 at a
reading of around 175 kNim2. This error, of about 2.3o
,
would be acceptable srnall
fo1 many engineering
pulposes, although Bourdon gaLlges with a much higher
accuracy are available for accurate work.
Questions for
Further Discussion
t. What suggestions have you for improving the appalatus?
No correction has been made for the difference in elevation of the piston of the
dead weight tester and of the pressure gallge. If the centre of the gallge were
200 mm higher than the base of the piston, should a colrection be made, and if
so, how big would it be?
3. What alterations would
you make to the dimensions of the piston if it wele
desired to calibrate a gauge with a fr-rll scale reading of 3500 kN/m2, using the
same weights?
i
The hysteresis can be put to good use ir.r the case of the arieroid barometer, which is atl att.uospllelic
pr.rrur. gauge that uses a r.uechanical linkage. Tlie hysteresis in the mechanislll callses the barouretet'
io r.acl low on a rising atmospheric pressure. So, if the baroureter is tapped gently with a firlger or
perhaps better still by au ur.ublella handle, the reacling will be seell to be incteased slightly by the
vibraiion.
Conversely, if tlie reading is seen to fall, the barometric pressure will be declining. The
tappilg therefore
gives not only a better indication of atr-nosphet'ic pressul'e, but also an indication of
whether it is rising or falling.
t7