778

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 3, JUNE 2006

A Low-Cost Noncontact Capacitance-Type Level

Transducer for a Conducting Liquid
Satish Chandra Bera, Jayanta Kumar Ray, and Subrata Chattopadhyay

AbstractIn this paper, an attempt has been made to design a

low-cost noncontact capacitance-type level sensor for a conducting
liquid. The sensor is in the form of a uniform circular cylinder
made of insulating material like glass, ceramic, plastic, etc. The
sensor is connected with the metallic- or nonmetallic-type liquid
storage tank, in which the conducting liquid column is taken as
one electrode, and a noninductively wound short-circuited outside
coil is taken as the other electrode of a variable capacitor. The
change in capacitance due to the change of liquid level is measured
by a modified linear operational-amplifier-based De Sauty bridge
network with adjustable bridge sensitivity. The bridge output
after amplification and rectification may be used to drive a direct
current indicator calibrated in level. The studies have been made
with high-density polyethylene and glass tube sensors separately
in both metallic and nonmetallic storage tanks with tap water as
the conducting liquid, and the experimental results of the static
characteristics of the level sensors with percentage error from
linearity are presented in the paper. These results are found to have
good linearity and repeatability within acceptable limits.
Index TermsCylindrical capacitor, error capacitance, glass
tube level sensors, high-density polyethylene (HDPE) tube level
sensors, liquid level measurement, modified De Sauty bridge,
noncontact capacitive level sensor.

I. I NTRODUCTION

IQUID level in a storage vessel is one of the important

variables [10], [11] that is required to be measured and
controlled in any process industry. The liquid properties [10],
[11], [13], [14] like buoyancy, pressure at a depth, relative electrical permittivity, electrical conductivity, thermal conductivity,
absorption of radiation, liquid surface reflection of sound or
light waves, etc., are related with the liquid level. These properties of liquid are used to design the different inferential types
of liquid level transducers [10][12] such as float, displacer,
pressure sensor, capacitance probe, partially immersed resistance wire probes, etc. Of these, the contact-type level-sensing
probes [10], [11] like float, displacer, capacitance probe, etc.,
have the disadvantage that their characteristic properties with
liquid level may change due to the physical or chemical reaction
between the liquid and the probe material, and hence, their
life period may be limited. The noncontact-type level-sensing
probes [10], [12] like ultrasonic probe, absorption probe, noncontact capacitance probe, etc., may have longer life period,

but they are comparatively costly and require various environmental and experimental precautions during the measurement.
The optical fiber liquid level transducer [5], [6] is a contact-type
transducer, which produces optical pulses when the liquid level
moves along the cladding and uncladding zones of the optical
fiber. This transducer has the further advantage that it can be
used with intrinsic safety in explosive or inflammable liquids.
However, this type of measuring system is much costlier than
the other similar instruments.
In this paper, a low-cost new type of noncontact capacitancetype liquid level measuring technique has been designed and
fabricated. In the conventional noncontact capacitance-type
liquid level transducer [10], [11], the air column between the
conducting liquid level and the sensing electrode of a capacitor
is used as the dielectric. This technique appears to suffer from
the error due to the frequent changes of the dielectric properties
of air as well as due to the low value of permittivity of air and,
thus, appears to have limited applications.
In the proposed technique as described in this paper, the
material of a uniform right circular cylinder made of insulting
material like glass, ceramic, nylon, polyvinyl chloride (PVC),
etc., is used as the dielectric of a cylindrical capacitor, and
thus, the defects of the conventional noncontact capacitancetype level-sensing probes may not arise in the proposed sensor.
Now, the change in capacitance of a capacitive transducer due
to a change in process variable is generally very small. Hence,
various attempts have been made by different investigators
[1][4], [7][9] to accurately measure this change in capacitance. In the present paper, an operational-amplifier (op-amp)based modified De Sauty bridge network has been designed
and fabricated to measure the change in capacitance of the
proposed level-sensing probe. In the conventional De Sauty
bridge network [15], the effect of the stray capacitance between the output leads of the bridge network may produce
error. This defect in measurement may be minimized in the
proposed bridge network with the output nodal points of the
bridge network to be at the same potentials. The experiment
is performed in both metallic and nonmetallic storage tanks
using high-density polyethylene (HDPE) tube and glass tube
level sensors. The experimental results are found to have good
repeatability, linearity, and resolution.
II. P RINCIPLE OF THE S ENSOR

Manuscript received August 17, 2004; revised November 19, 2005. This
work was supported by the All India Council of Technical Education (AICTE),
Ministry of Human Resource Development (MHRD), Government of India.
The authors are with the Department of Applied Physics, University of
Calcutta, Calcutta 700 009, India (e-mail: scb152@indiatimes.com).
Digital Object Identifier 10.1109/TIM.2006.873785

The level-sensing probe consists of a uniform hollow cylinder made of insulating material like glass, ceramic, PVC,
Teflon, etc. This cylinder is connected with a metallic storage
tank as shown in Fig. 1(a).

0018-9456/$20.00 2006 IEEE

BERA et al.: LOW-COST NONCONTACT CAPACITANCE-TYPE LEVEL TRANSDUCER FOR A CONDUCTING LIQUID

Fig. 1.

779

Noncontact capacitance-type level sensor in a metallic storage tank.

The insulating material of the sensing cylinder is taken as the

dielectric of the cylindrical capacitor, whereas the conducting
liquid column inside the cylinder is taken as the grounded
electrode of the capacitor because the liquid is generally stored
in a metallic grounded vessel. Thus, the vessel itself may be
taken as one electrode of the cylindrical capacitor, whereas
an outer double-layer noninductively wound short-circulated
coil may be taken as the other electrode of the capacitor.
Inasmuch as the coil is short-circuited, all the turns will be at the
same potential, and the interwinding capacitance and interlayer
capacitance of the coil may be assumed to be negligible. Moreover, the short-circuited outer layer coil and the short-circuited
inner layer coil have the same conducting material, and hence,
the effect of the inner layer coil on the capacitance of the whole
two-layer coil with respect to the liquid column may be ignored.
Thus, the capacitance between the liquid column and the shortcircuited noninductively wound two-layer coil may be assumed

to be equivalent to the capacitance between the liquid column

and only the outer layer of the coil.
The cross section of the level sensor is shown in Fig. 1(b).
Let the coil be made from a PVC- or enamel-insulated singlestrand copper wire of internal conductor radius r with insulation
thickness t and the internal and external diameters of the
sensing cylinder as d and D, respectively. Now, a single-turn
coil of conductor radius r of the outer layer may be assumed to
be a very small cylinder of height 2r, and a cylindrical capacitor
will be formed between the coil and the inner conducting liquid
column of diameter d, which is equal to the inner diameter
of the level-sensing cylinder. Hence, the capacitance formed
between this coil and inner liquid column may be given by
C1 =

1
1

 

ln  D  +
d

1
2

4r

 


ln  (D+2t)(D+6t+4r)

D(D+2t+4r)

(1)

780

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 3, JUNE 2006

where 1 is the permittivity of the insulating material of the

sensing cylinder, and 2 is the permittivity of the insulating
cover material of the winding wire.
Inasmuch as both ends of the coil are short-circuited, the
capacitances of the different turns of the coil with respect to
the conducting liquid column of height h may be assumed to be
connected in parallel. Hence, the capacitance of the part of the
outer layer of the coil surrounding the liquid column of height
h above the datum level may be given by
Ch1 =

h
C1 .
2(r + t)

(2)

Combining (1) and (2), we have

Ch1 =

(r + t)



1
1

2rh
 D 
1   (D+2t)(D+6t+4r)  .


+
ln
ln 


D(D+2t+4r)

(3)
This capacitance may be assumed to be equal to the capacitance
between the liquid column of height h and the portion of the
noninductively wound two-layer coil, embraced by this liquid
column.
Inasmuch as D
t and D
r, the capacitance of the
short-circuited noninductively wound two-layer coil may be
approximately given by
2r1
  h.
Ch =

(r + t) ln  D
d

(4)

There is another capacitance Co between the metallic storage

vessel or other grounded metallic objects and the coil of the
sensing cylinder. This capacitance may be assumed to be the
parallel combination of three parts, namely 1) a parasitic or
fringe part Cparasitic upper between the metallic vessel or other
metallic objects and the upper part or edge of the coil above the
liquid level, 2) a second parasitic or fringe part Cparasitic lower
between the metallic vessel or the other metallic objects and the
lower part or edge of the coil below the datum level, and 3) a
third part Cair between the metallic vessel and the whole part of
the sensing coil with air as the dielectric in the space between
the coil and the wall of the metallic vessel. Thus, Co may be
given by
Co = Cparasitic upper + Cparasitic lower + Cair .

Ch = kh + Co

(8)

where
k=

21 r
  = constant.

(r + t) ln  D
d

(9)

If the storage vessel be designed to have vertical wall equidistant from the sensing cylinder at all points and the sensing coil
be shielded by a grounded coaxial metal screen surrounding
the sensing coil, then all the components of Co in (5) may
be assumed to be constant under constant ambient conditions.
Thus, the capacitance of a two-layer noninductive coil for a
conducting liquid in a metallic storage tank may be assumed
to be linearly related with level as shown in (8).
Now, if the metallic storage tank be coated inside by slurries
or metallic oxides due to long use, then an error capacitance
Ce1 between its metal body and the conducting liquid column
of height h may be added in series with the sensor capacitance
Ch1 and may be given by
Ce1 =

23 h


 ds 
ln  ds 2t

s

(10)

where ds is the inside diameter of the storage tank without any

coating, ts is the thickness of the deposited slurries, and 3 is
the average permittivity of the slurries.
Similarly, the error capacitance due to the coating of slurries
of thickness tm in the sensing cylinder may be given by
Ce2 =

2
 4 h
 d 
ln  d2tm 

(11)

where 4 is the average permittivity of the coating material in

the sensing cylinder of internal diameter d. Capacitances Ch1 ,
Ce1 , and Ce2 as shown in (4), (10), and (11), respectively, may
be assumed to be connected in series, and hence, the effective
capacitance of the sensing probe with respect to the liquid
column of height h above datum level may be given by
Che =

1
+ Co .
1/Ch1 + 1/Ce1 + 1/Ce2

(12)

(5)
Combining (4), (10), (11), and (12), we have

Capacitance Co may again be assumed to be connected in parallel with the test capacitance Ch1 between the liquid column
of height h and the part of the sensing coil embraced by the
liquid column. Hence, the effective capacitance of the sensing
coil with respect to the liquid column of height h may be
given by
Ch = Ch1 + Co .

or

Che =

r+t
r1

  

ln D  +
d

1
3

2h

 

 
 ds 
 d  +Co
1
ln ds 2t
ln

+


d2t
s
4
m
(13)

or
Che = k1 h + Co

(6)

(14)

where
Combining (4) and (6), we have
Ch =

21 r
  h + Co

(r + t) ln  D
d

k1 =

(7)

r+t
r1

 

ln  D  +
d

1
3

2


 

 ds 
 d .
1
+
ln  ds 2t
ln




d2t
s
4
m
(15)

BERA et al.: LOW-COST NONCONTACT CAPACITANCE-TYPE LEVEL TRANSDUCER FOR A CONDUCTING LIQUID

Fig. 2.

781

Noncontact capacitance-type level sensor in a nonmetallic storage tank.

The value of k1 as shown in (15) is less than that of k as

shown in (9). Thus, (14) indicates that the coating of the slurry
material on the sensing probe and the storage vessel reduces the
sensitivity of the sensing probe, but still, a linear performance
of the probe is maintained.
For a nonmetallic storage tank, the level-sensing cylinder
of insulating material may be fabricated with a noncorrosive
conducting electrode in the form of a plate or a small piece
of rod fixed at the base of the cylinder in contact with the
conducting liquid as shown in Fig. 2.
In this case, the capacitance between the short-circuited
noninductive coil and the base electrode of the level-sensing
cylinder may be given by the same expression as shown in (3),
and under the condition D
t and D
r, it may be given
by the similar expression in (4). Hence, the capacitance of the
part of the level-sensing coil surrounding the liquid column of
height h for a nonmetallic tank may be given by

=
Ch1

21 r
  h.

(r + t) ln  D
d

(16)

Due to the coating of slurries of thickness tm inside the sensing

cylinder, the effective capacitance will be given by

=

Che

r+t
1 r

2h
 D 
1   d 


ln d + 4 ln  d2tm 

(18)

where
k2 =

r+t
1 r

2
 D 
1   d  .
ln  d  + 4 ln  (d2tm ) 

(19)



Co = k3 h + Cparasitic
upper + Cparasitic lower

(20)


Co = k3 h + Cparasitic

(21)

or


where Cparasitic
is the total parasitic component, and k3 is a
constant, which may be very small compared with k2 because

the permittivity of air is very small. Again, Cparasitic
upper
in (20) is due to the capacitance of the upper part of the
sensing coil and the upper edge of the liquid column, whereas

Cparasitic
lower is due to the capacitance between the lower part
of the coil below the datum level and the immersed electrode
lead wire. Hence, the effective capacitance of the sensing coil
with respect to the liquid column for a nonmetallic storage tank
may be given by

(17)

where 4 is the average permittivity of the coating material

deposited inside the sensing cylinderor

Che
= k2 h

Like the metallic tank, there is another capacitance Co con

, similar to the
nected in parallel with the capacitance Che
capacitance Co shown in (5). However, because the storage

of Co will not be
vessel is nonmetallic, the component Cair
constant but may be proved to be directly proportional to the
test liquid level h. Hence, Co may be given by



= (k2 + k3 )h + Cparasitic
Che

(22)



Che
= k4 h + Cparasitic

(23)

or

where k4 = k2 + k3 .

Now, the term Cparasitic
in (23) may be assumed to be
constant for a screened sensing probe under constant ambient

of the sensing probe
conditions, and hence, the capacitance Che
for a nonmetallic storage tank may also be assumed to be
linearly related to the liquid level like the metallic storage tank.

782

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 3, JUNE 2006

Fig. 3. Modified De Sauty bridge circuit for the measurement of capacitance of the liquid level transducer.

III. M EASURING C IRCUIT

Also

A modified De Sauty bridge circuit using op-amp has been

designed, as shown in Fig. 3, to measure the capacitance of the
proposed level transducer.
The bridge nodal points B and D as shown in Fig. 3 are at
the virtual grounds considering the very high open loop gain of
the op-amps A1 and A2. Hence, for a sinusoidal supply voltage
V from a Wein bridge oscillator at a stable frequency f , the
currents in the bridge arms AB, AD, BC, and CD of the bridge
circuit ABCD in Fig. 3 are respectively given by I1 = jV C,
I2 = jV Cx , I3 = V1 /P , and I4 = V1 /Q, where V1 is the
output of the op-amp A1. The current in the feedback circuit of
the op-amp A2 is given by If = Vo /Rf , where Vo is the bridge
output at the op-amp A2.
Now, from Kirchhoffs current law
I1 + I3 = 0

(24)

V1 = jV CP.

(25)

or

I4 + I2 + If = 0

(26)

or
Vo =

jRf
[CP Q(Co + C)] V
Q

(27)

where Co is the capacitance of the level-sensing cylinder for the

datum level of the liquid, and C is its change in capacitance
for a level h above that datum level. Thus, Cx = Co + C.
Now, if the bridge be balanced for this minimum value Co
of the level sensor at the datum level, then CP = Co Q, and
therefore, the bridge output may be given by
Vo = jRf V C.

(28)

Thus, the bridge output is the voltage, which is linearly related

with the change in the capacitance C of the level sensor, and
hence, it is linearly related with the level h. This output voltage
may be converted to a current signal by using the conventional
circuit and may be transmitted to the control room.

BERA et al.: LOW-COST NONCONTACT CAPACITANCE-TYPE LEVEL TRANSDUCER FOR A CONDUCTING LIQUID

783

Fig. 5. Percentage error from linearity of the HDPE and glass tube level
sensors in a metallic storage tank.

Fig. 4. Static response of the HDPE and glass tube level sensors in a metallic
storage tank.

IV. E XPERIMENTAL R ESULTS

The experimental work was carried out in both metallic and
nonmetallic storage tanks separately using two types of the
proposed capacitive level sensor with the normal tap water
as the conducting liquid. One type of the level sensor was
made from HDPE tube having 900 mm length, 40 mm internal
diameter, and 3 mm wall thickness. The PVC-insulated hookup
copper wire was used for the two-layer noninductive winding of
the short-circuited coil of this level sensor. The second type of
level sensor was made of Pyrex glass tube of length 800 mm
with 12 mm internal diameter and 2 mm thickness, and the
super enameled copper winding wire of 18 standard wire gauge
(SWG) was used for the short-circuited coil.
The two layers are uniformly wound on the sensing cylinder
one above the other in the same direction. The starting ends of
the both layers of winding are electrically shorted. The finishing
ends are also electrically shorted and taken as one terminal of
the sensing capacitor. The stray magnetic field induces currents
in the opposite directions, and thus, the self-inductance of the
coil is eliminated.
The metallic storage tank for the experiment is made from
galvanized iron, and each level sensor is connected as shown
in Fig. 1. In the case of the measurement with nonmetallic
storage tank, a metal electrode in the form of copper disk of
diameter of 30 mm and thickness of 2 mm is placed at the
bottom of the HDPE tube level sensor, whereas for glass tube
level sensor, the brass connector between the bottom of glass
tube and nonmetallic tank was used as the electrode.
The HDPE tube level sensor was fabricated without screen,
whereas the glass tube level sensor was fabricated with coaxial
copper screen. The grounded screen has been used to keep
the value of parasitic capacitance constant irrespective of the
distance of the tank from the sensing cylinder as well as to
shield the electrical noise [electromagnetic interference (EMI)
and radio frequency interference (RFI)]. This screen increases
the parasitic capacitance. However, this increase in capacitance

has no effect on the measurement because the developed bridge

is balanced at datum level. A bridge excitation voltage of 5 V ac
at a stabilized frequency of 1000 Hz was made constant in each
experiment, but the bridge sensitivity factor resistance Rf was
arbitrarily selected at different values for different experiments.
The liquid level in each experiment was varied in steps under
the static condition, and the bridge output voltage of each step
was measured by digital 41/2 digit multimeter and graphically
plotted for both the level sensors in metallic and nonmetallic
storage tanks. The static characteristic curves of the developed
sensors for metallic and nonmetallic storage tanks with HDPE
and glass tube level sensors are shown in Figs. 4 and 6,
respectively. From these experimental data, best fit straight-line
graph was drawn in each case, and the percentage error of the
experimental data from this optimum linearity was calculated.
The percentage errors of the level sensors for metallic and
nonmetallic storage tanks are graphically shown in Figs. 5
and 7, respectively.
V. D ISCUSSIONS
The static characteristics of the developed HDPE and glass
tube level sensors in the metallic and nonmetallic storage tanks
as shown in Figs. 4 and 6 are found to be linear with very good
repeatability within a resolution of about 1 mm and appear to
conform to the theoretical equations [(14), (23), and (28)]. The
percentage errors from the best fit linear graph, as shown in
Figs. 5 and 7, are found to be within tolerable limits except
some data around 4050 cm of level for both metallic and
nonmetallic storage vessels in the case of the HDPE tube-type
level sensor only. This may be due to some manufacturing
defects of the sensing cylinder near that level, such as an air
bubble inside the material or nonuniformity in thickness near
that level.
From the measurement data as shown in Figs. 4 and 6, it
is found that there is a wide difference between the bridge
outputs for metallic and nonmetallic storage tanks with the
same sensing probe as well as for the HDPE and glass tube
level sensors with the same storage tank. It is due to the fact
that the experiments were performed independently only to
know the linear nature of the sensing probe using arbitrarily
selected different values of the bridge sensitivity adjustment
potentiometer Rf but with the same 5-V, 1-kHz excitation

784

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 3, JUNE 2006

Fig. 6. Static response of the HDPE and glass tube level sensors in a
nonmetallic storage tank.

Fig. 7. Percentage error from linearity of the HDPE and glass tube level
sensors in a nonmetallic storage tank.

voltage signal. Moreover, the two storage tanks and the two
sensing tubes had different dimensions. Also, the level sensor
sensitivity factors (k1 for the metallic storage tank and k4 for
the nonmetallic storage tank) are different for different storage
tanks and sensing probes. Again, the HDPE and glass tube
level sensors had different dimensions with or without the
screen. Hence, the value of the datum level capacitance Co
was different for different tanks and different level sensors,
and the bridge output near balance was different for different
experiments.
The almost exactly identical results were obtained when
each experiment was repeated several times in both increasing
and decreasing modes. Hence, the repeatability of the sensor
is acceptable for any industrial application of the sensor. The
overall uncertainty in measurement of level with this developed
sensor in both the short-term and long-term bases has been
studied, keeping supply voltage, frequency, and temperature

constant, and has been found negligible with this sensor after
having a large number of repeated tests for measurement of
liquid level carried out in the laboratory. The overall sensitivity
of the developed level sensor has been experimentally found
to be 0.95 pF/mm for glass and 0.65 pF/mm for HDPE tube.
Inasmuch as this sensitivity is satisfactorily high, sensor fouling due to low excitation frequency (1000 Hz) has not been
observed during experimentation. This low operating frequency
also decreases the effect of parasitic capacitances.
The permittivity of the sensing tube material varies with
temperature and operating frequency significantly. Using a
frequency-stable oscillator, the effect of frequency is eliminated
in our present design, but the permittivity correction is required
for applications with a wide range of varying temperature.
However, it has been found experimentally that within the
ambient temperature region (2545 C), the error due to permittivity variation lies between 0.25% and 1.5%. However, this
error due to the variation of permittivity with temperature may
be easily eliminated by using a suitable temperature correction
technique from the knowledge of the variation of the permittivity with temperature.
The modified De Sauty bridge circuit was found to give
more stable results than the conventional De Sauty bridge
circuit used in the capacitance-type level probe. Here, it may
be mentioned that for the bridge setup as shown in Fig. 3, the
circuit common and the storage tank ground should be isolated
from each other. Hence, during the design and experimental
work, proper care should be taken so that this isolation may
not be disturbed by any means; otherwise, the results may be
erroneous. To avoid the incidental or accidental short circuiting
between these two grounds, a noninverting-type op-amp-based
circuit may be used as explained in Fig. 8 in the Appendix.
The metallic screen used as electrical shield of the sensor
does not require any bootstrap type of active guard because the
bridge is balanced at datum level by proper selection of bridge
arm components. The op-amps used in the bridge network
have 1-MHz bandwidth. Thus, the selected operating frequency
(1 kHz) does not produce any bandwidth error.
The use of the metallic sheet instead of the sensing coil is
simpler. The experiment was first attempted by wrapping a
thin copper sheet around the sensing cylinder. However, the
sensitivity was found to be very small with wide deviation from
linearity. This may be due to the possibility of the existence of
an air film capacitance between the copper sheet and the sensing
tube connected in series with the test capacitance. Moreover,
the effective area conductive plate of the sensor is large, and the
eddy current induced in the metallic plate tends to disturb the
electrostatic field pattern inside the dielectric material. Thus,
the noninductively wound coil used instead of the metallic sheet
eliminates this effect.
The basic materials for making this level sensor are simple
glass tube and enameled copper wire. The signal conditioning
circuit involves low-cost semiconductor devices. The construction technique does not involve any high-cost technology compared with existing noncontact level sensors such as ultrasonic
gauge and nuclear absorption level gauge. From the experimental study, the repeatability, linearity, and resolution are satisfactory within the tolerable limit of industrial level measurement.

BERA et al.: LOW-COST NONCONTACT CAPACITANCE-TYPE LEVEL TRANSDUCER FOR A CONDUCTING LIQUID

Fig. 8.

785

Modified bridge circuit for the level transducer of grounded storage tank.

The life of the level sensor developed is expected to be long

compared with other noncontact level gauges because there is
no wear and tear of the sensor during operation. Hence, the
present technique of noninductively wound short-circuited coil
may be treated as a low-cost linear alternative technique of level
measurement of a conducting liquid.

The feedback current of the amplifier A2 is given by

If = (Vo V )/Rf .

(b)

Inasmuch as I1 I3 = 0 and I4 + If I2 = 0, the bridge

output is given by

A PPENDIX

Vo = V [1 + jRf {(Co + C)Q CP } /Q] .

To avoid the incidental and accidental short circuiting between the circuit ground and tank ground, a bridge circuit
diagram as shown in Fig. 8 may be used.
The branch currents in the arms AB, AD, BC, and CD for
bridge supply voltage V are, respectively, given by

If the bridge be balanced for a minimum output at the minimum

datum level for the level sensor capacitance Co , then Co Q =
CP . Hence, the bridge output for any level h above the datum
level is given by
Vo = V [1 + jRf C].

I1 = jV C
I2 = jV Cx

(d)

Hence, the final bridge output after the differential amplifier

circuit is given by

I3 = (V1 V )/P
I4 = (V1 V )/Q.

(c)

(a)

Vo = V jRf C.

(e)

786

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 3, JUNE 2006

Thus, the bridge output is linearly related with the change

in capacitance C of the level sensor. The advantage of this
bridge circuit is that no isolation is required between the circuit
common and tank ground.
ACKNOWLEDGMENT
The authors would like to thank the Department of Applied
Physics, University of Calcutta, for providing the facilities to
carry out this work.

Satish Chandra Bera was born in West Bengal,

India, in 1952. He received the B.Tech., M.Tech.,
and Ph.D. (Tech.) degrees from the University of
Calcutta, Calcutta, India, in 1976, 1978, and 1988,
respectively, all in instrumentation engineering.
He worked as an Instrumentation Engineer with
the Hindusthan Fertilizer Corporation Ltd., India, for
nine years and then joined the University of Calcutta
faculty in 1991. Presently, he has been working as
a Reader in instrumentation engineering at the same
university and is involved in carrying out a number of
research projects on process plant instrumentation, biomedical instrumentation,
process modeling, etc., leading to Ph.D. degrees of a number of scholars. He is
the holder of a number of patents on transducers.

R EFERENCES
[1] A. J. Jaworski, T. Dyakowski, and G. A. Davies, A capacitance probe for
interface detection in oil and gas extraction plant, Meas. Sci. Technol.,
vol. 10, no. 3, pp. L15L20, Mar. 1999.
[2] W. R. Yang, M. R. Brant, and M. S. Beck, A multi-interface level measurement system using a segmental capacitance sensor for oil separators,
Meas. Sci. Technol., vol. 5, no. 9, pp. 11771180, Sep. 1994.
[3] S. M. Herang, R. G. Green, A. Plaskowski, and M. S. Beck, A high
frequency stray immune capacitance transducer based on the charge transfer principle, IEEE Trans. Instrum. Meas., vol. 37, no. 3, pp. 368373,
Sep. 1988.
[4] A. H. Falkner, The use of capacitance in the measurement of angular
and linear displacement, IEEE Trans. Instrum. Meas., vol. 43, no. 6,
pp. 939942, Dec. 1994.
[5] G. Betta, A. Pietrosanto, and A. Scaglione, A Gray code based fiber
optic liquid level transducer, IEEE Trans. Instrum. Meas., vol. 47, no. 1,
pp. 174178, Feb. 1998.
[6] , Microcontroller-based performance enhancement of an optical
fiber level transducer, IEEE Trans. Instrum. Meas., vol. 47, no. 2,
pp. 489493, Apr. 1998.
[7] P. Holmberg, Automatic balancing of linear AC bridge circuits for capacitive sensor elements, IEEE Trans. Instrum. Meas., vol. 44, no. 3,
pp. 803805, Jun. 1995.
[8] D. Marioli, E. Sardini, and A. Taroni, High accuracy measurement techniques for capacitance transducers, Meas. Sci. Technol., vol. 4, no. 3,
pp. 337343, Mar. 1993.
[9] C. Kolle and P. OLeary, Low cost, high precision measurement system
for capacitive sensors, Meas. Sci. Technol., vol. 9, no. 3, pp. 510517,
Mar. 1998.
[10] E. O. Doeblin, Measurement System Application and Design, 4th ed.
New York: McGraw-Hill, 1990.
[11] B. G. Liptak, Process Measurement and Analysis, 3rd ed. Oxford, U.K.:
Butterworth Heinman, 1999.
[12] D. M. Considine, Process Instruments and Control Hand Book, 2nd ed.
New York: McGraw-Hill, 1974.
[13] J. P. Bently, Principles of Measurement Systems, 3rd ed. Singapore:
Longman Singapore Publishers (Pte) Ltd., 1995.
[14] Level Measurement, p. 27, Dec. 1993. Control Instrum. (UK), Survey
Directory.
[15] E. W. Golding and F. C. Widdis, Electrical Measurements and Measuring
Instruments, 5th ed. London, U.K.: Pitman, 1963.

Jayanta Kumar Ray was born in West Bengal,

India, in 1954. He received the M.Sc. degree in
physics and the Ph.D. (Tech.) degree in instrumentation engineering from the University of Calcutta,
Calcutta, India, in 1979 and 2005, respectively.
Since 1984, he has been self-employed as an entrepreneur of an electronic-based instrument manufacturing company, where he is currently the Managing
Director. He is currently involved in carrying out
a number of research projects on the development
of smart sensing techniques, virtual instrumentation,
and supervisory control and data acquisition (SCADA) systems and data
networking for a rural water supply scheme. He has more than 30 years of
engineering design and practical experience in instrumentation and information
technology.

Subrata Chattopadhyay was born in West Bengal, India, in 1965. He received the B.Tech. degree
in electrical engineering and the M.Tech. degree
in instrumentation engineering from the University
of Calcutta, Calcutta, India, in 1991 and 1993,
respectively.
He is currently an Assistant Professor in Electrical
Engineering and Professor-in-charge, P. G. Studies,
National Institute of Technical Teachers Training
and Research, Kolkata, India, under the Ministry of
Human Resource Development (MHRD), Government of India, Calcutta. He is currently involved in research for the investigation
on some new/modified techniques of measurement and control instrumentation.