2017 International Conference on Electrical, Electronics, Communication, Computer and Optimization Techniques (ICEECCOT)
Design of an adaptive soft calibration circuit for
LVDT
Santhosh KV, Akanksha
Department of Instrumentation and Control Engineering,
Manipal Institute of Technology,
Manipal University, India
[email protected] Abstract—This paper presents a soft calibration technique for vibration caused due to chatters in the turning process is
displacement measurement using Linear Variable Differential measured by measuring the displacement of shaft using
Transducer (LVDT). The objective of the proposed work is to resistive technique is reported in [11]. Image processing
design a circuit, which will calibrate automatically the output of techniques are used to compute the displacement in a structure
LVDT based on the range set by the user. The output of LVDT is in order to measure the stiffness in paper [12]. From the
drawn from both the secondary coils, and is acquired on to the literature reported it is clear that many techniques are
system through the DAQ card (NIElvis). The acquired AC signal incorporated to measure displacement in any structure. In the
is first converted to pure DC by the help of rectifiers and filters. proposed we concentrate on using LVDT as displacement
The two secondary coil signals which have been obtained from
transducer because of ruggedness and output sensitivity.
filters are passed through comparators circuits. The gain of
filters and comparator circuits are controlled based on the user A brief survey on the reported work of LVDT is studied to
choice of range. The circuit once design is tested using practical understand its working, behavior, etc. From the literature it is
setup, proposed setup produced accurate results and avoided seen that paper [13], discusses the derivation of model for
repeated calibration which otherwise would be needed whenever LVDT. A design is reported so as to reject interference of
the range of measurement was changed. external magnetic field on LVDT output in [14]. Design of
rectangular coil for use to LVDT is reported for planar
Keywords—adaptive; calibration; Displacement measurement;
measurement in [15]. In [16], discusses the effect of core and
LVDT; Smart sensor
coil material on the performance of LVDT. Design of analog
I. INTRODUCTION circuits for measurement of displacement to be used in LVDT
for improvement of sensitivity and range is discussed in [17].
Measurement of displacement is often carried out in many Neural network based calibration technique for range
applications during the day to day activities of human. Often improvement [18] and adaptation with physical material of
displacement is also used as an indicative measure to find LVDT is reported in [19]. Digital signal processing based
other related physical variables like force, pressure, mass, design of calibration technique is discussed in [20]. Though a
volume etc. There are different kinds of displacement lot of work has been reported in design of calibration circuit/
measurement, which can be selected based on the application. technique in literature, they are limited only to a desired range
A few of these are reported, like in [1] capacitance sensors are of measurement and sooner there is a change in system, the
used to measure the tactile force. Image based analysis is measurement system fails to produce accurate results as it is
reported in [2], for measurement of displacement for a two practically impossible to create two identical sensor. In view
dimensional structure. In [3], mechano-optical transducer is of these the proposed work reports a technique for
used to compute displacement/ inclination angle made by a measurement of displacement using LVDT which can
physical structure. Measurement of distance between fiber automatically adjusts its gain, so that on using different LVDT
probe and reflective plate using optical principles is reported still the measurement is accurate.
in [4]. Measurement of soil travel is measured from sweeps
with different cutting widths by using a 3D simulation model The proposed design consists of software driven
is reported in [5]. In [6], vibration measurement by analyzing calibration technique, which is implemented on a
the measured transverse displacement is discussed. In [7], programmable device. The designed system was tested to
ultrasonic signal is used to compute the elasticity of tissue, validate the performance. Results showed successful
transmission time is computed for travel of ultrasonic wave implementation of the proposed work.
from transmitter to receiver, by knowing the velocity of
ultrasound signal and by measuring the traversal time one can II. LINEAR VARIABLE DIFFERENTIAL TRANSDUCER
compute the displacement. MEMS based force sensors are Linear differential variable transducer is a displacement
used to measure the contact by measuring the deflection is transducer, which produces a change in voltage output for
reported in [8]. Measurement of piston position which is change in displacement. LVDT consists of soft iron core
placed inside a cylinder made of magnetic material is carried which is subjected to change in displacement. Soft core coil is
by measuring the change in inductance across the coil placed encapsulated inside one primary and two secondary coil as
on the exterior of the cylinder in [9]. In [10], piezo-resistive seen in Fig. 1. The secondary coils are connected in a series
sensors are used to measure the force exerted on the tooth by opposition manner so that the measurement can be made for
measuring the displacement due to pressure applied. A both positive and negative displacement.
978-1-5386-2361-9/17/$31.00 ©2017 IEEE
Fig. 1. Schematic of LVDT coil
Fig. 3. Circuit representation of LVDT
IV. PROBLEM SOLUTION
To work on the objective mentioned in the earlier section,
a LabVIEW program is designed for the purposed. The first
stage of the work is to extract the signal from the LVDT
connected to NIelvis, for that we use DAQ Assistant block,
Fig. 2. Circuit representation of LVDT further the AC signal acquired from the LVDT is converted to
DC by using a mean average function as shown in Fig. 4.
The emf produced by the LVDT which is given by eq (1). Front panel of the proposed program is shown in Fig. 5. The
error output acquired is as shown in Fig. 6 and Fig. 7.
(1)
The derived potential (Eo) is proportional to the
displacement of core.
III. PROBLEM STATEMENT
The characteristics of the LVDT are studied by subjecting
it to test by varying the input displacement. The plot of
characteristics from Fig. 3 shows that there exist nonlinear
relation between the input displacement and voltage obtained
from the LVDT. Theoretically LVDT offers a linear output
with regard to given displacement. But in practice it is not the
case as can be seen from the plot of input output
characteristics in Fig. 3. A calibration technique can be used to
linearize the output. Now, due to some if the LVDT is
replaced by a different whose measurement range is difficult, Fig. 4. Block diagram of DC estimation process
present system would produce an error. To overcome the error
we may have to recalibrate the system itself.
On analysis of, LVDT based displacement measurement
system, it is found that the LVDT output is not linear and
output values invariably depends on the input ranges. In this
context this paper proposes a measurement system with the
following objectives:
1. Output should be linear
2. Repeated calibration should not be needed when the
LVDT is replaced.
Fig. 5. Block diagram of DC estimation process
525
Output obtained by the use of signal conditioning circuit
using filters is as shown in Fig. 9 and Fig. 10. In the next stage
neural network logic is used. Input data set is the conditioned
DC output for different values of input displacement, target set
is the displacement in mm which is linear in relation and is
equivalent to the range of measurement. In the proposed work
back propagation network with genetic algorithm having two
hidden layers with 6 and 7 neurons respectively.
Fig. 6. Obtained DC estimate value for a displacement of 3mm
Fig. 9. Output obtained for a displacement of 3mm
Fig. 7. Obtained DC estimate value for a displacement of 5mm
It is seen from the DC estimate graph that the output has a
high level of disturbance which needs to be removed for the
sake of high accuracy and linearity for the purpose a
combination of filters are used. The signal is passed through a
band pass filter followed by a low pass filter. Intermediately to
increase the signal strength an amplifier is included as is seen
from the Fig. 8. The gain and frequency ration of the filter is a Fig. 10. Output obtained for a displacement of 5mm
variable whose values are determined by the range of LVDT
used in the experiment, which is computed by the neural V. RESULTS AND ANALYSIS
network block. The adaptive calibration technique designed for LVDT is
tested for its functionality. For testing different sets of LVDT
with ranges 0 to 20mm, -10 to 10mm, 0 to 5mm , 0 to 10 mm
and -5 to 5 mm is used in proposed work. The designed
technique is tested under practical conditions with physically
connecting LVDT through a NIElvis board. Result as obtained
from the front panel of designed LabVIEW program can be
shown in Fig. 11 and Fig. 12. Results obtained from the
proposed measurement technique are tabulated in Table 1. The
first column of the table is the actual displacement which is
subjected to LVDT, followed by filter output and the output
obtained from the neural network block. Lastly the error
obtained from the measurement system. Figure 13 shows the
input output characteristics of the proposed system, followed
Fig. 8. Block diagram of the designed program in LabVIEW
526
by the plot of error obtained by the proposed system is shown
in Fig. 14.
TABLE I. RESULT POST CALIBRATION
Input DC estimate Measured Error
(mm) (V) (mm)
0.0 0.0302 0 0.00
0.5 0.029034 0.480021 4.00
1.0 0.0278 0.987737 1.23
1.5 0.026456 1.5407 -2.71
2.0 0.025292 2.019918 -1.00
2.5 0.02405 2.530658 -1.23
3.0 0.022999 2.96321 1.23
3.5 0.021576 3.549074 -1.40 Fig. 12. Front panel view when subjected with 12mm of displacement
4.0 0.020409 4.029136 -0.73
10
4.5 0.019348 4.46601 0.76
Measured displacement in mm
5.0 0.017922 5.052469 -1.05
5.5 0.016735 5.541091 -0.75 8
6.0 0.015387 6.095885 -1.60
6.5 0.014478 6.470144 0.46 6
7.0 0.013136 7.022181 -0.32
7.5 0.011969 7.502366 -0.03 4
8.0 0.010737 8.0093 -0.12
8.5 0.009731 8.423436 0.90
2
9.0 0.008529 8.918066 0.91
9.5 0.007206 9.462346 0.40
10.0 0.0059 10.00 0.00 0
0 2 4 6 8 10
Actual displacement in mm
Fig. 13. Plot of input output characteristics of proposed system
5
Error percentage
Fig. 11. Front panel view when subjected with 12mm of displacement -5
0 2 4 6 8 10
Displacement in mm
Fig. 14. Plot of error characteristics of proposed system
527
VI. CONCLUSION force control”, IEEE Transactions on Biomedical Engineering, vol. 63,
no. 7, 2016, pp. 1517-1524.
An adaptive displacement measurement technique was [8] Hardik J Pandya, Jun Sheng, Jaydev P Desai, “MEMS- based flexible
reported in the present work. The objective of the reported force sensor for tri-axial catheter contact force measurement”, Journal of
work was to avoid repeated calibration to be done when the Microelectomechanical systems, vol. 26, no. 1, 2017, pp. 264 – 272.
[9] Satish Chandra Bera, Samik Marick, “A modified technique of
ranges of LVDT used for measurement was varied. Reported displacement measurement of a piston made of magnetic material inside
system was designed on LabVIEW platform using the basic a cylinder”, ISA Transcation, vol. 52, 2013, pp. 567-574.
DC estimate function, followed by the filters and neural [10] Yun Shi, Chaochao Ren, Wei Hao, Min Zhang, Yuxing Bai, Zheyao
network block. The trained technique was subjected to test Wang, “An ultra-thin piezoresistive stress sensor for measurement of
under varying conditions, plot of result, tabulation shows tooth orthodontic force in invisible aligners”, IEEE Sensors Journal, vol.
12, no. 5, 2012, pp. 1090 – 1097.
successful implementation of objective proposed. It was seen [11] Haifeng Ma, Jianhua Wu, Liuqing Yang, Zhenhua Xiong “Active chatter
that the linearity was achieved and system produced a root suppression with displacement – only measurement in turning process”,
mean square of percentage of error of 1.35%, which is Journal of Sound and Vibration, vol. 401, 2017, pp. 255-267.
practically acceptable. [12] Dongming Feng, Maria Q Feng, “Identification of structural stiffness
and excitation forces in time domain using noncontact vision-based
REFERENCES displacement measurement”, Journal of Sound and Vibration, vol. 406,
2017, pp. 15-28.
[13] Debashis Sahu, Siddhartha Hazra, Prajit Nandi, “Modeling of Linear
[1] Guanhao Liang, Yancheng Wang, Deqing Mei, Kailun Xi, Zichen Chen, Variable Differential Transformer”, Int. Conf. on VLSI Design and
“Flexible Capacitive Tactile Sensor Array with Truncated Pyramids as Embedded System, 2016, India.
Dielectric Layer for Three-Axis Force Measurement”, Jr. of [14] Martino M, Danisi A, losito R, Masi A, Spiezia G, “Design of linear
Microelectromechanical systems, vol. 24, no. 5, 2015, pp. 1510-1519. differential transformer with high rejection to external interfering
[2] Chenlong Liu, Yongbo Yuan, Mingyuan Zhang, “Uncertainity analysis magnetic field”, IEEE Transactions on magnetics, vol. 46, no. 2010, pp.
of displacement measurement with imetrum video guage”, ISA 674-677.
Transactions, vol. 65, 2016, pp. 547-555. [15] Narendiran Anandan, Boby George, “Design and development of a
[3] Steiner H, Hortschitz W, Kainz A, Stifter M, Jachimowicz A, Schalko J, planar linear variable differential transformer for displacement sensing”,
Keplinger F, Kohl F, “MOEMS transduce with a non-linear transfer IEEE Sensors Journal, vol. 17, no. 16, 2017, pp. 5298-5305.
characteristics for static displacement application on the example of [16] Adrain Grima, Alessandro Danisi, Alessandro Masi, Nicholas Sammut,
inclination sensor”, Sensors and Actuators A: Physical, vol. 263A, 2017, “Influence of external conductive objective on the performance of an
pp. 727-732. ironless inductive position sensor”, IEEE Sensors Journal, vol. 17, no.
[4] Yuhe Li, Kaisen Guan, Zhaohui Hu, “Fiber optic displacement 14, 2017, pp. 4500-4507.
measurement model based on finite reflective surface”, Optics & Laser [17] Gang Chen, Bo Zhang, Pinkuan Liu, Han Ding, “An adaptive analog
Technology, vol. 84, 2016, pp. 32-39. circuit for LVDT’s nanometer measurement without losing sensitivity
[5] Songul Gursoy, Ying Chen, Bo Li, “Measurement and modelling of soil and range”, IEEE Sensors Journal, vol. 15, no. 4, 2015, pp. 2248-2254.
displacement from sweeps with different cutting widths”, Biosystems [18] Saroj Kumar Mishra, Ganapati Panda, Debi Prasad Das, “A novel
engineering, vol. 61, 2017, pp. 1-13. method of extending the linear range of linear variable differential
[6] Charly Faure, Frederic Ablitzer, Jerome Antoni, Charles Pezerat, tranformer using artificial neural network”, IEEE Transaction on
“Empirical and fully Bayesian approaches for the identification of Instrumentation and Measurement, vol. 59, no. 4, 2010, pp. 979-985.
vibration sources from transverse displacement measurements”, [19] Santhosh KV, B K Roy, “Online implementation of an adaptive
Mechanical systems and Signal processing, vol. 94, 2017, pp. 180-201. calibration technique for displacement measurement using LVDT”,
[7] Muyinatu A Lediju Bell, Shalki Kumar, Lily Kuo, H Tutkun Sen, Iulian Applied Soft Computing, vol. 53, 2017, pp. 19-26.
Iordachita, Peter Kazanzides, “Towards standardized acoustic radiation [20] Ralph M Ford, Robert S Weissbach, David R Loker, “A novel DSP-
force (ARF) – based ultrasound elasticity measurements with robotic based LVDT Signal Conditioner”, IEEE Transactions on
Instrumentation and Measurement, vol. 50, no. 3, 2001, pp. 768-773.
528
