Lecture (3)

Sensor Characteristics
(Part Two)

Prof. Kasim M. Al-Aubidy

Philadelphia University-Jordan

AMSS-MSc Prof. Kasim Al-Aubidy 1

3. Computation of Stimulus:
The main objective of sensing is to determine a value of the input stimulus (s) from the
value of the sensor output signal (S). This can be done by two methods.
1. From the inverted transfer function {s = F(S)} or its approximation, or
2. From a direct transfer function {S = f(s)} by use of the iterative computation.

Computation from Linear Piecewise

Approximation:
Consider triangles p1p2p3 and p1p5p2:
Both triangles are similar, a linear equation
is used for computing the unknown stimulus
(sx) from the measured value (nx):

AMSS-MSc Prof. Kasim Al-Aubidy 2

Example: Assume the thermistor is used
to measure temperature from 0C to +60C.
The output count from the ADC can be
modeled by a nonlinear function n(T) of
temperature:

Where;
T is the measured temperature,
T0 is the reference temperature,
R0 is the resistance of the thermistor at T 0

 The inverted transfer function which enables us to compute analytically the input
temperature is;

AMSS-MSc Prof. Kasim Al-Aubidy 3

Sensor Calibration:
 Calibrate the sensor at two temperatures Tc1 and Tc2 in order to find out values of
the constants R0 and β.
 Select two calibrating temperatures in the middle of the operating range as
T0=Tc1= 293.15 K and Tc2= 313.15 K, which correspond to 20C and 40C.
 The thermistor sequentially is immersed into a liquid bath at these two
temperatures and the ADC counts are registered as nc1=1863 and nc2=1078.
 By substituting these pairs into Equ.1, we find the values of R0=8.350 kΩ and
β=3,895 K.
 The second equation can be used for computing temperature from any reasonable
ADC count.
Example: A thermistor is used to measure some
unknown temperature and receive counts nx= 1505.

To check how far this computed temperature 29.12C deviates from that computed
from a “true” temperature, we plug the same nx=1,505 into second equation and
compute tx =28.22C.

AMSS-MSc Prof. Kasim Al-Aubidy 4

Iterative Computation of Stimulus (Newton Method):
Numerical iterative methods for finding roots of this typically nonlinear equation can be
used for calculating the unknown stimulus (s) without the knowledge of the inverse
transfer function.
If a sensor transfer function is f(s), the Newton method prescribes computing for any
measured output value (S) the following sequence of the stimuli values which after
several steps converges to the sought input s.

Example: If a 3rd degree polynomial with coefficients a=1.5, b=5, c=25, d=1 is used to
illustrate the Newton method.

AMSS-MSc Prof. Kasim Al-Aubidy 5

 If the measured sensor’s response is S=22 and our initial guess of the stimulus is
s0=2, then;

 At step 4, the Newton algorithm stops and

the stimulus value is deemed to be s=0.716.
 To check accuracy of this solution, plug this
number into polynomial Equa and obtain
f(s)= S=22.014, which is within the
resolution error (0.06%) of the actually
measured response S=22.

AMSS-MSc Prof. Kasim Al-Aubidy 6

4. Full Scale Input (Span):
Full scale input represents the highest possible input value, which can be applied to the
sensor without causing unacceptably large inaccuracy.
A decibel scale is used with nonlinear response characteristic, represents low level
signals with high resolution while compressing the high level numbers.

5. Full-Scale Output:
Full-scale output is the algebraic difference between the electrical output signals
measured with maximum input stimulus and the lowest input stimulus applied.

AMSS-MSc Prof. Kasim Al-Aubidy 7

6. Accuracy:
A very important characteristic of a sensor is accuracy, which really means inaccuracy.
Inaccuracy is measured as a highest deviation of a value represented by the sensor from
the ideal or true value of a stimulus at its input.

 The deviation is a difference

between the value, which is
computed from the output
voltage, and the actual input
value.
 All runs of the real transfer
functions must fall within the
limits of a specified accuracy.
These permissive limits
differ from the ideal transfer
function line by . The real
functions deviate from the
ideal by δ, where δ.

AMSS-MSc Prof. Kasim Al-Aubidy 8

Inaccuracy rating may be represented;
 Directly in terms of measured value (): It is used when error is independent on
the input signal magnitude. For example, it can be stated as 0.15C for a temperature
sensor.
 In % of the input full scale: Itis useful for a sensor with a linear transfer function.
 In % of the measured signal: It is useful for a sensor with a highly nonlinear
transfer function.
 In terms of the output signal: It is useful for sensors with a digital output format
so the error can be expressed, for example, in units of LSB.

Absolute & Relative Errors:

Accuracy is the capacity of a measuring instrument to give RESULTS close to the
TRUE VALUE of the measured quantity.

AMSS-MSc Prof. Kasim Al-Aubidy 9

7. Calibration Error:
 It is inaccuracy permitted by a manufacturer when a sensor is calibrated in the
factory.
 This error is of a systematic nature and is added to all possible real transfer
functions. It shifts the accuracy of transduction for each stimulus point by a
constant.
 This error is not necessarily uniform over the range and may change depending on
the type of error in calibration.
Example: Two-point calibration of a real
linear transfer function.
Determine the slope (b) and the intercept (a)
of the function using two stimuli, s1 & s2.

AMSS-MSc Prof. Kasim Al-Aubidy 10

8. Hysteresis:
A hysteresis error is a deviation of the sensor’s output at a specified point of the input
signal when it is approached from the opposite directions.

Example: A displacement sensor;

 When the object moves from
left to right at a certain point
produces voltage, which differs
by 20 mV from that when the
object moves from right to left.
 If sensitivity of the sensor is 10
mV/mm, the hysteresis error is
2 mm.

AMSS-MSc Prof. Kasim Al-Aubidy 11

9. Nonlinearity:
 Nonlinearity error is specified for sensors
whose transfer function may be
approximated by a straight line.
 A nonlinearity is a maximum deviation (L)
of a real transfer function from the
approximation straight line.
 There are several ways to specify
 nonlinearity, depending how the line is
superimposed on the transfer function;
1. Use terminal points to determine output
values at the smallest and highest stimulus
values and to draw a straight line through
these two points.
2. Use best straight line, which is a line
midway between two parallel straight lines
closest together and enveloping all output
values on a real transfer function.

AMSS-MSc Prof. Kasim Al-Aubidy 12

10. Saturation:
The sensor exhibits a span-end nonlinearity or saturation if further increase in stimulus
does not produce a desirable output.

AMSS-MSc Prof. Kasim Al-Aubidy 13

11. Repeatability:
Repeatability error is caused by the inability
of a sensor to represent the same value
under presumably identical conditions. (The
same output signal S1 corresponds to two
different input Signals).
It is usually represented as % of FS:

12. Dead Band:

It is insensitivity of a sensor in a specific
range of the input signals. In that range, the
output may remain near a certain value (often
zero) over an entire dead band zone.

AMSS-MSc Prof. Kasim Al-Aubidy 14

13. Resolution:
Resolution describes smallest increments of stimulus, which can be sensed.
The resolution of digital output format sensors is given by the number of bits in the data
word.

14. Output Impedance (Zout):

Output impedance is important to know to better interface a sensor with the electronic
circuit. The output impedance is connected to the input impedance (Zin) of the circuit
either in parallel (voltage connection) or in series (current connection).

AMSS-MSc Prof. Kasim Al-Aubidy 15

15. Sensor Output Format:
 Output format is a set of the output electrical characteristics that is produced by the
sensor.
 The characteristics may include voltage, current, charge, frequency, amplitude,
phase, polarity, shape of a signal, time delay, and digital code.

16. Sensor Excitation:

 Excitation is the electrical signal needed for operation of an active sensor.
 Excitation is specified as a range of voltage and/or current.

AMSS-MSc Prof. Kasim Al-Aubidy 16

17. Sensor Dynamic Characteristics:
A sensor is characterized with a time-dependent characteristic.
Zero-order Sensor:
 It is characterized by a transfer function that is time independent.
 Such a sensor does not incorporate any energy storage devices, like a capacitor.
 A zero-order sensor responds instantaneously. Such a sensor does not need any
dynamic characteristics to be specified.

AMSS-MSc Prof. Kasim Al-Aubidy 17

17. Sensor Dynamic Characteristics:
First-order Sensor:
 A sensor that incorporates one energy storage component is specified by a first-
order differential equation .
 An example: a temperature sensor where energy storage is thermal capacity.

AMSS-MSc Prof. Kasim Al-Aubidy 18

First-order sensor response:

AMSS-MSc Prof. Kasim Al-Aubidy 19

A first order system response is as follows:

where Sm is steady-state output, and t is time.

Substituting t = , we get;

This means that after a time equal to one time constant, the response reaches about 63%
of its steady-state level. Similarly, it can be shown that after two time constants, the
height will be 86.5% and after three time constants it will be 95% of the level that
would be reached at infinite time.
As a rule of thumb, a simple formula can be used to establish a connection between the
cutoff frequency (fc) and time constant () in a first-order sensor:

AMSS-MSc Prof. Kasim Al-Aubidy 20

AMSS-MSc Prof. Kasim Al-Aubidy 21
A second-order Sensor:
It incorporates two energy storage components, and represented by a second-order
differential equation. The relationship between the input s(t) and output S(t) is;

We can express this second-order transfer function as;

AMSS-MSc Prof. Kasim Al-Aubidy 22

AMSS-MSc Prof. Kasim Al-Aubidy 23
AMSS-MSc Prof. Kasim Al-Aubidy 24
18. Reliability:
 It is the ability of a sensor to perform a required function under stated conditions
for a stated period.
 It is expressed in statistical terms as a probability that the device will function
without failure over a specified time or a number of uses.
 The procedure for predicting in-service reliability is the MTBF (mean-time
between-
 failure) calculation.
 The qualification tests on sensors are performed at combinations of the worst
possible conditions. One approach (suggested by MIL-STD-883) is 1000 h, loaded
at maximum temperature. Three goals are behind the test:
1. to establish MTBF;
2. to identify first failure points that can then be strengthened by design changes;
3. to identify the overall system practical life time.

AMSS-MSc Prof. Kasim Al-Aubidy 25

References:
1. Jacob Fraden, “Handbook of Modern Sensors; Physics, Design, and Applications”, Fourth
Edition, Springer Press 2010.
2. Kelley CT (2003) Solving nonlinear equations with Newton’s method, No. 1 Fundamentals of
Algorithms. SIAM, Philadelphia, PA
3. ISO guide to the expression of uncertainty in measurements (1993) International
Organization for Standardization, Geneva, Switzerland
4. Taylor BN, Kuyatt CE (1994) Guidelines for evaluation and expressing the uncertainty of
NIST measurement results. NIST Technical Note 1297. US Government Printing Office,
Washington DC.
5. R. Pallas-Areny and J. G. Webster, 1991, Sensors and Signal Conditioning, Wiley, New York.
6. J. G. Webster, 1999, The Measurement, Instrumentation and Sensors Handbook, CRC/IEEE
Press , Boca Raton, FL.
7. H. R. Taylor, 1997, Data Acquisition for Sensor Systems, Chapman and Hall, London, UK.
8. J. Fraden, 1997, Handbook of Modern Sensors. Physics, Designs and Applications, AIP,
Woodbury, NY.
9. J. Brignell and N. White, 1996, Intelligent Sensor Systems, 2nd Ed., IOP, Bristol, UK

AMSS-MSc Prof. Kasim Al-Aubidy 26