Sensors and Actuators

Sensor Physics

Sander Stuijk
(s.stuijk@tue.nl)

Department of Electrical Engineering

Electronic Systems
4

SENSOR CHARACTERISTICS
(Chapter 2)
5 Resistance

V
 resistance of a material is defined as R 
i
l
 resistance depends on geometrical factors R  
a
 length of wire (l)
 cross-sectional area (a)
l m l
 resistance depends on temperature R    2
a ne  a
 number of free electrons (n)
 mean time between collisions (τ)

 resistor as temperature sensor

 some types have almost linear
relation between temperature t (°C)
and resistance R (Ω)
 example: platinum (PT100) sensor
 
Rt  R0 1  39.08 104 t 
6 Transfer function

 sensors translates input signal to electrical signal

 transfer function gives relation between input and output signal

Rt
R5 (3.01kΩ) Vout  V1
V1 (5V) Rt  R5
Rt, R0 = 100 Ω Vout

 
Rt  R0 1  39.08 104 t 
7 Signal processing

R5 (3.01 kΩ)

R2 (11.8 kΩ)
R1 (11 kΩ)
+
-
V1 (5V) R3 (105 kΩ)
Rt
Vout
R4 (12.4 kΩ)

R5 // Rt R2 // Rt R4
V  Vout  V1 V  Vout
R5 // Rt  R2 R2 // Rt  R5 R3  R4
R2 // Rt
V1
R2 // Rt  R5
V  V  Vout 
R4 R5 // Rt

R4  R3 R5 // Rt  R2
8 Signal processing

Vout

sensor
voltage (v)

output

temperature (°C)
 sensitivity increased from 0.63mV/°C to 6.67mV/°C
 non-linearity has also been decreased...
9 Nonlinearity

 assumption: resistance has a linear dependency on temperature

“ideal” linear transfer function

voltage (V)

error (V)
“real” transfer function

temperature (°C) temperature (°C)

10 Nonlinearity

 assumption: resistance has a linear dependency on temperature

 error can be expressed as deviation from actual temperature
error (°C)

error (V)

temperature (°C) temperature (°C)

11 Nonlinearity

 nonlinearity is the maximal deviation from the linear transfer function

 nonlinearity must be deduced from the actual transfer function or
from a calibration curve

ideal transfer function

error (°C)

nonlinearity

real transfer
function
temperature (°C)
12 Nonlinearity

 nonlinearity is the maximal deviation from the linear transfer function

 nonlinearity must be deduced from the actual transfer function or
from a calibration curve

raw sensor output

R5 (3.01kΩ)

R2 (11.8kΩ)
R1 (11kΩ)
error (°C)

+
-
V1 (5V)
PT100 R3 (105kΩ)
(100Ω)
Vout
R4 (12.4kΩ)

temperature (°C)

 nonlinearity can be reduced with signal processing electronics

13 Nonlinearity

 nonlinearity is the maximal deviation from the linear transfer function

 nonlinearity must be deduced from the actual transfer function or
from a calibration curve

raw sensor output network output

error (°C)
error (°C)

temperature (°C) temperature (°C)

 ~10x reduction in nonlinearity due to signal processing electronics

14 Errors

 errors are deviations from the “ideal”

transfer function

 sources
 nonlinearity
 materials used
 construction tolerances
 aging
 operational errors
 calibration errors
 impedance matching errors
 noise
 ....
15 Errors

 errors are deviations from the “ideal”

transfer function

 types of errors
 static errors: not time dependent
 dynamic errors: time dependent
 systemic errors: errors are
constant at all times and
conditions
 random errors: different errors in a
parameter or at different operating
times
16 Accuracy

 accuracy is a bound on the maximal

deviation of the true input for any output
of the sensor

 example: if the accuracy is ±3°C, the

measured temperature is the true value
±3°C

 accuracy may be represented

 in terms of measured value (Δ)
 in percent of full scale input (%)
 in terms of output signal (δ)
17 Accuracy

 accuracy may be represented

 in terms of measured value (Δ)
 in percent of full scale input (%)
 in terms of output signal (δ)

 LM135 - precision temperature sensor

 sensitivity: +10mV/°C
 range: -55°C to +150°C
 span: 150°C - (-55°C) = 205°C
 input full scale: 150°C
 output full scale: 4.2V
 uncalibrated temp error: ±1°C

 what is the accuracy of this sensor?

18 Accuracy

 accuracy may be represented

 in terms of measured value (Δ)
 in percent of full scale input (%)
 in terms of output signal (δ)

 accuracy of the LM135

 measured value: ±1°C
 percentage: ±1°C/(150°C)∙100 =
±0.7%
 output: ±10mV
19 Errors

 errors are deviations from the “ideal”

transfer function

 sources of errors (seen so far)

 nonlinearity

 other sources of errors

 calibration errors
 repeatability
 hysteresis
 saturation
 dead band
 ...
20 Calibration error

 calibration data is usually supplied by the manufacturer

 calibration error is the inaccuracy permitted by the manufacturer
when calibrating a sensor in the factory

slope error
 measurement
b  
s2  s1 with error

offset error accurate

 measurement
 a  a1  a 
s2  s1
21 Calibration and nonlinearity

 nonlinearity needs to be considered when calibrating sensor

 several calibration methods are used
 use range points (line 1)
 limit span to useful range and use these range points (line 2)
 use tangent of single calibration point (line 3)
 use linear best fit (line 4)

4
22 Errors

 errors are deviations from the “ideal”

transfer function

 sources of errors (seen so far)

 nonlinearity
 calibration errors

 other sources of errors

 repeatability
 hysteresis
 saturation
 dead band
 ...
23 Repeatability

 repeatability is the failure of a sensor to represent the same value

under identical conditions when measured at different times
 source: thermal noise, buildup charge, material plasticity, ...


r  100%
FS
24 Hysteresis

 hysteresis is the deviation of the sensor’s output at any given point

when approached from two different directions

 caused by electrical or mechanical properties

 mechanical friction
 magnetization
 thermal properties
 loose linkages

h  h
25 Example – magnetoresistive sensor

 sensor can be used to measure the position of magnetic objects

 resistivity of magnetoresistive sensor has relation with strength and
position of magnetic field

 sensor moved along X axis

 Hx provides auxiliary field
 variation in Hy is a measure for the displacement
 sensor output voltage V0 follows Hy curve

Hy

Hx
26 Example – magnetoresistive sensor

 sensor can be used to measure the position of magnetic objects

 resistivity of magnetoresistive sensor had relation with strength and
position of magnetic field

 hysteresis error
 too strong magnet or sensor to close to magnet
 Hx exceeds maximal Hx
 dipoles flip
 sensor has hysteresis loop: ABCD

Hy

Hx
36 Static and dynamic characteristics

 static characteristics
 values given for steady state measurement

 dynamic characteristics
 values of the response to input changes

 many sensors have a time-dependent behavior

 output signal needs time to adapt to change in input

 example - LM135 temperature sensor

 voltage step at input
 output needs time to settle
37 Dynamic error

 dynamic error is the difference between the indicated value and true
value of measured quantity when static error is zero
 difference in sensor response when input is constant or varies

 two important aspects

 magnitude of error
 speed of response (delay)

 different inputs considered when analyzing dynamic characteristics

 step (e.g., sudden temperature change)
 ramp (e.g., gradual temperature change)
 sinusoid (e.g., sound waves)

 any real signal can be described as superposition of these signals

38 Transfer function

 input-output behavior of sensor captured with constant-coefficient

linear differential equation (sensor is linear time-invariant system)
 general form linear differential equation

d n y(t ) d n1 y(t ) d 1 y(t ) d m x(t ) d m1 x(t )

an n
 an1 n 1
 ...  a1  a0 y(t )  bm m
 bm1 m 1
 ...  b0 x(t )
dt dt dt dt dt

 y(t) – output quantity

 x(t) – input quantity
 t – time
 ai, bi – constant physical parameters of system

 solution to equation can be computed using Laplace transform

 transfer function of a system is defined as
Y ( s) bm s m  bm1s m1  ...  b1s  b0

X ( s) an s n  an1s n1  ...  a1s  a0
39 Transfer function

 transfer function does not capture the instantaneous ratio of time-

varying quantities
Y ( s) bm s m  bm1s m1  ...  b1s  b0

X ( s) an s n  an1s n1  ...  a1s  a0

 inverse Laplace transform is needed to go back to time domain

dx(t ) d 2 x(t )
y(t )  Ax (t )  B C 2
 ...
dt dt
 Laplace form is convenient for combining transfer functions
 initial condition can be ignored in transformation when all initial
conditions are zero
 this is true for many practical systems
40 Transfer function

 complex sensors combine several transducers and a direct sensor

 combination of transfer functions of all transducers gives transfer
function of complex sensor
transducer direct sensor amplifier

measured kd
quantity kt voltage
s 2 2s ka
s  1  1
n2
n

sensor

measured kt k d k a
voltage
 s 2 2s 
quantity
s  1 2   1
 n n 
41 Zero-order system

 general form linear differential equation

d n y(t ) d n1 y(t ) d 1 y(t ) d m x(t ) d m1 x(t )

an n
 an1 n 1
 ...  a1  a0 y(t )  bm m
 bm1 m 1
 ...  b0 x(t )
dt dt dt dt dt

 many systems are simpler ...

 example – potentiometric displacement sensor
d

(1-α)RT
t
vo
Vr D
αRT vo d
t

 this system is “memory” less d (t )

vo (t )  Vr
y(t )  k  x(t ) D
42 Zero-order system

 general form linear differential equation

d n y(t ) d n1 y(t ) d 1 y(t ) d m x(t ) d m1 x(t )

an n
 an1 n 1
 ...  a1  a0 y(t )  bm m
 bm1 m 1
 ...  b0 x(t )
dt dt dt dt dt

 many systems are simpler ...

 differential equation for zero-order systems

b0
a0 y(t )  b0 x(t )  y (t )  x(t )  k  x(t )
a0
 static sensitivity given by k
 note: S was used before when discussing static characteristics
 S=k

 zero-order system represents ideal or perfect dynamic performance

43 Zero-order system

 zero-order system represents ideal or perfect dynamic performance

 demonstrated with response to step at input

x(t) Ao/Ai
K
ci

t ω
y(t) φ

kci
t ω

step input frequency response

 no dynamic error present in zero-order systems

 none of the elements in the sensor stores energy
44 First-order system

 many systems are not ideal...

 (parasitic) capacitance or inductance
are often present

 example – liquid-in-glass thermometer xo

 input – temperature Ti(t) of
environment
 output – displacement xo of the
thermometer fluid xo=0
 liquid column has inertia (i.e.
transfer function is not ideal) Tf

Ti(t)
45 First-order system

 first-order system contains one energy storing element

 differential equation for first-order system
dy (t )
a1  a0 y(t )  b0 x(t )
dt
 engineering practice to only consider x(t) and not its derivatives
 solve equation to obtain transfer function
a1 dy (t ) b0
 y (t )  x(t ) 
a0 dt a0 
  s  1Y ( s )  k  X ( s ) 
Y ( s)

k
b
k  0 ,  1
a  X ( s) s  1
a0 a0 

 k – static sensitivity
 τ – time constant
46 First-order system

Y ( s) k b a
 , with k  0 ,   1
X ( s) s  1 a0 a0
 static input implies all derivatives are zero
 static sensitivity (k) is the amount of output per unit input when the
input is static (constant)
 time constant (τ) determines the lag of the output signal on a change
in the input signal

small τ large τ
x(t)
y(t)
ci
kci
t t

step at input response at output

47 Example – liquid-in-glass thermometer

 conservation of energy provides relation between

fluid temperature (Tf) and liquid temperature (Ti)
dT f xo
Vb C  UAbT f  UAbTi
dt
 Vb – volume of bulb [m3]
 ρ – mass density of thermometer fluid [kg/m3]
 C – specific heat of thermometer fluid [J/(kg°C)]
 U – overall heat-transfer coefficient across bulb
wall [W/(m2°C)] xo=0
 Ab – heat transfer area of bulb wall [m2]
Tf

Ti(t)
48 Example – liquid-in-glass thermometer

 conservation of energy provides relation between

fluid temperature (Tf) and liquid temperature (Ti)
dT f xo
Vb C  UAbT f  UAbTi
dt
 relation between liquid level (xo) and
liquid temperature (Ti)
K exVb
xo  Tf
Ac
xo=0
 xo – displacement from reference mark [m]
 Kex – differential expansion coefficient of fluid Tf
and bulb [m3/(m3°C)]
 Vb – volume of bulb [m3]
 Ac – cross sectional area of capillary tube [m2]
Ti(t)
 what are sensitivity (k) and time constant (τ)?
49 Example – liquid-in-glass thermometer

 conservation of energy provides relation between

fluid temperature (Tf) and liquid temperature (Ti)
dT f xo
Vb C  UAbT f  UAbTi
dt
 relation between liquid level (xo) and
liquid temperature (Ti)
K exVb
xo  Tf
Ac
 what are sensitivity (k) and time constant (τ)? xo=0
 combining equations gives differential equation
for whole system
Tf
dT f
Vb C  UAbT f  UAbTi 
dt  CAc dxo UAb Ac
  xo  UAbTi
K exVb Ac xo  K ex dt K exVb Ti(t)
xo  Tf  Tf  
Ac K exVb
50 Example – liquid-in-glass thermometer

 what are sensitivity (k) and time constant (τ)?

 combining equations gives differential equation
for whole system xo
CAc dxo UAb Ac
 xo  UAbTi
K ex dt K exVb

 general first-order system

a1 dy(t ) b b a
 y (t )  0 x(t )  k  0 ,   1 xo=0
a0 dt a0 a0 a0
K exVb Tf
 sensitivity [m/°C] k 
Ac
 time constant [s]   CVb
UAb Ti(t)
51 Example – liquid-in-glass thermometer

 what are sensitivity (k) and time constant (τ)?

K exVb
 sensitivity [m/°C] k  xo
Ac
CVb
 time constant [s]  
UAb
 sensitivity and time constant related to physical
parameters
xo=0
 larger sensitivity (k) requires larger bulb volume (Vb)
 larger bulb volume (Vb) increases time constant (τ)
Tf
 effect partially offset by increased contact area (Ab)

 careful selection of parameters is required Ti(t)