DYNAMIC CHARACTERISTIC OF MEASURING INSTRUMENTS

The input varies time to time so does the output. The behavior of the system under such conditions is described by its dynamic response. When dynamic or time varying quantities are to be measured, it is necessary to find the dynamic response characteristics of the instrument being used for measurement.

Dynamic Input
Periodic Harmonic Non-harmonic

Transient Random

The periodic signal varies with time and repeats after a constant interval. The input may be of harmonic or non-harmonic type. The transient signal varies non cyclically with time. The signal is of definite duration and becomes zero after a certain period of time. The Random signal varies randomly with with time, with no definite period and amplitude. This may be continuous, but not cyclic.

..dynamic characteristic of measuring instruments

Speed of Response Rapidity with which a measurement system responds to changes in the measured quantity. Measuring Lag Delay in the response to a change in input. It may be due to capacity, inertia or resistance. retardation type response of the system begins immediately after a change in measured quantity. time delay type response begin after a dead time after the application of input.

Dynamic System Response

Input/Output Model of Linear Dynamic Systems Time Response of Dynamic Systems Solutions to Differential Equations
Transient and Steady State Response

Frequency Response of Dynamic Systems Review of Complex Variables Frequency Response Function Gain and Phase Characteristics System Integration

Linear Systems
Input

x1(t) x2(t) A x1(t) + B x2(t)

Complicated Simple Input Inputs

Linear LinearSystem System

Output

y1(t) y2(t)

Complicated Output

Satisfies the Superposition Principal. Can be modeled by Linear Ordinary Differential Equations. Input a sinusoidal signal of frequency f1, the output will be a sinusoidal signal with the same frequency f1.

Generalized Mathematical Model of a Measuring System

The most widely useful mathematical model for the study of measurement-system dynamic response is the ordinary differential equation with constant coefficients.

d n qo d n 1qo dqo an + an 1 n 1 + ...... + a1 + ao qo n dt dt dt d m qi d m 1qi dqi = bm m + bm 1 m 1 + ...... + b1 + bo qi dt dt dt

y (t ) = y P (t ) 1 2 3 + y H (t ) 1 2 3

Particular Solution (Steady State Solution)

H om ogeneous Solution (Transient Solution)

(a

n n 1 D + a D + .... + a 1 D + a o q o n n 1

= a m D m + a m 1 D m 1 + .... + a 1 D + a o q i
With D operator method the complete solution is obtained in two parts

q o = q ocf + q opi
The qocf has n arbitrary constants; qopi has none. These n arbitrary constants may be evaluated numerically by imposing n initial conditions. The solution qocf is obtained by calculating the n roots of the characteristic equation

anD + an 1D
n

n 1

+ ......+ a1D + ao = 0

Real Roots Unrepeated

For each real unrepeated root s one term of solution is written as Cest, where C is an arbitrary constant. Thus, for example roots 1.5, +2.5 and 0 give a solution

C 1e

1 .5 t

+ C 2e

+ 2 .5 t

+ C

Real Root Repeated

For each root s which appears p times, the solution is written as

(C o + C 1 t + C 2 t + .... + C p 1 t
2

p 1

)e

st

Thus, if the roots are 1, -1, +2, +2, +2, 0 ,0, the solution is written as.

(C o + C1 t ) e t + (C 2 + C 3 t + C 4 t 2 )e 2 t + C 5 + C 6 t

Complex Root Unrepeated A complex root has a general form a+ib . They are always in pairs aib. For each such root pair, the corresponding solution is

Ce sin(bt + )
For example -2i4, 3i5 and 0i7 give a solution

at

C o e sin (4 t + o ) + C1e sin (5t + 1 ) + C 2 sin (7 t + 2 )

2t 3t

Complex Root Repeated

Co e at sin (bt + o ) + C1te at sin (bt + 1 ) + C2t 2 e at sin (bt + 2 ) + ....C p 1t

p 1 at

e sin bt + p 1

Operational Transfer Function

qo bm D + bm1D ..........+ b1D+ bo (D) = n n 1 qi a n D + a n1D .......... ... + a1D+ a o

m

m1

Sinusoidal Transfer Function

qo b m (i) + b m 1 (i) ..... + b1i+ b o (i ) = n n 1 qi a n (i) + a n 1 (i) ....... + a1i+ a o

m m 1

Operational transfer function

The frequency response of a system consists of curves of amplitude ratio and phase shift as a function of frequency.

Zero Order Instruments

zeroth order instruments

No mathematical law can exactly represent any system. There no pure resistance. Potentiometer shall have some inductance or capacitance Mechanical loading due to sliding contact.

FIRST ORDER INSTRUMENTS

dqo d n qo d n 1qo an + an 1 n 1 + ...... + a1 + ao qo n dt dt dt d m1qi dqi d m qi = bm m + bm1 m1 + ...... + b1 + bo qi dt dt dt

dq o a1 + a oq o = boq i dt
bo a 1 dq o + qo = qi a o dt ao

First Order Instruments

(D+ 1) q o

= Kq i

bo Static sensitivit y (K) = (output/in put) ao a1 time constant ( ) = sec ao

qo K (D) = qi D + 1

First Order Instruments

xo displacement from the reference mark temperature of fluid in the bulb; Ttf=0 when xo=0 expansion coefficient of the thermometer fluid and bulb glass volume of bulb cross section area of capillary Temperature of measurand heat transfer area of the bulb

Vb Ac Ti Ab

kexVb Ttf xo = Ac

First order instruments: Thermometer Heat in Heat out = energy stored

UA b (Ti Ttf )dt 0 = VbCdTtf

Assumptions The bulb wall and the fluid film on both sides is a pure resistance to heat transfer with no heat storage capacity. U is constant. The film coefficient and the bulb wall conductivity does not change with temperature. Ab is constant but the change may take place by contraction or expansion. No heat loss by thermometer by conduction upstream. Mass of fluid inside the bulb is constant. The specific heat of fluid is constant.

First order instruments: Thermometer

dTt f VbC + UA b Ttf = UA b Ti dt

CA c dx o UA b A c x o = UA b Ti + K ex dt K ex Vb

K ex Vb K= Ac

m/ o C

CVb
UAb

Step Response of First Order Instruments

initial conditions qo= qi = 0 at t = 0 For t>=0; qo=qi
t

(D + 1) q o = Kq is

qi

qocf = Ce

q opi = Kq is
q o = Ce + Kq is
t

Time, t

Step response of first order instruments

Applying initial conditions we get

0 = C + kq is C = Kq is
which finally gives
t q o = Kq is 1 e

qo = 1 e kq is

(in non-dimensional form)

Step response of first order instruments

qo measurement error, e m = q i K
t e m = q is q is 1 e

em = e q is

1. the response is faster for a small value of time constant 2. setting time is the time for the instrument to reach and stay within a tolerance band around the final value

Step response of first order instruments

qi

qis time, t 1 Kqis 2 1< 2

time, t

Step response of first order instruments

t/

t/

Ramp Response of First Order Instruments

q o = 0 qi = & is t q

t0 t0

(D + 1) q o

& is t = Kq
t

qocf = Ce
q opi

d(q is t ) & is t = K q is t & is (t ) = K(1 D) q = Kq dt

& is (t ) q o = Ce + Kq
Applying initial conditions

Ramp response of first order instruments

t & is e + t q o = Kq

Measurement error

em

q is e qo 424 3 + steady state error = qi =1 transient error K e m, ss e m, t

& is q {

Numerical-1
A thermometer is initially at a temperature of 20 oC and is suddenly plunged into a liquid bath, which is maintained at 150 oC. The thermometer indicated 95 oC after the interval of 3 seconds. Estimate the time constant of thermometer. Also indicate temperature after 5 time constants and comment upon this result. Ans: 3.49 s, 149.1 oC

Numerical-2
A thermometer is sudenly subjected to a step input of 200 oC from 0 oC. Calculate the temperature indicated by the thermometer after a time of 1.5 seconds. The thermometer may be idealized as a first order system with a time constant of 2.5 seconds. Would there be any change in the indicated temperature if the thermometer was initially held at 25 oC? Ans: 90.2 oC; 103.95 oC

Numerical-3
A temperature sensitive transducer used to measure the temperature of a furnace has been idealized as a first order system subjected to ramp input. Calculate the time constant of the transducer if the furnace temperature increases at a rate of 0.15 oC/s. The maximum permissible error in temperature measurement is limited to 4.5 oC. Ans: 30 s

Numerical-4
A weather balloon carrying a temperature sensing devise with time constant of 10 s, rises through the atmosphere at 6 m/s. The balloon transmits information about temperature and altitude through radio signals. At 3000 m height, a temperature indication of 35 oC has been received. Determine the true altitude at which 35 oC temperature occurs. It may be presumed that the temperature sensing device is of the first order and that the temperature varies with altitude at a uniform rate of 0.01 oC/m. Ans: 2940 m

Impulse Response of First Order Instruments

KA For 0 < t < T (D + 1)qo = Kqi = T Since, up until time T, this is not different from a step input of size A/T, our initial condition is qo=0 at t=0+, and complete solution is
KA qo = T
t 1 e

valid up to t = T

For t > T

(D + 1)qo = Kqi = 0
q o = Ce
t

Impulse response of first order instruments by imposing initial conditions

KA 1 e C= T Te
For T

T t KA 1 e e qo = T

& 0 as the case is of impulse response

Te

T T KA1 e t t 1 e = KAe L limT0q o = Lim T0 e im T 0 T T Te Te

Impulse response of first order instruments

1 e lim T 0 T

0 = , an indeterminate form 0

Applying LHospitals rule

1 e lim T 0 T

= lim T 0

(1 )e
1

1 =

Thus, finally for the impulse response of a first orderinstrument

qo

t KA = e

Impulse response of first order instruments We take A=1 and T=0.01.

t 100K 1 e qo = t 100K 1 e 0.01 e 0.01 e

0tT Tt

The shape of the pulse is immaterial for a short duration

dq o + q o = Kq dt
0+

(area under qi curve from t=0 to t=0+)

dq
0

+ q o dt = Kq i dt
0 0

0+

0+

(qo 0 qo 0 ) + 0 = K
+

(area under impulse)

qo

0+

K =

Impulse response of first order instruments

( )

qo k

t/

qo/(k/)

qo/(k/)

t/

Impulse response of first order instruments

Frequency Response of First Order Instruments

qo b m (i ) + b m 1 (i ) ..... + b1i + b o (i) = n n 1 qi a n (i ) + a n 1 (i ) ....... + a 1i + a o
m m 1

Ao qo K = (i ) = Ai qi 2 2 + 1

phase angle, = tan

qo (i ) = K 0 o qi

( )

The ideal frequency response (zero order instrument) would have

Frequency Response of First Order Instruments qi is often combination of several sine waves of different frequency

q i = 1sin 2 t + 0.3 sin 20 t

Suppose the is 0.2 sec. Since, this is a linear system, we may use superposition principle to find qo.

qo (i ) = 2 = qi qo (i ) = 20 qi

K tan 1 (2 x 0.2 ) = 0.93K 21.8o 0.16 + 1 K = tan 1 (20 x 0.2 ) = 0.93K 76o 16 + 1

q o = (1)(0.93 K )sin (2 t 21 .8 ) + (0.3)(0.24 K )sin (20 t 76 ) qo = 0.93 sin (2 t 21 .8 ) + 0.72 sin (20 t 76 ) = q i K

Numerical 5 A signal prescribed by the following relation is required to be measured by using first order system having a time constant of 0.1 s. Develop an expression for the corresponding output. Comment on the result.

i = 3 sin 2t + 0.4 cos10t

Frequency Response of First Order Instruments

SECOND ORDER INSTRUMENTS

A second order instrument is one that follows the equation

dq o d 2q o a2 + a1 + a oqo = boqi 2 dt dt
bo static sensitivity, K = ao
a1 damping ratio, = 2 a oa 2

ao undamped natural frequency, n = a2

Second Order Instruments

D 2 D 1 q Kq which gives = + + o i 2 n n
2

qo K operationa l transfer function is (D) = 2 D 2 D qi + +1 2 n n

Second Order Instruments

dx o d xo fi B Ksxo = M 2 dt dt

(MD + BD+ K )x
2 s

= fi

1 K= Ks
n = Ks M

B 2 KsM

Step Response of Second Order Instruments

D 2 2 D + + 1 q o = Kq i 2 n n

qo = 0 at t = 0+ dqo + = 0 at t = 0 dt

q opi = Kq is
D 2 D + +1 = 0 2 n n
2

Step response of second order instruments over damped ( real and unrepeated)

qo = e Kqis 2 2 1

+ 2 1

2 n t + 1

2 1 2 2 1

2 1 n t

+1

critically damped ( real repeated)

qo n t = (1 + n t ) e + 1 Kq is

under damped (complex)

w n t qo e = sin 1 2 n t + + 1 Kq is 1 2

= sin 1 1 2

Step response of second order instruments

Nondimensional step-function response of second-order instrument

nt

Step response of second order system n=10

Step response of second order instruments Frequency of under damped oscillation

d = n 1
n t

Dynamic error in the measurement, em

qo e qi = sin 1 2 n t + q is K 1 2
Steady state error for the second order system for step input is zero.

ess = t em = 0

Step response of second order instruments

The response of second order underdamped system is sinusoid with a decaying amplitude. For = 0

qo = 1 sin t + = 1 cos t Kq is 2
Thus system has constant oscillations. For >1 there are no oscillations but system is highly sluggish in response.

Step response of second order instruments

Step response of second order instruments

Time domain specifications how fast the system moves to follow the applied input? how oscillatory is the system? how long will it take the system to practically reach its final steady state value? Rise Time
Time required by the system to rise from 0 to 100 percent of its final value.
w n t qo e 2 = sin 1 n t + cos + 1 = 1 2 Kq is 1

Step response of second order instruments

rise time, t r =
Peak Time, tp

cos 1 n 1 2

It is time required for the output to reach the peak of time response or peak overshoot. Differentiate the equation and put derivative equal to zero

sin n 1

t =0

n 1 t = 0, ,2 ,3 ,...
2

tp =

n 1
2

Step response of second order instruments

peak overshoot, Mp
is the difference of output and the input at tp
w n t qo e P 1 = sin 1 2 n t + + 1 1 Kq is 1 2

Mp = e

1 2

Step response of second order instruments Settling Time for 2% tolerance band

n t s

1 2
approximate solution is

= 0.02

ts =
for 5% band

n
3

ts =

Terminated Ramp Response of Second Order Instruments

The devices with extremely high natural frequency and very light damping (<0.01) show large overshoot and strong oscillations under the step input.

Step response of lightly damped system

Terminated ramp response of second order instruments

D 2 2 D + + 1 q o = Kq i 2 n n
Since we are concerned with the lightly damped systems, we obtain the solution for only the underdamped case

t 0tT qi = T 1.0 T t dqo = 0 at t= 0+ qo = dt

qo t 2 1 = + e n t sin 1 2 n t + K T n T n T 1 2
0 t T

Terminated ramp response of second order instruments

qo t 2 1 n t e sin 1 2 n t + = + K T nT nT 1 2

) )

2 t T 1 T n 1 + e n ( t T ) sin 1 2 n (t T ) + 2 T 1 n

= 2 tan

1 2

Tt

Terminated Ramp Response of Lightly Damped System

Terminated ramp response of second order instruments

For 0 t T
There is steady state error of size (2/nT) The transient error can be no larger than 1 / t 1 2 n

Thus if =0 (no damping), the steady state error is zero and transient error is a sustained sine wave of Amplitude 1/(nT). Therefore if n is sufficiently large relative to 1/T. The transient error can be made very small even if the damping is Practically nonexistent.

Ramp Response of Second-Order Instruments

D 2 2 D & + = + 1 q K q t o is 2 n n
overdamped
2 2 2 1 2 2 1 + 1 n t 1 + e 2 4 1 & is qo 2q & is t =q 2 n 2 2 + 1 2 2 1 K + 1 n t e + 2 4 1

dq o qo = = 0 at t = 0 + dt

Ramp response of second-order instruments

critically damped

qo n t 2q is w n t & is t =q 1 e 1 + n K 2
under damped
w n t & is qo 2q e 1 & it sin 1 2 n t + =q 2 K n 2 1

2 1 2 tan = 2 2 1

Ramp response of second-order instruments

Ramp response of second order instrument

Ramp response of second-order instruments

Nondimensional ramp response

Impulse Response of Second Order Instruments

D2 2D 2 + + 1 qo = 0 n n
Over damped
+ qo 1 e = KAn 2 2 1

dq 2 d = o; = KAn at t = 0 dt
o

2 1 n t

2 1 n t

Impulse response of second order instruments critically damped

qo n t = n te KA n
under damped

qo = KA n

1 1
2

n t

sin

( 1

nt

Impulse response of second order instruments

Non-dimensional Impulse Response of Second Order Instrument

Frequency Response of Second Order Instruments

qo K (i ) = 2 qi i i +1 + 2 n n
qo K (i ) = 1 1 n
1

qi

2 2 + n

= tan

n n

Frequency response of second-order instrument

 =0 n =0 n = n

M =1 1 M= 2 M=0

=0 = -90 = -180

Frequency response characteristics of second order systems

Resonance Frequency
The frequency at which M has the highest value is known as resonant frequency, when

dM =0 d n

resonant frequency r = n 1 2

Resonance Peak
Maximum value of M, when =1 n

= r = n 1 2 2

Band width
The frequency at which M = 0.707 is called cutoff frequency c Above this frequency the M reduces below 0.707 or (1/2)0.5. The frequency c represents half power point. The band of frequency between zero to cutoff frequency is called bandwidth of the system, b Measurement systems are low pass filters as the amplitude ratio is unity at =0. As the frequency of input signal increases the, the output gets attenuated. Bandwidth is ,therefore, indicative of the satisfactory reproduction of output signal.

Dead Time Element

The dead time element is defined as a system in which the output is exactly same form as input but occurs dt seconds (dead time) later.

q o (t ) = kq i (t dt )
This type of delay is also known as pure delay or transport lag

HIGHER ORDER SYSTEMS

When the order of the governing equation of an instrument of a combination of instruments is high, it is convenient to plot the frequency response of the system by logarithmic plots, known as Bode diagram. The advantage of this method is that the frequency response of a complex system can be obtained by adding the response due to various first and second order terms occurring in the transfer function of the system.

Bode Plot

qi

M11

q1

M21

q2

M31

qo

q1 (i) = M11 qi q2 (i) = M 2 2 q1 qo (i) = M 33 q2

Bode Plot

qo (i) = M1M 2 M 3 (1 + 2 + 3 ) qi
Usually, decibel notation is used to express Decibel value = 20 log M

log M = log M 1 + log M 2 + log M 3

For the first order system, with governing equation

(1 + D )q o = q i
qi = Ai sin t

Bode Plot

Bode Diagram For First Order System

Bode Plot

Ao K = Amplitude ratio, M = Ai 1 + 2 2

K M in decibels = 20log 2 2 1+ 2 2 = 20 log K 10 log 1 + dB Further for For , << 1 M = 0; every decade = 1 M -3dB; of frequency, there is a = 10 M -20dB; reduction of 20 dB. = 100 M -40dB

Bode Plot Similarly the Bode diagram for a second order system with governing equation

D2 2 D 2 + + 1 q o = q i = A i sin t n n

Ao = 20 log M = 20 log Ai

(1 r ) + (2 r ) (1 r ) + (2 r ) } = 10 log {
2 2 2 2 2

Bode Plot

r being frequency ratio, r =

For r>>1, a change of frequency ratio by a decade, changes M by 40 dB.

2 r = tan 2 1 r
1

angle is 90o for r=1 and the range is from zero to 180 degree.

Bode Plot Bode Diagram For Second Order System

Figure For The Problem

PERIODIC INPUT
Non-harmonic Signals

Fourier series for a periodic function qi(t) is

1 2 2 nt q i = a o + a n cos nt + b n sin 2 T T
Where, T is the time period

2 a o = q i (t )dt T T
2

T 2

2 2 nt dt a n = q i (t ) cos T T
2 2 nt dt b n = q i (t )sin T T

Non-harmonic signals For the square wave signal

T qi (t ) = C , - < t < 0 2 T = C ,0 < t < 2

T 0 2 2 ao = Cdt + Cdt = 0 T T 0 2

Non-harmonic signals
T 0 2 2 2 2 nt dt + C cos nt dt = 0 a n = C cos T T T T 0 2

T 0 2 2 2 2 b n = C sin nt dt + C sin nt dt T T T T 0 2 C [2 cos n cos( n )] = n

Non-harmonic signals

4C , for n = 1, 3, 5, ... = n = 0, for n = 2,4,6...

4C 2nt Thus, q i (t ) = sin , n = 1, 3, 5... n T

RANDOM INPUT
Random signal does not have a definite time period or amplitude and has to be described statistically. Only stationary random signals will be discussed here. It is possible to describe such signals statistically over a certain period of time.

Random Signal The statistical properties that are of relevance are : (i) mean or average value of the random signal, (ii) rms value, (iii) mean square spectral density, and (iv) auto-correlation function.

Mean valueof random signal qi (t ) = limT 1 q i (t )dt 2T T

+T

Random Signal

The signal has to be fad in the filter which allows a component of a signal center frequency c, with a small bandwidth to pass through it. Ideally should be as small as possible The filter output is then squared and averaged over a certain time interval. Thus, we get the mean square value of the component of a certain frequency c. By varying the center frequency, the mean square value corresponding to various frequency components is obtained.

Random Signal From the mean square spectral density S() is defined as:

[ q ] S( ) =
2

S() represents the density, i.e. the amount per unit frequency band width of the mean square value.

Random Signal

Consider an instrument with frequency response function M() and input signal

q = q i sint
Output response,

qo (t ) = M ( )qi sin t
T

Mean square value of input signal,

1 2 q i (t ) = q i sin t dt T0
2

Random Signal Mean square value of output signal,

1 2 2 2 q o (t ) = [M( )] q i sin t dt T0
2

= [M( )] x 2 (t )
2

the mean spectral sensitivity of the output signal So() is given by

S o ( ) = [M ( )] S i ( )
2

COMPENSATION
In order to improve the dynamic characteristics of a measuring system, compensation is employed. This involves use of additional elements. First order system compensation Governing equation for thermocouples is

(1 + D )qo = K i (t )

In order to reduce the effective value of time constant, the voltage V1 can be applied to a circuit whose output is V2. The relationship between V2 and V1 can be easily derived

First order compensation

V1 V2 i1 = Ro

V2 = i 3 R = (i1 + i 2 )R
d i 2 = CD(V1 V2 ), D = dt

V1 V2 V2 = + CD(V1 V2 ) R Ro

First order compensation This equation can be written as

V2 (1 + o D ) = V1 (1 + o D )

R Where, = and o = R o C R + Ro

If o is so chosen, by choosing R o and C, that o =

V2 (1 + D ) = V1 (1 + D )

(1 + D)V2 = K i (t )

First order compensation

It is seen that with compensation The effective time constant is . Since <1, the effective time constant is <. Static sensitivity becomes K, which is less than K, the static sensitivity of uncompensated system.

Second Order System

The damping in the second order system affects the output response considerably. Usually, damping in practice is small and may be increased by additional means, like use of compensation network.

Second order system

1 V1 = R o + R + LD + i CD
V2 = V1 iR o
D D 2 +1 + 1 2 1 1 V2 LCD + RCD + 1 = = 2 V1 LCD + (R o + R )CD + 1 D 2 D +1 + 2 2 1 1
2

Second order system

1 =

1 LC

1 =

R 2 L C

2 =

R + Ro 2 L C

The relation between Vi and Xi for the second order system is:

D n

D + 2 1 n

V = kx + 1 i 1

If the compensation circuit is so designed that

1 = n and

1 =

Second order system

D n

D V = kx 2 1 + + 2 2 i 1
2

The effective damping has increased as 2 > 1