AE303 MEASUREMENT

TECHNIQUES

Lecture 2
GENERAL CHARACTERISTICS OF MEASUREMENT SYSTEMS

Dr. Osman Nuri Şahin

1
Generalized Measurement System

• A necessary part of planning an experiment is to

determine the specifications for the required
measurement systems.
• In any experiment the experimenter seeks to obtain
numerical values for certain physical variables.
• These unknown variables are known as
measurands. Examples of measurands include
temperature, velocity, and voltage. Temperature

2
Generalized Measurement System

• The measurement system senses the measurand

and produces a unique numerical value that
describes the measurand.
• The measurement system can be viewed as
consisting of three subsystems.
Mercury-in-glass thermometer

• The sensing element has a significant physical

characteristic that changes in response to changes
in the measurand.

3
Validity Of Measurement

• It is very important to the experimenter that the output of a

measurement system truly states the actual value of the measurand.
• The experimenter must be convinced that the output of the
measurement system is valid.
• No measurement system is perfect—there will always be some
deviation between the actual value of the measurand and the
measurement system output.
• The smaller the allowed deviation, the more expensive will be the
measurement system.

4
Measurement Error And Related Definitions

• The error of a measurement is defined as the difference between the

measured value and the true value of the measurand:

• The experimenter can never really know the error of a measurement.

• But can estimate uncertinity.
• The uncertainty is an estimate of the limits of error in the
measurement.
• Narrow uncertainty intervals are usually achieved by using calibrated,
high-quality measuring systems.

5
Measurement Error And Related Definitions

• Errors in experiments generally fall into two categories:

systematic errors (sometimes called fixed or bias errors)
random errors (sometimes called precision errors).
• The distinction between systematic and random errors is shown
graphically in Figure below.

6
Measurement Error And Related Definitions

• The scatter in the data represents random error, and the deviation
between the average of the readings and true value demonstrates the
systematic error.

• The number of readings forming the average must be large enough to

eliminate the effects of random error in individual measurements on the
average. 7
Measurement Error And Related Definitions

Determine if the following errors are of random or systematic type.

• A digital scale, that always shows 0.2 g when no weight is applied.
• Vibration of the needle of an automobile speedometer.

8
 Example 2.1

• In a calibration test, 10 measurements using a digital

voltmeter have been made of the voltage of a battery
that is known to have a true voltage of 6.11 V. The
readings are: 5.98, 6.05, 6.10, 6.06, 5.99, 5.96, 6.02,
6.09, 6.03, and 5.99 V. Estimate the systematic and
maximum random errors caused by the voltmeter.
• Solution: First, determine the average of the 10 readings:

• Then the estimate of the systematic error is computed as follows:

9
Example 2.1

• To estimate the maximum random error, we need to determine the

reading that deviates the most from the average reading. This is the
reading of 5.96 V. The maximum precision error is thus

10
Measurement Error And Related Definitions

• A measuring system is only designed to operate over a specified range of

measurands. For example, a voltmeter may have a range of 0 to 10 V and
would not give a correct response to measurands of -5 or 13 V.

11
Measurement Error And Related Definitions

• Accuracy is a common term used to

specify uncertainty and is defined as
the closeness of agreement between
a measured value and the true value.
• It is most often specified as a
percentage of full-scale output.
• At readings below full scale, the
percent uncertainty in the reading
will be greater.
• This problem with high uncertainty at
the low end of the range is a major
concern in selecting a measuring
system. 12
Measurement Error And Related Definitions

• Accuracy is a common term used to specify uncertainty and is

defined as the closeness of agreement between a measured value
and the true value.
• It is most often specified as a percentage of full-scale output.
• At readings below full scale, the percent uncertainty in the reading
will be greater.
• This problem with high uncertainty at the low end of the range is a
major concern in selecting a measuring system.

13
Measurement Error and Related Definitions

• Precision is another term frequently used to describe a measuring

system or component and characterizes the random error.
• A highly precise measuring system will give the same value each time
it is read, but it may not be very accurate.

Precision vs accuracy 14
Measurement Error and Related Definitions

• Repeatability is the ability of a device to produce the same output

reading when the same measurand is applied using the same
procedure.
• The random error known as repeatability error.
• An important characteristic of a measuring system is the sensitivity,
defined as the ratio of the change in magnitude of the output to the
change in magnitude of the measurand:

15
Measurement Error and Related Definitions

• In mechanical measuring devices, sensitivity is an important and

limiting characteristic.
• In systems with electrical output sensors, the sensitivity can normally
be increased using simple amplifiers.

16
Example 2.2

• An angular-velocity measuring device (tachometer)

can measure the speed of a mechanical shaft in the
range 0 to 5000 rpm. It has an accuracy of ±5% of full
scale. You notice that when the shaft speed is zero, the
device has a reading of 200 rpm. What is the
maximum error that you might estimate in reading a
shaft speed of 3500 rpm?
• Solution: The accuracy specification indicates an uncertainty of

• Thus, all readings have at least this uncertainty. However, there is a zero
offset of 200 rpm. This error is in addition to the accuracy uncertainty.
Thus, the reading might be as much as high.
17
Dynamic Measurements

• If a measurand is unchanging in time and if the measurement system

instantaneously shows an equilibrium response to the measurand, the
measurement process is said to be static.
• However, in the general case, when the measurand is changing in time
and the measuring system does not show instantaneous response, the
measurement process is said to be dynamic.

18
Dynamic Measurements

• The dynamic response of a measurement system can usually be placed

into one of two categories: first order, and second order.
• These categories are based on the order of the differential equation
needed to describe the dynamic response.
• As an example of first order system may be given as single spring.
• Common spring–mass systems are second order—the mechanical
bathroom scale is an example.

19
Order of a Dynamic Measurement System

• For static measurements, the mathematical relationship between the

measurand (input) and the measurement-system output can be
described with algebraic equations.
• For example, for a linear device, the input–output relationship can be
given by

• where y is the measurement-system output, x is the value of the

measurand, and K is the instrument static sensitivity.
• For time-varying measurements, however, the mathematical relationship
between the measurand and the output must be described with a
differential equation.
20
Order of a Dynamic Measurement System

• An ordinary differential equation (ODE):

• In this equation, n is the order of the system, x is the input signal (forcing
function), y is the output signal, and are constant coefficients that
depend on the characteristics of the measurement system.

21
Zero-order Measurement Systems

• If n = 0, Eq. (11.2) reduces to

• Although x varies dynamically, y responds proportionally, and the

measurement process is essentially a static measurement. This is the
ideal measurement process.
• No system is truly of order zero, but some systems approximate this
behavior in some modes of their operation.
• Strain-gage sensors can be considered as zero-order systems for most of
their range of operation.

22
First-order Measurement Systems

• For Eq.(11.2) becomes

• This equation can be rewritten in the form

• K is the static sensitivity, and which has units of time, is called the time
constant. The time constant, , is determined by the physical
characteristics of the measuring system.

23
Step Response of First-order Measurement Systems

• Consider a situation in which both the input and output to the device
are both zero until time .
• At as shown in Figure, the input x suddenly rises to a value

24
Step Response of First-order Measurement Systems

• The output of measurement system is

• Figure shows the variation of the output y

with respect to the time t. At the output
has reached 63.3% of its final value, and
at it has reached 98.2% of its final value.
• Since the larger the time constant of a
first-order system, the longer it will take
the output to approach its final value, the
system should have a small time constant
if rapid response is desired.

25
Step Response of First-order Measurement Systems

• In many cases, the output does not have an initial value of zero.
• If the initial value of y is and the equilibrium value (a long time after
the step change) is , then

26
Ramp Response of First-order Measurement Systems

• In this case both y and x are zero at time . As shown in Figure, starting at
the forcing function x takes the form

where A is the slope of the ramp.

27
Ramp Response of First-order Measurement Systems

• The output of measurement system is

• In most cases, y will not be zero at time zero. If is the initial value of y at
, the input ramp would be and then

28
Example

• Thermocouples are a very common first-order system and serve to

demonstrate the application of first-order dynamic analysis. the time
constant for a thermometer is given by

bare-junction thermocouple

29
Example

• A bare-junction thermocouple has an approximately spherical junction

with a diameter of 0.3 mm. It is used to measure the temperature of
gases in a combustion tunnel. When the flame is ignited, it produces an
approximate step increase of the gas temperature of 900 K. The average
heat-transfer coefficient on the surface of the thermocouple is . The gas
temperature before ignition is 300 K. (junction is made by platinum and
the properties of platinum are and .

30
Example

a) Find the time constant of the thermocouple.

b) After how much time and how many time constants will the
measurement error be less than 1% of the final temperature change?

Solving, we obtain which corresponds to 4.6 time constants.

31
Example

c) If the same thermocouple is used in an aqueous environment in which

the heat transfer coefficient is what will be the thermocouple time
constant?

this time constant is an order of magnitude smaller than the time

constant for the gas phase.

32
Example

Plot of the thermocouple response versus time for both cases

33
Second-order Measurement Systems

• Second-order behavior governs spring–mass mechanical systems as well

as capacitance–inductance electrical systems. Load cells, pressure
transducers, and accelerometers are second-order systems.
• For second-order measurement systems, the value of n in general ODE
is 2

• divide this equation by and to make the following definitions:

34
Second-order Measurement Systems

• K is the static sensitivity. the undamped natural frequency, is the

angular frequency at which the system would oscillate if there were no
damping . is the damping ratio.
• Using these definitions, equation becomes

• For a system described by this equation, the ideal response (response in

the absence of dynamic effects) would be .

35
Second-order Measurement Systems

• Consider the spring–mass–damper system shown in Figure.

• Using Newton’s second law we obtain

36
Step Response of Second-order Measurement Systems

• As with the first-order system, we will take the initial value of y to be

zero and let the forcing function change from to at . The solution
depends on the value of the damping ratio,

37
where,
Step Response of Second-order Measurement Systems

• If , the system will show oscillatory

behavior and is said to be
underdamped.
• If the system will show an asymptotic
response without overshoot and is
said to be overdamped.
• The condition is the dividing line
between these two responses and is
known as critical damping.
• Overall, dynamic effects are a
minimum with a damping ratio of
about 0.7—there is a fast response,
but the overshoot is less than 5%.
38
Sinusoidal Input Response of Second-order Measurement Systems

• In this case, a forcing function is applied that has a sinusoidal variation

with time:

39
Sinusoidal Input Response of Second-order Measurement Systems

• The actual continuing response of the second-

order system to a sinusoidal input is given by

• where the phase angle is

40
Example

• A load cell is sought for monitoring the off-balance bearing load of a

small motor. The load cell will be installed directly under the bearing at
one end of the shaft.The motor is operated in the range 1500 to 6000
rpm and has a mass of 10 kg. Check to see if a low damping ratio load
cell with a range of 0 to 200 N and a full-scale deflection of 0.1 mm is
appropriate for this dynamic application.
Solution: The stiffness (spring constant) of this transducer is

41
Example

• Considering that the motor will be supported by two bearings, the

effective mass for each bearing will be 5 kg.
• The natural frequency of the motor/bearing system will be

• and the maximum forcing frequency will be

• The ratio of the forcing frequency to the natural frequency is then

42
Example

• According to Figure, at low

damping ratios applying this
transducer will result in measured
force amplitudes far larger than
the actual force amplitude.
• Consequently, the recommended
load cell will not be appropriate
for this type of measurement.

43