Measurement and

Instrumentation– ME 3102

Ch1: Basic Concepts

1
 Measurements
• What ever that exists, exists in some amount Standard
• Determination of this amount => Measurement
• Gives a result
Comparison
• Experiment/Testing Measurand Result
• Examining the response of a system to an imposed input
• Measurement: Recording the system’s response
• Directly measured vs Derived quantities
• E.g., Acceleration could be measured and velocity could
be derived from it
• Mechanical measurements
• Motion, Force, Strain, Temperature, Pressure, Acoustics
etc.
• Electrical measurements Input Output
• Resistance, Voltage, Current A system
• In this course focus will be mainly on the measurement of Experimental Measured
mechanical quantities conditions values
• Often requires consideration of electrical aspects
(transducers)
 Instrumentation
Typically measured physical quantities
• Instrumentation
Pressure, Temperature, Flow, Viscosity, Level
• All the equipment used for observation,
measurement and recording of physical Voltage, Current, Resistance, Inductance,
quantities Capacitance
Weight, Force, Displacement, Position,
Velocity, Acceleration, Vibrations
• Instrumentation engineering Chemical composition, chemical properties
• Selection of an appropriate sensor
• Installation of sensors and integrating
them with displays, recorders and
control systems
• Designing process control systems
• Integrating all the sensors and providing
a central monitoring station
• Calibration and maintenance of sensors
3
 The Importance of Measurements
• Provide accurate information on the state of a system
• E.g: All the operating parameters in a chemical plant
• Essential for research and design and development of new products
• Verification of the performance of a design
• Essential for process/plant control
• To minimize the difference between the desired and the current state of a
system
• To monitor the proper performance of a system
• Especially for safety critical systems e.g. nuclear power plants, aircraft
• Must be reliable
➢ Having incorrect information is more damaging than having no information at
all
• E.eg, measured quantities in a car
• Driving speed, Engine RPM, Engine Temp., TPS, MAF, Exhaust O2, Crank
angle, Fuel level, Oil pressure etc.
4
 Methods of Measurement
• Direct comparison with a standard
• E.g., Measuring the length of an
object with a tape measure having
standardized markings
• Measuring the volume of a liquid with
standardized containers
• Sensitive to human perception =>
Less accurate
Measurand Result
Sensor

• Indirect comparison
• Needs a calibrated system Calibration
➢More accurate data
5
 Methods of Measurements
• Deflection method
• Linearity of the scale is important
• E.g.s: Mercury thermometers, Spring weighing scales

• Difference method
• Difference between a known reference and the
measured value Reference
pressure
• Results depend on the refence
• Linearity is important
• E.g., U-tube manometer

6
 Methods of Measurements, Contnd.
• Null method

• Measurement is obtained by matching

the measurand with a reference

➢Attaining the null state

• E.g: Weighing balance, liquid

measurement, Wheatstone bridge etc.

• Sensitivity and zero drift are important

• Linearity is not important

7
 Functional Description of an Instrument
System

• Primary sensing element

• Responds to the changes in the measured parameter
• Variable conversion element: Converts the measurement from one physical form
to another form, while preserving the measurement information
• E.g., Transducer
• Variable manipulation element: Chane the numerical value of the measurement
according to a defined rule
• E.g., signal amplification 8
Example: Pressure Thermometer

9
 Example: Digital Revolution Counter
• Measured variable: Shaft RPM
• Microswitch arm
• Primary sensing element
• Also servers as the variable conversion
element – i.e, shaft rotation to linear
displacement
• Micro switch contacts – variable conversion
• Mechanical to electrical oscillation
• Solenoid – Variable conversion
• Electrical pulses to mechanical
reciprocation of the plunger
• Counter
• Data presentation element
• Also variable conversation from
reciprocating to rotational
10
 Example: Galvanometer
• Used in oscillographs
• Time varying voltage is applied and
passed to the coil suspended in the
field of a permeant magnet
• The current results in magnetic forces
causing angular deflection of the coil
and the mirrors
• Light reflected off the mirrors scribes
the signal on a photosensitive chart
moving at a fixed speed => Voltage vs
time graph

11
 Generalized Measuring System
Auxiliary power (not
Auxiliary power
always)
Indicator
Measurand Signal Recorder
Sensor/Trans
ducer conditioner Controller
Transduced Analog signal
signal Processor
Stage 1 Stage 2 Stage3

• Sensor stage
• Ideally should be sensitive only to the measured variable and insensitive
to the other operating parameters
• E.g., Strain measurement => independent of the temperature
• Pressure measurement => independent of the vibrations/acceleration

12
 Generalized Measuring System
• Signal conditioning stage
• Modify the transduced information so that it is acceptable to the third stage
• Needs carryout signal processing such as amplification, filtering, integration,
differentiation, linearization etc.
• Must be designed to match with the first and third stages
• Readout stage (Stage 3)
• Should provide information in the form sought by a controller/human
• Display in digital form or the displacement of an indicator

Example: Velocity measurement from acceleration

Stage 2
Stage 1 Stage3
Acceler Low-pass Data
Amplifier Integrator
ometer filter acquisition
 Types of Measurement Inputs
➢Static (or quasi static)
• Nearly constant with time
• E.g., Temperature, flow rate
etc. ECG signal

➢Dynamic
• Steady state periodic
• E.g., AC voltage signal,
ECG Speech signa, aperiodic
• Transient aperiodic
• Acoustics, accelerations

14
 Analog vs Digital Signals Analog
• Analog signals

Sensor output
• Analogues to a continuous physical process
e.g., the speed of a car
• Has a value at every instance in time
• Varies smoothly in magnitude vs time
• Most signals from sensors are analog type
Time
• Effected by noise in the transmission systems

• Digital signals
• Insensitive to transmission noise Digital
• Analog signal is sampled at discrete time Base-3
intervals Decimal Binary
• The sampled value is converted to an

Sensor output
1 0.00 000
equivalent binary number with a specified 2 0.71 001
system (e.g., 4,8,12 etc. bit) 3 1.43 010
• n bit system : 2n levels 4 2.14 011
• E.g. Signal with 0-5 V converted to digital with 5 2.86 100
a 4 bit system
• 0 V ~ level 0 =0000 6 3.57 101 Time
• 5V ~ level 16=1111 7 4.29 110
8 5.00 111
 Calibration (static )
• All inputs except for one are kept constant and this one input Input of known Output
magnitude response
is varied over a specific range Instrument
system
• Output varies in some range
• This input-output relationship: Static calibration
• Repeat the procedure by varying other input parameters one
at a time Calibration chart/graph

• The calibrating system should have sufficient accuracy

• Considerably greater than the instrument under calibration

Input magnitude
• A rule of thumb: Calibration system uncertainty 4 times lower
than the instrument under calibration
➢ Hierarchy of standards
• Primary standards: the most accurate
• Eg: NIST standards
Sensor response (V, mA etc.)
• Output response could be linear or non-linear
• Linear response highly desirable
16
 Instrument Characteristics
• Static characteristics
• Sensitivity

• Zero error

• Range

• Resolution

• Linearity

• Drift

• Hysteresis

17
 Sensitivity
• The magnitude of change in the sensor output for a unit change in
the measurand Constant sensitivity
• Slope of the calibration curve Variable sensitivity

Sensor Output

Sensor Output
• Higher sensitivity is usually desirable Δy

• Very high sensitivity

Δx Δx
• Becomes too sensitive to the nominal perturbations in the input Sensitivity=
Δy
• Sensitivity to interfering or modifying inputs is also important Input Input

• Eg: Sensitivity of a pressure gage to temperature

• Relative expansion or contraction of the gage display hardware
• Change in the modulus of elasticity of the sensing element =>
change in sensitivity
• If the ambient temperature is measured
• Zero drift and sensitivity corrections could be made
• If the temp. is unknown
• Expressed as errors in the instrument
• Sensitivity drift determination
• Hold pressure constant and vary the temperature
18
 Linearity
• Linearity: Closeness of the calibration curve
to a straight line
• Linear calibration curve is desirable x
x
• Simple calculation of the output by Calibration curve x

Output
multiplying with a constant x
x
x Max
• A reasonably small amount of non linearity x
deviation
x
due to the inherent errors in measurements x
• Fit a straight line through the calibration data
using linear regression Input
• Definition: The maximum deviation of the
calibration curve from the straight line fit
expressed as a percentage of full scale
• Eg: Linearly = ± A % of F.S.
19
 Range (Span)
• The difference between the maximum and minimum possible measurement
value of an instrument
• Maximum measurement value : Full scale
➢Need to select an instrument with an appropriate range that covers the
minimum and maximum possible value in the measurement input
➢Reduces the measurement error
➢Increases the sensitivity
➢Too high a range – reduced accuracy
• A similar parameter – Dynamic range
• Ratio of the largest to smallest input an instrument can measure
accurately
• Expressed in dB, dB=20log(max input/min input)
• Eg: 60 dB dynamic range => ratio of max to min input ~1000

20
 Zero (Bias) Error
• The measured value for zero input
Ideal: 450
straight line

Measured value
Zero error

True value

Positive zero error Negative zero error 21

 Resolution (Error)
• The smallest change in the input that could cause a
measurable change in the instrument output Resolution

Sensor Output
• Aka, Error in the measurement, Least count
• Often times the error in measurements is a % of the
full scale reading
• E.g., Measured value: y±0.01yfs
• Need to select an instrument with sufficiently low
range that could still measure all the possible Input
variations in the input
Higher resolution
• Depends on the instrument, noise in the system and
the resolution of the digital conversion system

Sensor Output
• E.g., Resolution of the common ruler ~ 1mm
• Resolution of the vernier caliper ~ 0.01 mm

Input 22
 Drift
• Change in the calibration curve
over time

• Zero drift

• Sensitivity drift
• Change in the slope of the
calibration curve over time

➢Requires periodic calibration of

the instruments

23
 Hysteresis
• Not all the energy that is put into stressed parts during
loading is recoverable upon unloading
• 2nd law of thermodynamics – Most process are irreversible
• Deviation in the instrument output depending on the
direction of the input variation
• Usually expressed as a % of the FS reading on input or
output
• Threshold of an instrument
• If the input is increased slowly starting from zero a min
value must be reached before a discernible output is
produced
• The min value: threshold of the instrument
• Possible reasons
• Friction
• Residual memory effects

24
Measurement Errors

25
 Measurement Errors
• A measurement should always have some assessment on the accuracy of the measured data
• Errors are inherent in measurements
• Eg: Electronic noise, environmental disturbances could cause errors even if a sensor is highly
accurate
• Single sample vs multi sample data
• Multi-sample data could be used for estimating errors
• Multiple readings from a single pressure gage => still a single sample data since the inherent
instrument error will get repeated
• If it is possible to accurately determine the error magnitude
• The measured data could be corrected with this error
• Therefore no error at all in the presented result
• The real errors in measurements are unknown or vague
• Uncertainty instead of error
• Our task: determine the uncertainty
• Uncertainty definition: Possible values the error may have
• Since the magnitude of error is always uncertain the correct usage is to call experimental
uncertainty instead of experimental error 26
 Measurement Errors
• Error (ε) = Measured Value – True Value
Parallax error
• True value: Not known generally meniscus error

• How to estimate the measurement error?

• Need to estimate the bounds on the error
with some level of confidence after
measurements are done
• E.g., xm±Δxm (95% confidence
• xm- Δxm ≤ xtrue ≤ xm+Δxm with 95% confidence

𝐴𝑏𝑠𝑜𝑙𝑢𝑡𝑒 𝐸𝑟𝑟𝑜𝑟 = 𝑇𝑟𝑢𝑒 𝑣𝑎𝑙𝑢𝑒 − 𝑀𝑒𝑎𝑠𝑢𝑟𝑒𝑑 𝑣𝑎𝑙𝑢𝑒

𝐴𝑏𝑠. 𝐸𝑟𝑟𝑜𝑟
𝑅𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝐸𝑟𝑟𝑜𝑟 =
𝑇𝑟𝑢𝑒 𝑉𝑎𝑙𝑢𝑒
27
 Error Propagation
• The final parameter of
interest = fn (several
measured parameters) 𝐸
𝐼 =
𝑅1 + 𝑅2 + 𝑅3
• Individual errors in the
measured parameters
Differential pressure flow meter
➢Result in a much larger error
in the final parameter

1/2
2𝑃𝑖𝑛
𝑚ሶ = 𝐶𝑑 𝐴 Δ𝑃
𝑅𝑇𝑖𝑛
28
 Types of Errors
1. Systematic/ Bias errors
2. Precession/Random errors
3. Gross errors/ Blunders
Systematic errors (Bias errors )
• The same magnitude of error for every measurement
• No distribution => statistical methods can not be used for the error
quantification
• Could be estimated by comparing with an accurate standard
• Examples:
• Calibration errors
• Consistently recurring human errors
• Loading (i.e, insertion) errors
• Errors due to resolution 29
 Random errors (Precession errors)
• Different for each successive measurement
• For a large number of measurements have a specific mean value
• With sufficient reading a distribution i.e., clustering around a central point could be
obtained
• Could be quantified with statistical methods
• Examples:
• Errors caused by the disturbances to the equipment
• Errors due to fluctuating experimental conditions
• Errors due to insufficient measuring system sensitivity
Gross errors
• Mistakes during experiments
• Computational errors

• Both bias and random error will be present together in an experiment

• Total error = Bias error + Precession error
• Since the true value is often not known, errors are referred to as uncertainties
30
Errors Graphical Representation

Bias error
Precision error
distribution
Occurrence frequency
Total error

Precision
error

xtrue xmean xm

31
 Precession vs Accuracy
• Precision
• Deals with the random errors
• Accuracy
• Systematic errors

Accuracy  Accuracy ☺
• Uncertainty: Precision ☺ Precision 
• The true value of a measurement is
not known
• The magnitude of the error is not clear
i.e. uncertain
➢Uncertainty : The true values lies in
an interval about the mean value at
a certain confidence level Accuracy  Accuracy ☺
Precision  Precision ☺
32