Sensor Selection

The quality of a sensor is determined by numerous parameters, many of

which are interdependent and the significance of a particular parameter
depends on the overall system requirements. These parameters can be
specific to the sensing element or the sensor as a whole.
➢ Accuracy
• The accuracy of a sensor is determined by its nearness to the true
value.
• The challenge lies in obtaining the true value, which is typically
determined using standard references in a laboratory setting.
• The sensor is exposed to a known magnitude and frequency, and the
accuracy is calculated as the expected error, considering both
magnitude and frequency variations. After taking measurements, the
relationship between input and output is established in a probabilistic
sense.
• The accuracy of a sensor is calculated as the expected error (in
percentage terms) of the sensor, considering both the magnitude and
frequency variations of the measurand.
➢ Sensitivity
• The sensitivity, when it refers to the sensing element, is the minimum
magnitude of a measurand that can be picked up by the sensor to
produce a corresponding output.
• Sensitivity, when it refers to the entire sensor, is not necessarily the
quality of the sensing element alone but also the quality of the
conditioning circuit and, most importantly, the quality of the
preamplifier.
➢ Zero-offset
• The zero-offset of a sensor is the magnitude of the output when the
measurand is zero.
• It can be expressed in different ways, depending on the sensor’s
output.
• For most electrical sensors, it is expressed in millivolts or milliamps. It
can also be expressed as a unitless quantity, for example, as
percentages of the full-scale output.
• Zero-offset can be corrected by proper calibration
➢ Reproducibility
• Reproducibility refers to the a sensor’s ability to repeatedly yield the
same output (or comparatively the same) for the same input under
the same operational condition. This is an important aspect
particularly for:
• magnetic sensors, because of a potential hysteresis effect
• piezoelectric sensors
• strain gauges, because of the possibility of inelastic characteristics
in the sensing elements.
➢ Span
• The difference between the highest and the lowest magnitudes of a
measurand that can be detected by a sensor within an acceptable
accuracy and without damaging or significantly affecting its
reproducibility is called the sensing span.
• The span of a sensor is not merely a static value, as it is affected by
the frequency response
➢ Stability
• Stability (or long-term stability) refers to the physical changes the
sensor undergoes over time as a result of which there is a slow drift
or shift in its operational condition.
• Stability may also refer to the change in the spectrum of noise to
which the sensor is sensitive
• Factors that affect the stability of a sensor are
• ageing,
• adsorption or desorption of contaminants,
• stress,
• package leaks
• gassing and chemical reactions
➢ Resolution
• Resolution refers to the minimum change in the measurand that can
be detected by the sensor
• The resolution of a sensor is affected by the sensing element itself as
well as all the subsequent stages, including analogue-to-digital
conversion (ADC). At the ADC level, the resolution is affect by the
quantisation error, which in turn is affected by the number of bits
allocated to digitise the analogue input.
➢ Selectivity
• Selectivity is a sensor's ability to respond only to a measurand while
rejecting undesired signals. It is a quality of both the sensing element and
conditioning circuit.
• Difference amplifiers can improve selectivity when the measurand and
interfering inputs have overlapping spectra.
• Clippers and limiters can suppress noise if the measurand is large.
• Analogue and digital filters can separate signals.
• Selectivity becomes part of the sensing element's structure for more
subtle noise sources.
• In chemical and biochemical sensors, selectivity refers to a sensor's ability
to react to specific molecules or compounds while remaining
unresponsive to others.
➢ Response Time
• The duration a sensor requires to approach its true output when
subjected to a step input is referred to as its response time. This
parameter is typically affected by the frequency response of the
sensor.
• An ideal sensor will have a constant (flat) speed of response within its
operational band.
➢ Self-heating
• Self-heating is a significant issue in electrical sensing elements, as it
results from a voltage drop across inductive, capacitive, and resistive
components due to a current.
• This power dissipation, known as self-heating, is problematic due to
temperature-dependent characteristics of sensors, especially when
the element has a significant resistive characteristic.
➢ Hysteresis
• Hysteresis is the challenge a sensing
element faces in reproducing the same
one-to-one relationship between a
measurand and sensor output in both
opposite and forward directions. A
hysteresis-free sensing element
faithfully reproduces this relationship
in both directions. An example of
hysteresis' effect on a sensor's
reproducibility where the sensor
produces two different output curves
for the same measurand values.
➢ Ambient Condition
• Sensors require proper ambient conditions, including temperature,
exposure to magnetic fields, and vibration.
• Sensing devices like ECGs, EEGs, and SQUIDs require shielding from
external magnetic fields and power line signals.
• Thermal conditions can affect the stability of magnetic and optical
sensors.
• Ambient condition refers to the surrounding or environmental
conditions that a sensor should maintain for proper functioning.
➢ Overload Characteristics
• The maximum amount of current or power that can be drawn from a
sensor or the maximum amount of power that can be dissipated
inside it without significantly affecting its operation is referred to as
its overload characteristic
➢ Operating Life
• The operating life of a sensor can be described in many ways,
including in terms of its stability and accuracy. Generally it is the
expected number of hours it can operate without its stability or
accuracy deteriorating beyond a set value.
• In reality, the operating life of a sensor is not so easy to determine, as
it depends on so many factors.
• For example,the operating life of a chemical sensor depends on the
total amount of gas it is exposed to during its lifetime, as well as other
environmental conditions, such as temperature, pressure and
humidity.
➢ Cost, Size, and Weight
• Apart from the functional aspects of a sensor, its non-functional
aspects are important too. These include cost, size, and weight. This is
particularly true for those sensors that have to be seamlessly
embedded into physical bodies or processes.
• Non-functional aspects such as size and weight can have a direct
bearing on some functional parameters, such as self-heating, span,
and sensitivity.
• Micro-sized sensors, for example, can only allow a limited amount of
power dissipation, which in turn limits their span. The same can be
said of weight. Light sensors respond to surrounding heat and
radiation more quickly and more accurately than heavy sensors
Temperature sensor selection
• Resistance Temperature Detectors (RTDs)
• Thermistors
• Thermocouples
• Infrared
 Resistance Temperature Detectors
▪ RTDs are resistive sensors
▪ Resistance varies as a function of
temperature

▪ Can be produced as a wire coil or thin

film
Merits Demerits

• Linear response over a wide • Relatively high cost

range of temperatures • Slow response time
• Most stable, most accurate, and • Poor sensitivity to small
most reproducible temperature temperature changes (low
sensing element resolution)
• Highly resistant to • High sensitivity to external
contamination and corrosion vibrations (due to the
• Does not require recalibration piezoresistive effect)
after fabrication • High self-heating and power
dissipation
• Requires a small amount of
operating current
 Thermistors
▪ Thermistors are made from metal
oxides pressed into a semiconductor
chip
▪ Most have negative temperature
coefficients (NTC)
▪ Resistance decreases as temperature
increases
▪ High temperature coefficients and
high resistance
Merits Demerits

• High resolution and can be • Fragile

fabricated in various sizes and • Limited temperature range
shapes • Accuracy and stability vary along
• Relatively high resistance allows the sensing range
for a small amount of current
flow
• Less self-heating compared to
RTDs
Thermocouples
Merits Demerits
• Measure temperature directly, • Least stable and least
without requiring a biasing current reproducible
or voltage
• Simple to fabricate, rugged, and • Poor resolution
inexpensive • Easily affected by ambient noise
• Widest sensing range of all • Comparatively low accuracy
temperature sensors (from -273°C
to 2700°C)
• Fastest response times
• Enable point temperature sensing
• Do not exhibit problems associated
with contact resistance
 Infrared
A heat-producing object radiates energy in the form of infrared light.
A one-to-one relationship between the infrared energy and the temperature
of the object can be established by setting up an infrared detector.
The detector, for example, can be a photodiode or a phototransistor that
generates an electrical voltage proportional to the light it absorb
Merits Demerits

• No contact needed between the • High set-up costs

sensor and the object • More complex than other sensor
• Comparable response times to technologies
thermocouples (or even faster) • Require optoelectronic
• Good stability and high components
reproducibility • Accuracy can be affected by
• Not affected by corrosion or emissivity variations
oxidation • Field of view and spot size can
• Accuracy does not deteriorate limit usability
over time • Accuracy can be affected by dust,
smoke, and background radiation
Sensor Integration
➢Importance of Sensor Integration
• Sensors are rarely used alone
• Integrated into complex systems with advanced processing and
actuation units
• Effective integration requires considering electrical and non-electrical
aspects
• Electrical aspects (impedance matching) covered in Chapter 3
• Non-electrical aspects: ensure sensor functions as intended
(integrator's task)
Two Important Integration Issues

• Optimal interface with the measurand (what's being measured)

• protecting the sensor from interference (internal and external)

Dead volume
• most sensors also require housing for various reasons
• This housing or shield not only minimises the exposure of the sensor to
the measurand but also creates a dead volume, which can trap the
measurand for some time
• The larger the dead volume, the longer the housing will hold a portion
of the measurand.
• As a result, there will be a delay before the sensor perceives a change in
the measurand.
• For example, for a temperature sensor, the dead volume can trap hot
air, which may not represent the actual state of the process being
monitored. The same can be said of magnetic or humidity sensors.
Consequently, this volume has to be minimised