Sensors, Instrumentation, and Electronic Systems

Introduction

Sensors connect physical processes to electronic systems. They convert temperature, displacement, pressure, light, magnetic field, chemical activity, or physiological activity into electrical quantities that can be conditioned, digitized, interpreted, communicated, and used for control.

A sensor alone rarely constitutes a measurement system. Practical instrumentation must also provide stable excitation, amplification, filtering, protection, calibration, data conversion, computation, communication, and—in closed-loop systems—actuation. Each stage contributes error, noise, delay, loading, and uncertainty. Good design therefore requires a system-level approach.

Learning Objectives

After studying this chapter, students should be able to:

• classify and interface resistive, capacitive, inductive, optical, magnetic, temperature, and biomedical sensors;
• design voltage-divider, bridge, excitation, amplification, filtering, and protection circuits;
• calculate instrumentation-amplifier gain and practical filter components;
• select ADC resolution, sampling rate, and reference arrangements;
• perform calibration and measurement-uncertainty analysis;
• distinguish measurement, monitoring, communication, and feedback-control architectures;
• develop sensor subsystems for embedded, IoT, biomedical, and industrial applications.

1. Measurement-System Architecture

A complete electronic measurement chain typically contains:

1. Measurand: the physical quantity to be determined.
2. Sensor or transducer: converts the measurand into an electrical parameter or signal.
3. Excitation: supplies current, voltage, optical energy, or an alternating carrier where required.
4. Signal conditioning: amplifies, filters, linearizes, isolates, and protects the signal.
5. Data conversion: converts an analog signal into digital data.
6. Processor or controller: applies calibration, estimation, decision, communication, or control algorithms.
7. Output subsystem: displays, stores, or communicates measurements.
8. Actuator: modifies the physical process when closed-loop control is required.

Figure: The processor/controller separately sends information to the display or network and commands to the actuator; the actuator does not receive commands from the display subsystem.

In an open-loop measurement system, the actuator and feedback path are absent. In a closed-loop system, the controller compares a measured value with a setpoint and computes the actuator command.

2. Sensor Characteristics and Measurement Quality

2.1 Static characteristics

For an approximately linear sensor,

where x is the measurand, y is sensor output, S is sensitivity, and b is zero offset. The units of S are output units divided by input units.

Accuracy describes closeness to a reference value, whereas precision describes the dispersion of repeated measurements. High precision does not guarantee low bias.

2.2 Dynamic characteristics

Many thermal, mechanical, and chemical sensors can be approximated as first-order systems:

Here τ is the time constant in seconds, K is static gain, and x and y are input and output. For a step input and zero initial output, the response reaches 63.2% of its final change after one time constant. This model applies only when one dominant energy-storage mechanism determines the dynamics.

3. Resistive Sensors and Bridge Circuits

Resistive sensors include strain gauges, resistance temperature detectors (RTDs), thermistors, potentiometric displacement sensors, piezoresistive pressure sensors, and photoresistors.

3.1 Voltage-divider interface

For a sensor resistance R_s connected below a fixed resistor R_f, the unloaded output is

where V_s is the excitation voltage. A following circuit with finite input resistance loads the divider; its input resistance must therefore be included in parallel with R_s. High-impedance buffers reduce this loading.

3.2 Wheatstone bridge

A bridge provides differential measurement and is particularly useful for small resistance changes. With left-arm resistors R₁,R₂, right-arm resistors R₃,R₄, and excitation V_s,

The bridge is balanced when R₁/R₂ = R₃/R₄. For a quarter bridge with R₁=R+ΔR and the other three resistors equal to R,

The approximation requires |ΔR| ≪ R. For a strain gauge,

where GF is the dimensionless gauge factor and ε is strain in m/m. Thus, for this quarter-bridge arrangement,

Half- and full-bridge arrangements offer greater sensitivity and improved temperature compensation. Ratiometric conversion, in which the bridge and ADC use the same reference supply, reduces sensitivity to excitation-voltage variation.

Worked Example 1: Bridge and Instrumentation-Amplifier Design

A quarter bridge uses a gauge with GF=2.1, R=350 Ω, excitation V_s=5.00 V, and strain ε=500 × 10^-6. Its approximate differential output is

Consider a three-operational-amplifier instrumentation amplifier whose matched differential stage has unity gain. If each first-stage feedback resistor is R and the gain-setting resistor between the two inverting nodes is R_G, its ideal differential gain is

This formula applies to the stated symmetrical topology; commercial instrumentation amplifiers may use a different manufacturer-specified constant.

Selecting R=10.0 kΩ and R_G=200 Ω gives

The first-stage output magnitude is therefore

To obtain approximately 1.00 V, a second non-inverting stage requires

Since G₂=1+R_F/R_I, choosing R_I=10.0 kΩ and R_F=65.4 kΩ gives G₂=7.54. The overall nominal gain is about 762, producing approximately 1.00 V.

In practice, resistor tolerance, bridge offset, input-offset voltage, common-mode range, output swing, bias-current effects, and drift must be checked. Excessive gain in one stage can cause saturation before offset correction; staged gain is therefore often preferable.

4. Capacitive and Inductive Sensors

4.1 Capacitive sensing

For parallel plates with effective area A, separation d, and dielectric permittivity ε,

This ideal expression neglects fringing fields. Capacitive sensors may detect displacement, pressure, humidity, touch, liquid level, or proximity. Common interfaces include:

• RC oscillators, where frequency changes with capacitance;
• AC bridges;
• switched-capacitor and charge-transfer circuits;
• capacitance-to-digital converters;
• differential capacitive structures that reject common-mode effects.

Parasitic cable capacitance can exceed the sensor change. Guard rings, driven shields, short traces, differential layouts, and remote conversion electronics reduce this problem. Moisture, contamination, and nearby conductors can alter the electric field and must be addressed during calibration.

4.2 Inductive sensing

Inductive sensors depend on changes in inductance, mutual coupling, eddy currents, or magnetic reluctance. Applications include proximity sensing, metal detection, current measurement, and position measurement.

A linear variable differential transformer (LVDT) uses an AC-excited primary and two secondary windings connected in series opposition. At the null position, the secondary voltages cancel. Core displacement changes their difference; synchronous demodulation converts the amplitude and phase into a signed displacement signal. LVDTs provide low friction and high endurance but require carrier generation, shielding, and demodulation.

5. Optical and Magnetic Sensors

5.1 Photodiodes and transimpedance amplifiers

A reverse-biased photodiode produces photocurrent approximately proportional to incident optical power over its linear range:

where I_p is photocurrent in amperes, P is incident optical power in watts, and ℜ is responsivity in A/W at the specified wavelength.

A transimpedance amplifier converts current to voltage. Ideally,

where R_F is feedback resistance and V_ref establishes the output baseline. A small feedback capacitor is often required for stability because photodiode and input capacitances introduce additional phase shift. Ambient-light subtraction, optical filtering, synchronous detection, and modulation improve performance in noisy environments.

5.2 Magnetic sensors

Hall-effect sensors produce a signal related to magnetic flux density. Integrated Hall devices commonly contain amplification, compensation, and digital threshold circuits. Applications include brushless-motor commutation, wheel speed, current measurement, position sensing, and contactless switching.

Magnetoresistive sensors change resistance in response to magnetic field and can offer greater sensitivity than basic Hall elements. Fluxgate sensors are used for precise low-field measurements. Magnetic designs must consider nearby currents, ferromagnetic materials, temperature drift, magnetic hysteresis, and alignment.

6. Temperature Sensors

For a platinum RTD over a restricted range near 0 °C, a linear approximation is

where R₀ is resistance at reference temperature T₀, α is the temperature coefficient in °C⁻¹, and T is temperature in °C. Precision work over a wide range requires the appropriate standardized nonlinear RTD equation.

An excitation current I causes self-heating power

If the dissipation constant is D in W/°C, the approximate self-heating error is ΔT=P/D. This is a steady-state approximation and depends on mounting and fluid flow.

Thermocouples measure a temperature difference, not an absolute junction temperature. Cold-junction compensation estimates the connector temperature and applies the thermocouple reference tables or polynomials.

7. Biomedical Sensors and Patient Safety

Biomedical instrumentation measures signals such as electrocardiograms, electromyograms, blood oxygen saturation, blood pressure, temperature, and respiration. These signals may be only microvolts or millivolts and are susceptible to electrode polarization, motion artifact, mains interference, and common-mode voltage.

An ECG front end commonly requires:

• high input impedance;
• high common-mode rejection ratio;
• low input noise and bias current;
• input-current limiting and defibrillation protection where applicable;
• galvanic isolation or appropriately certified patient protection;
• lead-off detection and controlled common-mode feedback;
• filtering that preserves diagnostically important waveform components.

Patient-connected equipment must comply with applicable medical-device standards, such as the relevant parts of IEC 60601, and with jurisdiction-specific regulation. Ordinary laboratory isolation practices are not automatically adequate for clinical use. Privacy, informed consent, cybersecurity, data integrity, and algorithmic bias are also engineering responsibilities.

8. Signal Conditioning Design

8.1 Common-mode rejection and grounding

Instrumentation amplifiers amplify differential signals while rejecting common-mode voltage. Their common-mode rejection ratio is

where A_d and A_cm are differential and common-mode gains. Resistor matching strongly influences CMRR. Cable shields, twisted pairs, differential routing, star grounding where appropriate, and separation of analog and switching-current paths reduce interference.

8.2 Practical low-pass filter sizing

An analog anti-alias filter attenuates frequencies that could fold into the sampled band. For a second-order low-pass Butterworth response,

where K is DC gain and f_c is the −3 dB cutoff frequency.

A Sallen–Key stage with equal resistors R, equal capacitors C, and non-inverting amplifier gain K has

under the equal-component assumptions. A Butterworth response requires Q=1/√2, giving K=1.586.

Worked Example 2: A 20 Hz Anti-Alias Filter

Choose C=100 nF and f_c=20 Hz. Then

The required amplifier gain can be obtained using K=1+R_F/R_I. Choosing R_I=10.0 kΩ and R_F=5.90 kΩ gives K=1.590, close to the desired 1.586.

If sampling occurs at 200 samples/s, the Nyquist frequency is 100 Hz. At 100 Hz, f/f_c=5, so

The attenuation relative to passband gain is approximately

Whether this is sufficient depends on the out-of-band signal and required alias rejection. Higher-order filtering or oversampling may be needed. Component tolerance and the filter's DC gain must also be included in the system error budget.

8.3 Input protection calculation

Protection must limit fault current without degrading normal measurements. Common elements include series resistors, current-limiting resistors, rail clamps, transient-voltage suppressors, RC filters, fuses, and isolation barriers.

Suppose a 3.3 V ADC input may accidentally receive 24 V. An external Schottky clamp limits the node to approximately 3.6 V, and the specified safe clamp current is 2 mA. The required series resistance is at least

Selecting R_s=12 kΩ limits the current to

The resistor's fault power is

A 0.125 W resistor provides reasonable steady-state power margin, subject to surge rating and fault duration. However, the 12 kΩ source impedance may violate the ADC's acquisition-settling requirement. A buffer can isolate the protection resistance from the ADC's sampling capacitor. Clamp leakage, normal input range, transient energy, and power-supply back-feeding must also be checked.

9. Sampling and Data Conversion

For an ideal N-bit ADC with unipolar full-scale range V_FS, one least significant bit is

For a 12-bit, 3.3 V ADC, LSB=3.3/4096=0.806 mV. Ideal quantization error is bounded by approximately ±0.5 LSB, excluding offset, gain, nonlinearity, reference, and noise errors.

The sampling theorem requires f_s>2f_max for a signal strictly band-limited below f_max. Practical systems require a transition band, so the sampling rate is normally chosen appreciably higher than twice the desired bandwidth and combined with an anti-alias filter.

Delta-sigma ADCs provide high resolution at modest bandwidth and are common in precision instrumentation. Successive-approximation ADCs provide a useful balance of speed, power, and resolution. ADC selection should consider effective number of bits, input topology, reference noise, latency, channel multiplexing, common-mode range, and interface timing—not nominal resolution alone.

10. Calibration and Uncertainty

Calibration establishes the relationship between instrument indication and reference values. It does not automatically adjust an instrument, although calibration data may be used for correction.

For two reference points (x₁,y₁) and (x₂,y₂), a linear model has

Multiple-point least-squares calibration is preferred when nonlinearity, noise, or residual analysis is important. Calibration should cover the operating range and relevant environmental conditions.

10.1 Uncertainty propagation

For a result q=f(x₁,x₂,…,x_n) with independent input quantities, the combined standard uncertainty is approximated by

This first-order propagation law assumes sufficiently small uncertainties, a differentiable model, and negligible correlations. Covariance terms must be included when inputs are correlated. Type A uncertainty is evaluated statistically from repeated observations; Type B uncertainty uses certificates, specifications, resolution, prior data, or engineering knowledge.

Worked Example 3: Calibrated Pressure and Uncertainty

A transmitter gives 0.498 V at 0 kPa and 4.502 V at 100 kPa. Its calibrated sensitivity and intercept are

At V=2.501 V,

Assume independent standard uncertainties u(V)=0.002 V, u(b)=0.003 V, and u(S)=0.00004 V/kPa. For p=(V-b)/S,

The three contributions are approximately 0.04995 kPa, 0.07493 kPa, and 0.04998 kPa. Therefore,

Using coverage factor k=2, the expanded uncertainty is approximately U=0.208 kPa. The result may be reported as 50.03 kPa ± 0.21 kPa for the stated model and an approximate 95% coverage interpretation.

11. Measurement, Control, IoT, and Communication

A monitoring system estimates and reports process conditions. A controller additionally modifies the process. For proportional-integral-derivative control,

where e(t) is setpoint minus measured output and u(t) is actuator command. Derivative action is sensitive to measurement noise, while integral action can accumulate during actuator saturation; filtered derivative and anti-windup are therefore common.

IoT sensor nodes add identification, timestamps, local storage, networking, remote configuration, and power management. Communication options include I²C and SPI within a device; RS-485, CAN, or industrial Ethernet within machinery; and Wi-Fi, Bluetooth Low Energy, cellular, or low-power wide-area networks for remote connectivity.

Engineering requirements include:

• end-to-end timing and acceptable control latency;
• packet loss and fail-safe behavior;
• authentication, encryption, and secure firmware updates;
• clock synchronization and traceable timestamps;
• power budgeting and battery-life estimation;
• local operation when cloud communication fails.

Cloud analytics and AI can detect anomalies, estimate sensor drift, fuse multiple sensors, and support predictive maintenance. An AI estimate should not be treated as a calibrated measurement unless its uncertainty, validation domain, failure modes, and traceability are established.

12. Illustrative Case: Instrumented Fermentation Vessel

Background

An illustrative food-processing facility requires automatic temperature control, level monitoring, optical turbidity measurement, batch records, and remote alarms for a fermentation vessel.

Problem

The temperature must remain within ±0.3 °C of setpoint. Long sensor cables run near motor drives, the vessel is washed regularly, and internet connectivity is intermittent. A heater is the principal actuator. Incorrect heating could spoil a batch or create a safety hazard.

Analysis

A Pt1000 RTD is selected for stability. Four-wire connection minimizes lead-resistance error. At 0.5 mA and approximately 1.1 kΩ, self-heating power is

If the installed dissipation constant is 5 mW/°C, the estimated self-heating rise is only 0.055 °C. The assumption must be validated in the actual thermowell and fluid-flow condition.

A differential capacitive probe estimates liquid level, while a modulated LED and photodiode with synchronous detection estimate turbidity. Shielding and differential transmission reduce motor-drive interference. Each channel is filtered before conversion by a precision ADC.

Approach and Solution

The local controller performs calibrated conversion, plausibility checks, alarm generation, and PID temperature control. Crucially, heater commands originate in the local controller—not in the display or cloud subsystem. A hardwired independent over-temperature switch removes heater power if the software control fails.

Measurements and calibration identifiers are stored locally with timestamps. Authenticated records are uploaded when the network becomes available. Remote users may request a setpoint change, but the local controller checks authorization, allowable limits, sensor validity, and safety interlocks before applying it.

Outcome

The architecture continues safe control during network failure, provides traceable batch data, and distinguishes process-control functions from informational functions. Redundant temperature protection prevents a single sensor, processor, or network fault from directly causing uncontrolled heating.

Learning Points

• Sensor choice must include installation and self-heating effects.
• Local control should not depend on cloud availability.
• Signal integrity, calibration, cybersecurity, and functional safety are coupled design concerns.
• Independent protection is required where the consequence of control failure is unacceptable.

13. Challenges, Governance, and Best Practices

Sensor systems face offset, drift, aging, cross-sensitivity, electromagnetic interference, mechanical stress, condensation, contamination, and component obsolescence. Calibration performed in a laboratory may not represent installed conditions. Communication and software add cybersecurity, version-control, and data-governance risks.

Professional best practices include:

• define the measurand and required uncertainty before selecting hardware;
• prepare an error and uncertainty budget early;
• verify common-mode, output-swing, bandwidth, and settling limits;
• use anti-alias filtering before sampling;
• separate control commands from display and reporting pathways;
• design protection from credible fault voltage, current, energy, and duration;
• retain calibration traceability and configuration history;
• test open-circuit, short-circuit, stale-data, and communication-loss behavior;
• minimize collection of personal data and protect retained records;
• validate AI models against representative operating and fault conditions.

Chapter Summary

Sensors convert physical or physiological variables into electrical signals, but accurate measurement requires excitation, conditioning, conversion, calibration, and uncertainty analysis. Resistive sensors commonly use dividers and bridges; capacitive and inductive sensors require careful management of parasitics and excitation; optical sensors often require transimpedance amplification; and magnetic, temperature, and biomedical sensors have distinct conditioning and safety requirements.

Instrumentation-amplifier gain, filter response, ADC resolution, source impedance, and input protection must be calculated rather than selected descriptively. Calibration determines the measurement model, while uncertainty analysis identifies the influence of reference, noise, resolution, drift, and model parameters. In control systems, actuator commands branch from the processor or controller, whereas display, storage, and network functions form a separate information path. Reliable IoT instrumentation combines local autonomy, cybersecurity, traceability, fail-safe behavior, and appropriate governance.

Key Terms

Actuator: A device that converts a command into physical action.

Anti-alias filter: An analog filter that attenuates frequencies capable of folding into the sampled band.

Calibration: Establishment of the relationship between indications and reference values.

Common-mode rejection ratio: A measure of an amplifier's ability to reject signals common to both inputs.

Drift: A gradual output change with time when the input is unchanged.

Excitation: Electrical or optical energy supplied to operate a sensor.

Gauge factor: Fractional resistance change per unit strain.

Hysteresis: Dependence of sensor output on the direction of input change.

Instrumentation amplifier: A precision differential amplifier with high input impedance and high common-mode rejection.

LVDT: An AC-excited inductive transducer used for displacement measurement.

Measurand: The quantity intended to be measured.

Ratiometric measurement: Measurement in which sensor excitation and converter reference vary proportionally.

Resolution: The smallest input change that can be distinguished.

Sensitivity: Change in output per unit change in input.

Signal conditioning: Amplification, filtering, isolation, protection, or linearization applied to a sensor signal.

Transducer: A device that converts energy or information from one form to another.

Transimpedance amplifier: A circuit that converts input current to output voltage.

Uncertainty: A parameter characterizing the dispersion of values attributable to a measurand.

Wheatstone bridge: A four-resistor differential network for detecting resistance changes.

Review Questions

Short Answer Questions

1. Distinguish sensitivity, resolution, accuracy, and precision.
2. Why is ratiometric conversion useful with a resistive bridge?
3. What is the purpose of synchronous demodulation in an LVDT interface?
4. Why can a large ADC input-protection resistor create a settling error?
5. Distinguish calibration from adjustment.

Long Answer Questions

1. Explain the complete measurement chain from measurand to digital result, identifying major sources of error.
2. Derive the approximate output of a quarter-bridge strain-gauge circuit and state the approximation condition.
3. Compare RTDs, thermistors, thermocouples, and integrated temperature sensors.
4. Explain the design requirements of a biomedical front end, including patient safety and data governance.
5. Discuss the relationship among anti-alias filtering, ADC sampling rate, bandwidth, and effective resolution.

Analytical / Application Questions

1. Design a bridge and two-stage amplifier for a pressure sensor whose maximum differential output is 2 mV and whose required ADC input is 2.5 V. Check offset and saturation constraints.
2. Design a second-order Butterworth Sallen–Key filter with a 50 Hz cutoff using 47 nF capacitors. Calculate resistor values, amplifier gain, and attenuation at 500 Hz.
3. Develop an architecture for a remotely monitored water tank with level sensing, pump control, local safety interlocks, calibration records, and secure communication. Clearly identify which subsystem issues actuator commands.

Multiple Choice Questions with Answers

Question 1

In a correctly structured closed-loop electronic system, which subsystem directly generates the actuator command?

A. Display subsystem

B. Processor or controller

C. Data-storage subsystem

D. Network dashboard

Correct Answer: B. Processor or controller

Question 2

A quarter bridge has V_s=5 V, gauge factor 2.0, and strain 400 × 10^-6. What is the approximate output magnitude?

A. 0.25 mV

B. 0.50 mV

C. 1.00 mV

D. 4.00 mV

Correct Answer: C. 1.00 mV

Question 3

For the symmetrical three-op-amp instrumentation amplifier described in the chapter, R=10 kΩ and R_G=200 Ω. What is the ideal gain?

A. 51

B. 100

C. 101

D. 201

Correct Answer: C. 101

Question 4

A first-order sensor subjected to a step input reaches approximately what percentage of its final change after one time constant?

A. 36.8%

B. 50.0%

C. 63.2%

D. 86.5%

Correct Answer: C. 63.2%

Question 5

Which interface directly converts photodiode current into voltage?

A. Schmitt trigger

B. Transimpedance amplifier

C. Wheatstone bridge

D. Differentiator

Correct Answer: B. Transimpedance amplifier

Question 6

A first-order RC low-pass filter uses R=7.96 kΩ and C=1.00 μF. Its cutoff frequency is closest to:

A. 2 Hz

B. 20 Hz

C. 125 Hz

D. 796 Hz

Correct Answer: B. 20 Hz

Question 7

A 24 V fault is clamped at 3.6 V. What minimum series resistance limits clamp current to 2 mA?

A. 1.8 kΩ

B. 7.2 kΩ

C. 10.2 kΩ

D. 13.8 kΩ

Correct Answer: C. 10.2 kΩ

Question 8

What is one LSB for an ideal 12-bit ADC with a 3.3 V unipolar range?

A. 0.403 mV

B. 0.806 mV

C. 1.61 mV

D. 3.30 mV

Correct Answer: B. 0.806 mV

Question 9

Which statement about a thermocouple is correct?

A. It directly measures absolute junction temperature without compensation.

B. It requires a bridge because its output is resistance.

C. It produces a Seebeck voltage related to a temperature difference.

D. It cannot operate above the range of silicon sensors.

Correct Answer: C. It produces a Seebeck voltage related to a temperature difference.

Question 10

For independent input quantities and a differentiable measurement model, combined standard uncertainty is commonly estimated using:

A. The arithmetic sum of all measured values

B. First-order sensitivity coefficients and root-sum-square contributions

C. Only the instrument resolution

D. The largest uncertainty contribution alone

Correct Answer: B. First-order sensitivity coefficients and root-sum-square contributions