Electrical Measurements, Sensors and Instrumentation

Introduction

Mechanical engineers routinely measure voltage, current, temperature, pressure, displacement, speed, force, torque, strain, vibration and flow. Design verification, machine protection, predictive maintenance and automated control all depend on measurements that are not merely numerical, but also traceable, repeatable and appropriate to the variable being observed.

An instrumentation system converts a physical measurand into useful information. A typical chain is:

measurand → sensor/transducer → signal conditioning → transmission or data acquisition → processing → display, control or storage

The quality of the final decision depends on every stage. A highly accurate sensor cannot compensate for poor wiring, electrical noise, unsuitable sampling, calibration drift or an instrument that significantly loads the measured system.

Learning objectives

After studying this chapter, the reader should be able to:

1. Distinguish accuracy, precision, sensitivity, resolution, linearity, repeatability and uncertainty.
2. Explain the operation and limitations of analog and digital electrical instruments.
3. Select sensors for common mechanical variables.
4. Analyze bridge circuits, loading effects and signal-conditioning stages.
5. Apply calibration and uncertainty concepts to measurement results.
6. Design the principal elements of a basic data-acquisition system.
7. Evaluate measurement systems in practical mechanical-engineering applications.

1. Measurement concepts and terminology

A measurand is the quantity intended to be measured. The true value is an ideal concept; in practice, the best estimate is obtained from a calibrated reference and a stated uncertainty. A measurement result should therefore include both a numerical value and its unit, and, where necessary, an uncertainty interval.

Accuracy and precision

Accuracy describes closeness to the accepted reference value. Precision describes the closeness of repeated observations to one another. A system may be precise but inaccurate when it has a stable calibration bias. Random variation reduces precision; systematic error reduces accuracy.

The error of a measured value xₘ relative to a reference xᵣ is:

The percentage error, when xᵣ is nonzero, is:

The signed result preserves whether the instrument reads high or low. Absolute error is the magnitude of the error.

Sensitivity, resolution and range

Sensitivity is the change in output per unit change in input:

Here x is the measurand and y is the instrument output. For a linear sensor, S is constant and the output may be written as:

where y₀ is the zero output. Resolution is the smallest input change that can be distinguished. A digital display with a 0.1 °C least significant digit does not necessarily have 0.1 °C accuracy.

Range is the interval over which an instrument is specified to operate. Span is the upper range value minus the lower range value. Linearity describes how closely the calibration curve follows a specified straight line. Hysteresis is the difference in output at the same input when approached from increasing and decreasing directions.

Repeatability, reproducibility and drift

Repeatability concerns repeated measurements under the same conditions. Reproducibility concerns measurements under changed conditions, such as different operators, instruments or laboratories. Drift is a gradual change in output for a constant input, often caused by temperature, ageing or mechanical stress.

Static and dynamic characteristics

Static characteristics apply when the measurand is constant or changing slowly. Dynamic characteristics describe response to time-varying inputs. A first-order instrument can often be approximated by:

where τ is the time constant, K is static sensitivity, x is input and y is output. For a step input, the output reaches approximately 63.2% of its final change after one time constant and about 95% after three time constants. This model is useful for thermometers and many fluid-pressure systems, but not for every sensor.

2. Analog and digital electrical instruments

Analog instruments

Analog instruments indicate a continuously varying quantity. A permanent-magnet moving-coil (PMMC) meter produces deflecting torque approximately proportional to current. It has good sensitivity and a uniform scale for direct current, but it requires rectification or another mechanism for alternating current.

Moving-iron instruments respond to both direct and alternating current. Their scale is nonuniform and they generally consume more power than PMMC instruments. Electrodynamometer instruments can measure power and alternating-current quantities because their fixed and moving coils interact.

An analog voltmeter is fundamentally a sensitive current meter connected in series with a large resistance. If the movement requires full-scale current Iₘ and the desired voltage range is V, the total resistance is approximately:

For a voltmeter with internal resistance Rᵥ connected across a circuit resistance R, the measured voltage is affected by the parallel combination:

This is the loading effect. It is small when Rᵥ is much larger than R.

An ammeter is connected in series and should have very low resistance. A shunt resistor extends its current range. Incorrect connection of an ammeter across a voltage source can cause excessive current and instrument damage.

Digital instruments

A digital multimeter generally contains an input protection network, attenuator or shunt, analog-to-digital converter (ADC), reference, processor and display. Advantages include high input resistance, direct numerical readout, automatic range selection and data logging. Limitations include quantization, aliasing in sampled signals, update-rate limitations and possible confusion caused by excessive displayed digits.

For an ideal N-bit ADC with input range 0 to Vᵣ, the quantization step is approximately:

The ideal quantization uncertainty is commonly modeled as ±q/2 when the signal is not correlated with the conversion code. Real accuracy also includes reference error, offset, gain error, integral nonlinearity and noise.

A true-RMS meter is necessary when an AC waveform is nonsinusoidal and the heating or power effect is important. For a periodic voltage v(t), its RMS value over period T is:

For a sinusoid with peak value Vₚ, V_RMS = Vₚ ÷ √2, but that special result must not be applied to arbitrary waveforms.

3. Measurement error, uncertainty and calibration

Sources of error

Errors may arise from:

• Instrument error: imperfect calibration, hysteresis, drift and finite resolution.
• Environmental error: temperature, humidity, vibration, electromagnetic interference and supply variation.
• Observational error: parallax, poor timing or incorrect interpretation.
• Loading error: the instrument changes the system being measured.
• Random error: unpredictable variation around a mean.
• Systematic error: repeatable bias that can often be corrected by calibration.

A result should not report more significant figures than the instrument and uncertainty justify.

Combining independent uncertainties

If a result y depends on independent input quantities x₁, x₂, …, xₙ, and their standard uncertainties are u(xᵢ), the first-order propagation law is:

This assumes small uncertainties, a sufficiently smooth model and negligible covariance. If inputs are correlated, covariance terms must be added. For y = a + b, independent uncertainties combine as u(y) = √(u²(a) + u²(b)). For a product or quotient, relative uncertainty is often the useful approximation:

Expanded uncertainty is often reported as U = ku(y), where k is a coverage factor, commonly near 2 for approximately 95% coverage when a normal model is appropriate. It is not automatically a guarantee of exactly 95% probability.

Calibration

Calibration compares an instrument with a reference of known traceability over selected points. A practical calibration procedure is:

1. Define the measurand, range, environmental conditions and acceptance limits.
2. Select a reference whose uncertainty is sufficiently smaller than the unit under test.
3. Allow thermal and electrical stabilization.
4. Apply increasing and decreasing input points where hysteresis is possible.
5. Record reference and indicated values, including units and conditions.
6. Fit or determine correction values, evaluate uncertainty and document traceability.
7. Label the instrument with status, date and due date, then control the calibration record.

A two-point linear correction can be obtained from zero and span checks, but it should not conceal significant nonlinearity. Calibration intervals should reflect stability history, use severity and consequences of failure rather than an arbitrary calendar alone.

4. Sensors and transducers for mechanical engineering

A sensor responds to a physical quantity. A transducer converts that quantity into another form, usually an electrical signal. A primary transducer directly responds to the measurand; a secondary transducer converts the primary response into an electrical output.

Resistive sensors

A metallic resistance temperature detector (RTD) is commonly modeled over a limited range by:

R₀ is resistance at reference temperature T₀, α is the temperature coefficient in °C⁻¹, and R_T is resistance at T. The approximation is not universally valid over a wide range; higher-order coefficients may be required.

Strain gauges exploit the change in resistance caused by deformation. Gauge factor is:

where ε is dimensionless strain and R is nominal resistance. Bonding quality, transverse sensitivity and temperature compensation strongly affect results.

Thermocouples

A thermocouple produces a voltage related to the temperature difference between its measuring junction and reference junction. The voltage is small and requires amplification, filtering and cold-junction compensation. It does not directly measure absolute junction temperature without reference information.

Inductive and capacitive sensors

An LVDT provides a differential output that changes approximately linearly with core displacement over its specified range. It has no sliding electrical contact and can provide high resolution, but requires AC excitation and phase-sensitive signal conditioning.

Capacitive sensors detect changes in capacitance caused by displacement, dielectric variation or geometry. They can be very sensitive, but stray capacitance, contamination and humidity may be significant.

Piezoelectric and semiconductor sensors

Piezoelectric sensors generate charge under dynamic force and are excellent for vibration and impact measurement. Because charge leaks through practical insulation, they are unsuitable for accurate static force measurement over long periods. MEMS accelerometers may use capacitive proof masses and integrated electronics; they are compact but can have cross-axis sensitivity, bias drift and limited shock tolerance.

5. Bridges and signal conditioning

Bridge circuits convert small changes in resistance, inductance or capacitance into a differential voltage. The Wheatstone bridge has four arms. If supply voltage is Vₛ and arm resistances are R₁, R₂, R₃ and R₄, the output between the divider midpoints is:

The sign depends on which midpoint is designated positive. The bridge is balanced when:

For a quarter-bridge strain circuit with one active gauge, nominal resistance R and small change ∆R, the approximate output magnitude is:

This assumes the other arms remain nominal, the change is small, and the bridge output is measured with high input resistance. Half- and full-bridge arrangements increase sensitivity and can compensate temperature effects when active gauges are correctly positioned.

Signal conditioning commonly includes:

• instrumentation amplifiers for high common-mode rejection and high input resistance;
• differential amplifiers for bridge outputs;
• excitation sources, preferably stable and low-noise;
• low-pass, high-pass or band-pass filters;
• isolation for safety and ground-loop control;
• linearization, cold-junction compensation and sensor fault detection;
• voltage-to-current conversion, such as industrial 4–20 mA transmission.

For an ideal first-order RC low-pass filter, the cutoff frequency is:

R is resistance in ohms, C is capacitance in farads and f_c is in hertz. The filter attenuates frequencies above the cutoff; it does not provide an infinitely sharp boundary.

6. Data acquisition and digital measurement

A data-acquisition system must match sensor bandwidth, amplitude, isolation and timing requirements. Its principal elements are sensor, excitation, analog conditioning, anti-alias filter, sample-and-hold, ADC, processor and storage or communication interface.

Sampling and aliasing

If a signal contains no frequency components above fₘₐₓ, ideal uniform sampling requires:

This is the Nyquist condition. In practice, the sampling frequency is selected higher than the theoretical minimum, and an analog anti-alias filter attenuates components near and above half the sampling frequency. Sampling faster cannot recover information already lost through inadequate sensor bandwidth or aliasing.

Grounding, shielding and interference

Low-level signals are vulnerable to electromagnetic interference, ground loops and common-mode voltage. Good practice includes twisted-pair differential wiring, cable shields connected according to the grounding architecture, physical separation from power conductors, single-point or carefully engineered multipoint grounding, proper sensor excitation and separation of analog and high-current return paths. Shielding alone cannot correct poor reference design or saturation of the input stage.

Data integrity and software

A measurement database should preserve raw data, calibration coefficients, timestamps, units, sensor identity and quality flags. Software should distinguish missing data from zero, detect out-of-range values and record configuration changes. AI-based anomaly detection can assist maintenance by identifying patterns in vibration or temperature, but it cannot replace sensor calibration, physical reasoning or human authorization for safety-critical decisions. Training data bias, false alarms and cybersecurity must be considered.

7. Worked examples

Example 1: Voltmeter loading

A circuit has Thevenin voltage V_th = 10 V and resistance R_th = 100 kΩ. A voltmeter has input resistance Rᵥ = 1 MΩ. The indicated voltage is the voltage across the parallel combination:

Thus:

The loading error relative to 10 V is approximately −9.09%. A meter with 10 MΩ input resistance would reduce this error substantially.

Example 2: Wheatstone bridge output

A quarter-bridge uses four nominal 120 Ω arms and a supply of 5 V. One strain gauge has ∆R = 0.12 Ω. Using the small-change approximation:

The signal is small compared with the supply, so an instrumentation amplifier, stable excitation and temperature compensation are required.

Example 3: Uncertainty in resistance-derived current

A current is calculated from I = V/R, where V = 12.0 V with standard uncertainty 0.1 V and R = 100 Ω with standard uncertainty 1 Ω. The calculated current is 0.120 A. For independent inputs:

Therefore the relative standard uncertainty is approximately 1.18%, giving u(I) ≈ 0.0014 A. A suitable statement is I = 0.120 ± 0.001 A, using an appropriate rounding and coverage convention. The stated uncertainty does not include unrecognized systematic errors.

8. Illustrative case: monitoring a motor-driven pump

Background

A manufacturing plant experiences intermittent pump failures. The pump is driven by an induction motor, and failures appear to involve bearing wear and inadequate cooling.

Problem

Existing operators record motor current once per shift. This misses short-duration vibration events and provides no direct bearing-temperature trend. The engineering team must design a measurement system that supports early warning without creating unsafe electrical connections.

Analysis

The relevant variables are vibration acceleration, bearing temperature, motor current, shaft speed and discharge pressure. Vibration requires a sensor with adequate frequency response and rigid mounting. Temperature can be measured using an RTD or industrial transmitter. Current measurement should preferably use an isolated current transformer or Hall-effect transducer. A pressure transmitter should be selected for the fluid, range and process connection.

Approach and solution

An accelerometer is mounted close to the bearing, an RTD is installed in the bearing housing, and isolated current and pressure transmitters feed a data-acquisition unit. The vibration channel uses a charge or IEPE-compatible front end, a defined band-pass range and a sampling rate selected after identifying the highest frequency of interest. The temperature channel is calibrated against a reference thermometer. All channels receive synchronized timestamps and retain raw data as well as alarm features.

The system initially establishes a healthy baseline. Alarm limits are not chosen solely from generic tables; they are based on baseline variation, manufacturer limits, process state and confirmatory inspections. A rising temperature trend combined with a change in vibration spectrum is treated as stronger evidence than either signal alone.

Outcome

In this illustrative scenario, the system identifies a progressive increase in bearing vibration before temperature reaches the shutdown limit. Inspection confirms lubrication degradation. Planned maintenance replaces the lubricant during a scheduled stop, avoiding secondary shaft damage.

Learning points

The case demonstrates that sensor selection, mounting, calibration, sampling and interpretation are inseparable. A data-rich system is not automatically reliable: traceability, alarm governance, cybersecurity, maintenance of the sensors and competent interpretation are essential.

9. Limitations, safety and professional practice

No measurement is perfectly independent of the system measured. Sensor mass can alter vibration; thermocouple installation can conduct heat away; pressure taps can disturb flow; and wiring resistance can change a resistance measurement. Engineers must state assumptions and assess whether the act of measurement changes the measurand.

Electrical measurements near machines require isolation, fusing, protective enclosures, appropriate categories of test equipment and lockout/tagout procedures. Instruments must be rated for the expected voltage, transient energy and environment. Calibration records should be controlled under a quality system, with traceability to recognized standards where required by the application.

Measurement data may identify personnel, production rates or equipment vulnerabilities. Access control, secure communication, audit trails and retention policies are therefore part of responsible instrumentation engineering. Safety-related measurements require fail-safe design, independent protection where appropriate and validation against applicable standards and regulations.

Chapter Summary

Electrical instrumentation transforms physical variables into usable information through sensing, conditioning, conversion, transmission and interpretation. Accuracy and precision are different properties; uncertainty must accompany significant measurement results. Analog meters provide intuitive continuous indication but may load a circuit, while digital instruments offer high input resistance and numerical processing but introduce quantization and sampling considerations. Sensors for temperature, strain, force, pressure, displacement, speed and vibration have distinct physical principles and limitations. Wheatstone bridges are fundamental for small resistance changes, and instrumentation amplifiers, filtering, isolation and stable excitation are essential for reliable signals. Calibration establishes correction and traceability, while uncertainty analysis quantifies known sources of variation. Data acquisition must address bandwidth, aliasing, grounding, data integrity and cybersecurity. In practice, a defensible measurement system is selected for the measurand, installed correctly, calibrated, protected and interpreted in its physical context.

Key Terms

Measurand: Quantity intended to be measured.

Accuracy: Closeness of a result to an accepted reference value.

Precision: Closeness of repeated results to one another.

Sensitivity: Change in output per unit change in input.

Resolution: Smallest distinguishable input change.

Linearity: Agreement of a calibration curve with a specified straight line.

Hysteresis: Difference in output for the same input approached from opposite directions.

Drift: Gradual output change for a constant input.

Loading effect: Change in the measured system caused by connecting the instrument.

Calibration: Comparison with a reference to determine indication error or correction.

Traceability: Documented connection of a result to recognized measurement standards.

Uncertainty: Quantified doubt associated with a measurement result.

Sensor: Device that responds to a physical quantity.

Transducer: Device that converts a measurand into another usable form.

Wheatstone bridge: Four-arm circuit used to compare or measure impedance changes.

Instrumentation amplifier: High-input-resistance differential amplifier used for low-level signals.

Aliasing: Misrepresentation of frequency caused by inadequate sampling.

Anti-alias filter: Analog filter that limits frequency content before sampling.

Data acquisition: Collection and conversion of sensor signals for processing or storage.

Review Questions

Short Answer Questions

1. Distinguish accuracy from precision and give one example of each.
2. Why is a voltmeter designed with high input resistance?
3. State two limitations of thermocouple measurements.
4. What is the purpose of an instrumentation amplifier in a strain-measurement system?
5. Define aliasing and state one method of reducing it.

Long Answer Questions

1. Explain the complete measurement chain from measurand to engineering decision, including sources of error at each stage.
2. Compare analog and digital instruments with respect to operating principle, loading, resolution, accuracy and practical use.
3. Explain Wheatstone-bridge operation and derive the small-change quarter-bridge relation.
4. Describe a rigorous calibration procedure and discuss traceability, drift and uncertainty.
5. Design a data-acquisition system for monitoring temperature and vibration in a rotating machine.

Analytical / Application Questions

1. A 10 MΩ voltmeter measures a 1.00 MΩ Thevenin source with open-circuit voltage 5.00 V. Calculate the indicated voltage and percentage loading error.
2. A 12-bit ADC has a 0–10 V input range. Calculate its ideal quantization step and explain why this is not the total measurement uncertainty.
3. Select suitable sensors and signal-conditioning stages for measuring force, shaft speed and bearing temperature in a test rig. Justify the choices in terms of range, bandwidth, loading, calibration and safety.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes accuracy from precision?

A. Accuracy is repeatability, while precision is closeness to a reference.

B. Accuracy is closeness to a reference, while precision is repeatability.

C. Accuracy applies only to analog instruments, while precision applies only to digital instruments.

D. Accuracy and precision always have identical meanings.

Correct Answer: B. Accuracy is closeness to a reference, while precision is repeatability.

Question 2

A voltmeter measures a Thevenin source with V_th = 10 V, R_th = 100 kΩ, and input resistance 900 kΩ. What voltage does it indicate?

A. 0.99 V

B. 5.00 V

C. 9.00 V

D. 10.90 V

Correct Answer: C. 9.00 V

Question 3

For an ideal N-bit ADC with a 0–5 V input range, the approximate quantization step is

A. 5N V

B. 5 ÷ N V

C. 2ᴺ ÷ 5 V

D. 5 ÷ 2ᴺ V

Correct Answer: D. 5 ÷ 2ᴺ V

Question 4

A Wheatstone bridge is balanced when

A. R₁ + R₂ = R₃ + R₄

B. R₁R₄ = R₂R₃

C. R₁ ÷ R₂ = R₃R₄

D. all four resistances are zero

Correct Answer: B. R₁R₄ = R₂R₃

Question 5

Which sensor is particularly suitable for dynamic vibration and impact measurements but generally unsuitable for long-duration static force measurement?

A. Piezoelectric sensor

B. Potentiometric sensor

C. RTD

D. Thermocouple

Correct Answer: A. Piezoelectric sensor

Question 6

The principal purpose of an anti-alias filter is to

A. increase sensor mass

B. remove DC power from a circuit

C. limit frequency content before analog-to-digital conversion

D. convert current directly into mechanical displacement

Correct Answer: C. limit frequency content before analog-to-digital conversion

Question 7

For a first-order RC low-pass filter, the cutoff frequency is

A. RC ÷ 2π

B. 2πRC

C. 1 ÷ 2πRC

D. R ÷ 2πC

Correct Answer: C. 1 ÷ 2πRC

Question 8

Which instrument connection is normally correct?

A. An ammeter is connected in parallel with the supply.

B. A voltmeter is connected in series with the load.

C. An ammeter is connected in series with the measured branch.

D. Both meters must always be connected in series.

Correct Answer: C. An ammeter is connected in series with the measured branch.

Question 9

A quarter-bridge strain circuit has Vₛ = 4 V, R = 100 Ω and ∆R = 0.20 Ω. Using the small-change approximation, the output magnitude is

A. 0.002 V

B. 0.020 V

C. 0.200 V

D. 2.000 V

Correct Answer: A. 0.002 V

Question 10

Which practice most directly improves the traceability of a measurement result?

A. Displaying more decimal places

B. Using a documented calibration comparison with a reference standard

C. Increasing the cable length

D. Removing all uncertainty statements

Correct Answer: B. Using a documented calibration comparison with a reference standard