What Is Sensitivity?

Sensitivity is a measure of how much the output of a system changes in response to variations in its parameters, inputs, or environmental conditions. It quantifies the relationship between small perturbations at the input or within the system and the resulting deviations observed at the output, and it appears as a foundational concept in control systems engineering, circuit design, and measurement science. The term originates from classical control theory and circuit analysis but now applies broadly across electrical engineering, signal processing, and physical measurement.

In its most general form, sensitivity is expressed as a ratio: the fractional change in an output quantity divided by the fractional change in the parameter that caused it. When that ratio is small, the system is robust to variation; when it is large, small changes in components or operating conditions produce large swings in performance. Engineers use sensitivity as a design target, aiming to minimize it for parameters that are uncertain or likely to drift.

Sensitivity in Control Systems

In control engineering, the sensitivity function describes how a closed-loop system responds to changes in the plant transfer function. It is formally defined as S(s) = 1/(1 + L(s)), where L(s) is the loop transfer function. This function also governs how external disturbances propagate to the system output, so a controller that reduces sensitivity to plant uncertainty simultaneously attenuates disturbances. A complementary relationship holds between the sensitivity function S and the complementary sensitivity function T: S + T = 1, which means that reducing sensitivity in one frequency band necessarily increases it in another. Engineering LibreTexts on sensitivity and robustness in control systems provides a detailed treatment of these trade-offs and their practical consequences. Robustness, defined as the ability of a control system to maintain stability and performance despite parameter variation, is therefore the inverse quality to sensitivity: a system is robust when its sensitivity is low across the operating range.

Sensitivity in Circuit Analysis

In circuit design, sensitivity describes the degree to which an output voltage, frequency, or impedance depends on the nominal values of individual components. Real resistors, capacitors, and transistors carry manufacturing tolerances and drift with temperature, aging, and electromagnetic interference. Sensitivity analysis identifies which components contribute most to output variation, guiding decisions about component grade, layout, and circuit topology. Two main methodologies apply: deterministic methods, which compute partial derivatives of the output with respect to each component value, and probabilistic methods such as Monte Carlo simulation, which sample component values from statistical distributions to estimate the distribution of circuit outputs. Cadence's treatment of circuit sensitivity analysis describes how modern numerical simulators apply both approaches to circuits that do not admit closed-form solutions.

Sensitivity in Measurement and Instrumentation

In sensors and instrumentation, sensitivity refers to the ratio of the output signal change to the input physical quantity change. A pressure sensor with a sensitivity of 10 millivolts per pascal delivers ten times as much output shift per unit of applied pressure as one rated at 1 millivolt per pascal, making it more suitable for small-signal detection. The IEEE 1451 family of standards for smart transducer interfaces includes provisions for encoding sensitivity metadata in a transducer electronic data sheet (TEDS), enabling automated calibration and unit conversion. High sensitivity in measurement instruments must be balanced against dynamic range and noise floor, since amplifying the signal of interest also amplifies noise.

Applications

Sensitivity has applications in a wide range of engineering and scientific fields, including:

• Feedback controller design for stability and disturbance rejection
• Analog and mixed-signal integrated circuit tolerance analysis
• Sensor calibration and instrument specification
• Structural health monitoring and fault detection systems
• Optimization of filter and amplifier topologies