Temperature as an equilibrium relation. Classical thermodynamics begins with the zeroth law of thermodynamics: thermal equilibrium is transitive. A thermometer can therefore assign a common empirical temperature to systems in mutual thermal equilibrium. Different thermometer properties initially give different monotone empirical scales; a resistance reading by itself does not define the absolute scale.
The Second law of thermodynamics provides a substance-independent thermodynamic scale. A reversible engine operating between reservoirs has a heat ratio depending only on their temperatures. One chooses absolute temperature so that , with the positive heat magnitudes. Its reversible efficiency is . Equivalently, reversible heat defines entropy through . For an equilibrium simple compressible system,
Classical thermodynamics uses this relation without needing a microscopic account of entropy. A dilute-gas thermometer realizes an empirical scale approaching the thermodynamic scale after corrections for nonideal behavior.
Statistical meaning of the same derivative. In the microcanonical description, an equilibrium energy shell containing accessible states has . Its statistical temperature is consequently
To see why this is an equilibrium temperature, let two weakly coupled systems exchange a small energy while their total energy stays fixed. The combined multiplicity is . Maximizing its logarithm gives
which is equality of the two inverse temperatures. Stability requires the total entropy maximum, not only stationarity. The overwhelmingly dominant macrostate in a large system therefore realizes the thermodynamic equilibrium condition.
A large reservoir gives a small subsystem the canonical probabilities , with . In that ensemble, and agree with macroscopic thermodynamic relations in their usual equilibrium regime. The equipartition identity for a classical dilute monatomic gas, , is a useful operational connection between microscopic motion and temperature, but is not a universal definition: quantum modes need not obey equipartition, and a system away from equilibrium need not have one temperature at all.
For finite systems one must specify the entropy convention and energy resolution; surface and fluctuation corrections can distinguish definitions that agree in the ordinary thermodynamic limit. Bounded-energy population-inverted systems can also have a negative entropy derivative and hence a negative statistical temperature. That possibility lies outside an ordinary positive-temperature ideal-gas thermometer; it does not mean temperature is merely an arbitrary average kinetic energy.
Obtaining the energy–temperature conversion. The Boltzmann constant connects entropy measured in with the dimensionless logarithm of a state count. Historically it could be determined from the ideal gas law : measure the molar gas constant and independently determine the Avogadro constant , then obtain . The independent particle-number determination matters; using alone would not fix energy per particle.
An electrical method uses Johnson-Nyquist noise. A resistor of resistance at known temperature has, in the classical range , a one-sided open-circuit voltage-noise spectrum . Over a calibrated flat bandwidth ,
A practical measurement uses an equilibrium resistor, a calibrated low-noise amplifier and a known transfer function, subtracts or independently characterizes amplifier noise, and replaces a nominal bandwidth by the actual effective noise bandwidth. Varying temperature and fitting the mean-square voltage is more informative than treating one amplified voltage sample as the thermal variance. Other primary methods relate temperature to sound speed in a dilute gas or to calibrated particle diffusion and mobility.
There is a historically relevant unit-definition qualification: by the date of this exam, the revised SI had already taken effect on 20 May 2019. The kelvin is now defined by fixing
The same experiments now realize or compare thermodynamic temperatures using this fixed constant, whereas historically they measured the conversion constant on an independently defined temperature scale. The physical distinction between equilibrium temperature and microscopic state counting is unchanged.[ref]
Use an RC low-pass filter: put the resistor in series with the input and the capacitor between the output node and the return conductor. Measure the output voltage across the capacitor, with negligible output loading. The voltage divider then has transfer function
At the high frequency, the attenuation requirement is
At the low frequency, approximately unchanged amplitude means , so
The upper relation is an asymptotic design condition, because “approximately” specifies no numerical tolerance. For example, retaining at least of the low-frequency amplitude would sharpen the upper bound to approximately .
A practical illustrative choice is . Its voltage gain is approximately at and at : the latter amplitude is reduced by a factor of about . Its cutoff frequency is .
Figure 1.
An original RC low-pass circuit and its amplitude response for a 22 microfarad capacitor
. Details.
Measure the mass's displacement from its static equilibrium height. This removes the constant spring extension supporting its weight. If the spring's top moves with the box, its incremental extension is , so its incremental upward force is . The specified damping force is against the mass's velocity in the static frame. Newton's second law therefore gives
In absolute height coordinates an additional constant on the right expresses the spring's rest length, its fixed attachment height and gravity; translating to static equilibrium removes it. Replacing the given damping by would describe a different base-excited oscillator.
After the decaying transient, write the steady-state response as . Using a complex sinusoid gives
so
For positive frequency, the phase lag lies between zero and , passing through at . For small , the frequency response begins at , develops a narrow resonance near , and falls as
The precise amplitude-peak frequency is when that expression is positive.
Figure 1.
Absolute and relative motion of the stated seismograph model, and the phase lag of its inertial mass
. Details.
The pen records the relative displacement . Its complex transfer function is
At high frequency, is small while . The inertial mass then stays nearly stationary and the recorded relative motion reproduces the ground displacement, with reversed sign. At low frequency the mass follows the box and their relative motion is small; near resonance the response is strongly frequency-dependent. Here “high” means compared with the suspension's natural frequency, rather than an absolute seismic-frequency threshold.
In the phasor convention, the total series RL circuit impedance is . The resistor voltage follows the current, so its transfer function relative to the applied voltage is
At the stated condition ,
Thus the resistor-voltage amplitude is and it lags by . For a cosine input it is .
Write the dimensionless scale factor as and the density contrast as . Gravity amplifies overdensities while expansion and pressure oppose growth. In a sub-Hubble Newtonian single-fluid approximation, a comoving Fourier mode obeys
For pressureless matter in an Einstein-de Sitter background, the growing solution is and the decaying solution is . Jeans support suppresses growth when the pressure term dominates. The equation is not a substitute for relativistic perturbation theory on horizon-sized scales.
cold dark matter clusters without baryonic pressure support. Sub-horizon growth is much weaker during radiation domination, and baryons are initially tied to the photon fluid, producing acoustic oscillations rather than unrestricted collapse. Horizon entry, radiation-matter equality, acoustic evolution and diffusion therefore give a scale-dependent transfer function. The primordial curvature spectrum and the later matter-density spectrum are not the same function: evolution and the gravitational constraints relate them. Late dark-energy domination reduces the growth rate.
When approaches unity, linear theory fails. Collapse, virialization, mergers and tidal interactions form a hierarchy of haloes. Gas must cool and lose angular momentum to make dense stellar systems; feedback then changes its supply and distribution. Simulations combine collisionless gravitational dynamics with increasingly complex gas physics rather than extending the linear growing mode indefinitely. Galaxy clustering measures biased tracers of matter; gravitational lensing and CMB constraints provide complementary tests. Approximately adiabatic, nearly scale-invariant initial perturbations explain the coherent acoustic pattern and supply seeds for this later growth.[ref]
Ratiometric cancellation with matched signal and reference filters preserves a resistance ratio despite changing excitation when both paths apply the same dynamics to the shared electric current. Matching only the steady gain is insufficient during a transient.
Let constant sensor and reference resistances be , and let both paths be linear time-invariant systems with impulse responses . With consistent initial conditions, their filtered electric potential differences are
where is convolution. If , simultaneous sampling gives whenever the denominator is nonzero. Arbitrary unequal stored initial states need not cancel. Nor can a time-varying generally be pulled outside the convolution.
For unit steady gains, write after settling at . Define and . Then
This displays exactly which excitation component survives. For small fluctuations the fractional residual is , so both amplitude and timing differences between the transfer functions matter.
As an original transient example, step the excitation up by ten percent from a settled state. Two first-order low-pass filters with time constants and give, for ,
At the reading is high by approximately , although both filters have the same final gain and the error tends to zero. Setting both time constants to cancels this particular excitation step at every time. Sampling skew after matched ratiometric filters can still break simultaneous cancellation even with identical hardware.
The primary ADS1248 experiment used a resistance decade box, a excitation and a reference. Reference-only filtering produced signal-dependent noise; matching the signal and reference paths substantially reduced that dependence. The report provides the circuit and noise-versus-input plots, rather than merely recommending a common nominal cutoff.[ref]
A series resistor and capacitor feed a parallel pair with the same values. Measuring the parallel pair's voltage relative to the return gives this unloaded transfer function for phasors proportional to . The voltage divider applied to and proves the formula. The amplitude peaks at with ratio and zero relative phase. At low and high positive frequencies, the transfer function tends respectively to and . Reversing output polarity changes its sign, while loading changes the response itself.
Sinusoidal steady state 2026-11-05
The sinusoidal steady state of a stable linear time-invariant system driven by a sinusoid is its persistent response at the driving frequency, after the relevant transient response has decayed. A phasor calculation finds that response; it does not generally satisfy the circuit's switch-on initial conditions. The late-time assumption is explicit in Berggren's treatment of sinusoidal circuit analysis.[ref]
Using peak phasors, write the source as . If is the relevant transfer function, its particular solution is
The complete response adds a homogeneous solution determined by the initial conditions. Approaching the displayed expression requires the modes affecting the measured output to decay; an unchanged amplitude and a fixed phase shift then characterize its frequency response.
For an original turn-on example, drive a series RC circuit with for , measure the capacitor's electric potential difference, and take . With time constant , Kirchhoff's voltage law gives the first-order linear differential equation
Put . Substitution verifies the sinusoidal particular solution . Adding and imposing the initial condition fixes , giving
Its sinusoidal part has amplitude and lags by . The exponential settling bound for RC sinusoidal turn-on, relative to that amplitude, is
With chosen , , and , we have and . At turn-on the phasor approximation predicts 0.5 volt while the capacitor is initially at zero. After 5 milliseconds, their difference is about 3.37 millivolts, or 0.476 percent of the sinusoidal amplitude. A 1 percent settling time threshold for this specific switch-on is about .
Figure 1.
A phasor response omits the exponential switch-on transient
. Original RC example with time constant 1 millisecond and driving angular frequency 1000 radians per second. The exact capacitor response starts at zero, approaches its sinusoidal particular solution and crosses a one-percent relative transient threshold after about 4.259 time constants. Details.
A useful measurement compares a long, low-frequency square-wave response with sinusoidal responses at several frequencies. An oscilloscope can identify the exponential decay and then measure amplitude ratios and phase shifts. MIT's laboratory uses that sequence to connect the time constant to the sinusoidal response; probe loading belongs in the measured circuit model.[ref]
Undamped resonant forcing has no bounded sinusoidal steady state. For an ideal series LC circuit, let be electric charge and . At resonance, the equation
has the particular solution . Its growing envelope prevents a bounded attracting sinusoidal response. This ideal counterexample explains why decay of the relevant homogeneous solutions is part of the physical steady-state assumption.
The source-filter model of speech production represents a speech sound as an acoustic source shaped by a filter describing the vocal tract. During voiced speech, vibration of the vocal folds supplies a roughly periodic source; constrictions can also generate noise. The tract's resonances emphasize particular frequency regions called formants.[ref]
Paul Boersma credits Gunnar Fant's 1960 account as the theoretical basis of this construction.[ref]
In a simplified linear model, output spectrum is , where describes the source and the tract's transfer function. Keeping the source fixed while changing the filter can change vowel quality without changing the source's fundamental repetition rate. The Praat tutorial makes this test concrete by constructing a source signal and a formant filter separately, then combining them. The model's separation is an approximation; real sources and tracts can interact.[ref]
Transfer function Created 2026-09-24 Updated 2026-10-13
A transfer function packages the response of a linear time-invariant system into a rule for turning a plant input into a plant output. Instead of solving a differential equation separately for every signal, take a Laplace transform, multiply by , then transform back. This also makes it possible to compare slow and fast oscillations through the continuous-time frequency response.
Here consider a continuous-time, causal, finite-dimensional linear time-invariant system with one plant input and one plant output . With every initial state set to zero, its transfer function is defined by
The ratio is used where in a common Laplace convergence region; describes the system, independently of that particular plant input. Zero initial state matters: stored motion or another nonzero initial condition can contribute to the plant output even when the plant input vanishes. [ref]
The first-order low-pass filter below shows the same transfer function predicting both a step and a sustained sinusoid. Its pole gives a decay rate, its value at zero gives the eventual gain, and its values along the imaginary axis give the continuous-time frequency response. These descriptions make a useful bridge between a physical relaxation model and control theory calculations.
Unmodeled dynamics 2026-11-05
Unmodeled dynamics are time-dependent behaviors omitted from the nominal model of a controlled plant or its interconnection. Examples include an omitted actuator lag, an additional flexible mode or a delay. They matter when the neglected behavior acts on time scales relevant to the controller. This differs from changing only the numerical parameters in a fixed equation model.
For an original example, a nominal plant satisfies , and state feedback control commands . The nominal prediction is , with decay rate . Now suppose the actual applied input follows a first-order actuator:
Eliminating gives
With and in chosen consistent units, the exponents are : the actual response can oscillate while decaying, whereas the nominal model predicts a single exponential. The additional state changes the transient even though the nominal static command remains the same.
The omitted actuator also admits a frequency-domain description. Relative to an ideal actuator with gain1, its transfer function is . The corresponding multiplicative uncertainty is
The relative discrepancy is small when is small but approaches1 at high frequency. A low-frequency fit alone therefore need not justify a high-bandwidth design.
Dynamic uncertainty models can describe a family by bounds on its frequency response, with a shaping filter expressing which frequency ranges are poorly known. MathWorks's ultidyn documentation describes this construction for uncertain linear time-invariant systems.[ref] Such an uncertainty family must be chosen to cover the omitted behavior; its mere presence in a model does not prove that every physical discrepancy has been included.
Zero-input response 2026-11-05
The zero-input response is the output due to a system's initial internal state when its externally applied input is set to zero. It describes memory left by earlier excitation. For a linear time-invariant system, linearity separates this from the zero-state response due to the subsequent input.
For constant matrices in , , setting and gives
The matrix exponential solves the homogeneous state equation. With an arbitrary input, the complete solution adds the zero-state response:
A zero initial output alone does not establish that the whole state is zero.
For an RC circuit with capacitor difference and no applied source, , so . The capacitor energy falls as ; the resistor receives the dissipated energy. The primary MIT lecture explicitly distinguishes the capacitor's initial state, zero external input and zero initial state.[ref]
At a time constant and , the modeled difference after two seconds is . Zero input has not meant zero output or zero stored energy. A direct experiment charges the same capacitor to different known initial differences, removes the charging source and records the decay across a fixed resistor. The amplitudes should scale with while the normalized decay remains the same in the linear model.
A zero-input response need not decay in every system: an ideal undamped oscillator retains its initial motion, and an unstable homogeneous system can amplify a nonzero initial state. The transfer function describes the zero-state input-output map, so its product with a zero input cannot represent a separate nonzero initial-state contribution.
Zero-state response 2026-11-05
The zero-state response is the output produced by an applied input when the system starts with zero internal state. For a linear time-invariant system it separates the effect of the input from the zero-input response caused by an initially stored state. A zero initial output alone need not mean zero internal state.
For a causal linear time-invariant system with impulse response , the zero-state response is the convolution
An instantaneous feedthrough term, if present, is represented by a Dirac delta term in . Equivalently, for constant matrices in , with ,
For nonzero , linearity adds the zero-input response . Here the matrix exponential propagates the state left by the earlier input.
For example, with , and a constant input for has the step response . At one time constant it has reached about of its final value. If the initial value instead equals , the total output is . A charged capacitor in an RC circuit provides a direct physical example: discharge it before repeating the same input step to isolate the zero-state response, and compare with a run from a known initial electric potential difference.[ref]
With compatible transforms, the zero-state input-output relation is , where is the transfer function. For a nonzero initial state, a unilateral Laplace transform includes additional initial-condition terms. Those terms cannot be absorbed into an input-independent transfer function multiplying .