Electronics Guide

Voltage Coefficient of Resistance

The voltage coefficient of resistance (VCR) describes how a resistor's value changes with applied voltage. This effect is most pronounced in carbon composition and thick-film resistors, where the conduction mechanism involves tunneling or hopping between conductive particles through insulating material.

VCR is typically specified in parts per million per volt (ppm/V) and can range from essentially zero for wirewound resistors to several hundred ppm/V for carbon composition types:

• Wirewound resistors: Negligible VCR (less than 1 ppm/V) because conduction occurs through continuous metal wire
• Metal film resistors: Very low VCR (typically less than 5 ppm/V) due to metallic conduction
• Thick-film resistors: Moderate VCR (10-100 ppm/V) depending on formulation
• Carbon composition resistors: High VCR (200-500 ppm/V) due to granular conduction mechanism

Published VCR figures depend on geometry as much as on material, because the effect tracks the voltage gradient along the resistive element rather than the terminal voltage alone. A physically longer element at the same terminal voltage sees a lower gradient, so a higher-power part in a family usually specifies a lower VCR than a small chip of the same resistance. Comparing figures across families therefore requires knowing the test voltage and the element size behind each one.

In a voltage divider using high-VCR resistors, the division ratio becomes voltage-dependent, introducing distortion for AC signals. For precision applications, metal film or wirewound resistors should be selected to minimize this effect. When high-VCR resistors must be used, the voltage applied across them should be minimized.

Temperature Coefficient of Resistance

All resistors change value with temperature, characterized by the temperature coefficient of resistance (TCR). This parameter, typically expressed in ppm per degree Celsius, indicates how much the resistance changes per degree of temperature change.

Different resistor technologies exhibit vastly different TCR values:

• Precision metal film: TCR of ±5 to 25 ppm per degree Celsius, with matched temperature tracking available
• Standard metal film: TCR of ±50 to 100 ppm per degree Celsius
• Thick-film chip resistors: TCR of ±100 to 200 ppm per degree Celsius
• Carbon composition: TCR of ±500 to 1500 ppm per degree Celsius, often with non-linear behavior
• Wirewound: TCR varies widely; special alloys like manganin achieve less than 20 ppm per degree Celsius

The relationship is not always linear. Many resistors exhibit a parabolic temperature characteristic, with minimum resistance near room temperature and increasing resistance at both temperature extremes. Precision resistors are often specified with a maximum resistance change over a temperature range rather than a simple TCR value.

For ratio-dependent circuits like voltage dividers and differential amplifiers, matching the TCR of paired resistors is more important than achieving low absolute TCR. Resistor networks on a common substrate achieve excellent TCR matching because all elements experience the same temperature.

Parasitic Inductance and Capacitance

Every resistor possesses parasitic inductance from its lead wires and body construction, as well as parasitic capacitance between its terminals and between the resistive element and any surrounding conductors.

The equivalent circuit of a real resistor includes:

• Series inductance: Typically 1 to 10 nH for leaded resistors, arising from lead wires and the helical path of current in film resistors
• Parallel capacitance: Typically 0.1 to 1 pF for leaded resistors, arising from capacitance between terminals and distributed capacitance along the resistive element

These parasitics cause the impedance to deviate from the nominal resistance at high frequencies. For a typical 10 kilohm metal film resistor with 5 nH series inductance and 0.5 pF parallel capacitance:

• At 1 MHz: The impedance deviates from nominal by less than 0.1%
• At 10 MHz: The deviation approaches 5%
• At 100 MHz: The parallel capacitance dominates, reducing the impedance magnitude to about 3 kilohms, roughly a third of nominal

Wirewound resistors have the highest parasitic inductance, often tens of microhenries, making them unsuitable for RF applications unless specially wound in non-inductive configurations. Carbon composition and metal film resistors have lower inductance and are preferred for high-frequency circuits.

Surface-mount chip resistors have lower parasitics than leaded types due to their smaller physical size and lack of lead wires. This makes them the preferred choice for high-frequency applications.

Power Coefficient and Self-Heating

When current flows through a resistor, power dissipation causes the component to heat above ambient temperature. This self-heating changes the resistance according to the TCR, creating a power-dependent resistance effect.

The thermal resistance from junction to ambient, typically expressed in degrees Celsius per watt, determines how much temperature rise occurs for a given power dissipation. For a resistor with a thermal resistance of 200 degrees Celsius per watt and a TCR of 100 ppm per degree Celsius dissipating 100 mW:

Temperature rise = 0.1 W × 200 degrees Celsius per watt = 20 degrees Celsius

Resistance change = 20 × 100 ppm = 2000 ppm = 0.2%

The 20 in the second line is the self-heating rise computed in the first, so the powered resistor settles 0.2% above its unpowered value. The rise is a temperature span rather than an absolute temperature, which is what the temperature coefficient multiplies.

In precision applications, this effect can be significant even at modest power levels. Solutions include using resistors with low TCR, oversizing resistors to reduce temperature rise, or using constant-temperature mounting techniques.

The power coefficient, expressed in ppm per watt, combines thermal resistance and TCR to give a single specification for resistance change with power dissipation. Lower values indicate better stability under varying power conditions.

Noise in Resistors

All resistors generate thermal noise (Johnson-Nyquist noise) due to random electron motion. The open-circuit noise voltage spectral density is:

e_n = √(4kTR)

Here k is Boltzmann's constant, T the absolute temperature, and R the resistance. The result is a spectral density in volts per root hertz, not a total: converting it to an RMS voltage requires multiplying by the square root of the noise bandwidth. At 300 K a 1 kΩ resistor produces about 4.07 nV/√Hz and a 1 MΩ resistor about 129 nV/√Hz, the density growing only as the square root of resistance. Over a 1 MHz noise bandwidth the 1 kΩ figure becomes about 4.1 microvolts RMS.

This fundamental noise source cannot be eliminated; it depends only on resistance and temperature. However, resistors also generate excess noise above the thermal noise floor, primarily in the form of current noise (also called 1/f noise or flicker noise).

Current noise arises from fluctuations in the resistive material when DC current flows. Unlike thermal noise, it depends strongly on resistor construction:

• Wirewound resistors: Lowest current noise; metallic conduction produces minimal fluctuations
• Metal film resistors: Very low current noise; excellent for precision applications
• Thick-film resistors: Moderate current noise; acceptable for most applications
• Carbon composition: Highest current noise; the granular structure creates significant fluctuations

Current noise is specified as a noise index in decibels, referenced to one microvolt of RMS noise per volt of applied DC voltage in each decade of frequency; zero decibels corresponds to 1 microvolt per volt. Because the index equals 20 times the base-10 logarithm of this microvolt-per-volt ratio, a noise index of −20 dB corresponds to 0.1 microvolt of noise per volt of DC per decade, and −40 dB to 0.01 microvolt per volt. For low-frequency precision circuits carrying DC current, resistor current noise can exceed thermal noise and should be considered in component selection.

Voltage Coefficient of Capacitance

Many capacitor types, particularly those using ceramic dielectrics, exhibit significant capacitance change with applied DC voltage. This DC bias effect can remove well over half the nominal capacitance at rated voltage, and more than 80% of it in the smallest case sizes.

Ceramic capacitor classes exhibit dramatically different voltage coefficients:

• Class 1 (C0G/NP0): Negligible voltage coefficient (less than 0.1%); capacitance is essentially constant with voltage
• Class 2 (X7R, X5R): Strong DC bias effect; the loss at rated voltage ranges from roughly 20% for a modest capacitance in a large case to more than 80% for a high capacitance in a small one
• Class 3 (Y5V, Z5U): Severe voltage coefficient; capacitance can drop 50 to 80% at rated voltage

The spread within Class 2 is far wider than the dielectric code suggests, because the loss tracks the electric field in the ceramic rather than the voltage at the terminals. Packing more capacitance into a given footprint means thinner dielectric layers and a higher field per volt applied, so case size and capacitance density matter as much as the dielectric formulation. Two parts of the same capacitance and dielectric make the point: a 4.7 microfarad X5R rated at 25 V in an 0805 case retains only about 15% of its nominal value at that rated voltage, while a 4.7 microfarad X5R rated at 100 V in a 1210 case still holds about 70% at 25 V. The dielectric code predicts neither figure; only the manufacturer's DC bias curve for the specific part does.

The three-tier numbering here follows the IEC 60384 convention. The EIA RS-198 scheme that defines the C0G, X7R, and Y5V letter codes instead groups all high-permittivity dielectrics, including Y5V and Z5U, as Class 2, reserving Class 3 for the largely obsolete barrier-layer types.

This effect has critical implications for decoupling and filtering applications. A 10 microfarad Y5V capacitor rated at 16 V might provide only 2 to 3 microfarads when biased at 12 V. Operating well below the rated voltage recovers part of the loss, and moving to a larger case size or a higher voltage rating at the same capacitance recovers more, because both reduce the field in the dielectric.

Film capacitors and electrolytic capacitors have much lower voltage coefficients, typically less than a few percent over their operating range. For applications requiring stable capacitance regardless of voltage, C0G ceramics, film capacitors, or mica capacitors are preferred.

Temperature Coefficient of Capacitance

Capacitance varies with temperature due to dimensional changes and changes in dielectric properties. The temperature coefficient depends strongly on dielectric type:

• C0G/NP0 ceramics: Temperature coefficient of 0 ± 30 ppm per degree Celsius; extremely stable
• X7R ceramics: ±15% capacitance change from −55 to +125 degrees Celsius
• X5R ceramics: ±15% capacitance change from −55 to +85 degrees Celsius
• Y5V ceramics: +22% to −82% capacitance change from −30 to +85 degrees Celsius
• Polyester film: Approximately plus 400 ppm per degree Celsius; positive temperature coefficient
• Polypropylene film: Approximately −200 ppm per degree Celsius; negative temperature coefficient
• Aluminum electrolytic: Large variations; capacitance typically decreases at low temperatures

The three-character codes state a limit rather than a coefficient. X7R means only that the capacitance stays within ±15% across the range; it says nothing about the shape of the curve between the endpoints, which for Class 2 dielectrics peaks near room temperature and falls off at both extremes. The limit also applies with no DC bias applied, so the bias loss described above adds to it rather than being included in it.

For timing circuits and oscillators where frequency stability depends on capacitance stability, C0G ceramics or polystyrene capacitors (now rare) provide the best temperature performance. For general decoupling, the temperature variation of X7R capacitors is usually acceptable.

Equivalent Series Resistance and Dissipation Factor

Real capacitors have losses that can be modeled as an equivalent series resistance (ESR). This resistance dissipates power when AC current flows through the capacitor and limits the effectiveness of capacitors in filtering and decoupling applications.

ESR depends on the capacitor's construction and varies with frequency and temperature:

• Ceramic capacitors: Very low ESR, typically milliohms; ideal for high-frequency decoupling
• Film capacitors: Low ESR, typically tens of milliohms; good for audio and intermediate frequencies
• Aluminum electrolytic: Higher ESR, typically tens of milliohms to several ohms; ESR increases dramatically at low temperatures
• Tantalum electrolytic: Moderate ESR, typically tens to hundreds of milliohms

The dissipation factor (DF, also written tan δ) expresses losses as the ratio of ESR to capacitive reactance at a stated frequency:

DF = ESR/X_C = ESR (2πfC)

A dissipation factor of 0.01 (1%) means the capacitor dissipates 1% of the reactive power as heat. DF is the reciprocal of the capacitor's Q, so those same losses amount to a Q of 100. High-quality film and ceramic capacitors achieve a DF below 0.1%, a Q above 1000, while electrolytic capacitors may reach several percent, a Q below 50.

Because ESR varies strongly with frequency, a dissipation factor quoted without its measurement frequency is incomplete. Manufacturers measure it at conventional points that differ by family: 120 Hz for aluminum and tantalum electrolytics, 1 kHz for film capacitors and for ceramics above 1000 pF, and 1 MHz for ceramics of 1000 pF and below. A figure taken at 120 Hz and one taken at 1 MHz describe different operating conditions and cannot be compared.

ESR is particularly important in power supply applications. High ESR limits the capacitor's ability to filter high-frequency switching noise and can cause significant power dissipation in ripple current applications. Capacitor manufacturers specify maximum ripple current ratings based on the power dissipation from ESR heating.

Equivalent Series Inductance

Capacitor leads and internal construction create equivalent series inductance (ESL) that limits high-frequency performance. At some frequency the inductive and capacitive reactances cancel and the capacitor self-resonates. Because ESL and capacitance sit in series, that resonance is a series resonance: the impedance falls to the ESR alone, its lowest value anywhere. Above the self-resonant frequency the ESL dominates and the impedance climbs again, so the component behaves as an inductor rather than a capacitor.

Typical ESL values and self-resonant frequencies vary by construction:

• Leaded ceramic capacitors: ESL of 2 to 5 nH; with a 0.1 microfarad part that places self resonance near 11 MHz at 2 nH and near 7 MHz at 5 nH, dropping to about 5 MHz once 10 nH of lead and loop inductance is counted
• Surface-mount ceramics: ESL of 0.5 to 1 nH; the same 0.1 microfarad value then self-resonates between roughly 16 and 23 MHz
• Aluminum electrolytic: ESL of 10 to 50 nH; self-resonant frequency often below 1 MHz for large values
• Tantalum capacitors: ESL of 1 to 5 nH; moderate self-resonant frequencies

A capacitor presents its lowest impedance at its self-resonant frequency, so useful decoupling extends somewhat above that point; what ends it is the rising ESL reactance, not the resonance itself. Above resonance a smaller value with lower ESL presents a lower impedance than a larger one, which is why decoupling networks combine several values.

Mixing values carries a penalty the simple argument omits. Two capacitors in parallel form a loop in which the smaller part's inductance resonates against the larger part's capacitance, and at that anti-resonance the pair presents an impedance peak rather than a minimum. A 10 microfarad part with 2 nH of loop inductance self-resonates near 1.1 MHz and a 100 nanofarad part with 1 nH near 16 MHz; between them, near 9 MHz, the parallel combination peaks. Holding the peak down means keeping loop inductance small and leaving enough resistance in the path to damp it, which is why several identical capacitors, or a deliberately damped spread of values, often outperform a decade ladder of single parts.

Dielectric Absorption

Dielectric absorption (also called soakage or battery effect) causes a capacitor to recover some of its charge after being discharged. If a capacitor is charged, then briefly shorted, it will develop a voltage after the short is removed. This effect arises from polarization mechanisms in the dielectric that respond slowly to electric field changes.

Dielectric absorption is quoted as the percentage of the original charging voltage that reappears at the terminals. The figure depends entirely on the timing of the measurement, so the standard fixes it: IEC 60384-1 charges the capacitor at its rated DC voltage for 60 minutes, discharges it through a low resistance for 10 seconds, and takes the recovery voltage that appears over the following 15 minutes. Figures obtained under other protocols are not comparable with these.

• Polypropylene and polystyrene film: About 0.02%; the best readily available dielectrics, and the reason these types dominate sample-and-hold service
• Polyphenylene sulfide film: 0.05 to 0.1%
• Polyester film: 0.2 to 0.5%; acceptable for many applications
• C0G/NP0 ceramic: About 0.6%
• X7R ceramic: 0.6 to 1.0%
• Z5U ceramic: 2.0 to 2.5%
• Solid tantalum: 2 to 3%
• Aluminum electrolytic: 10 to 15%; unsuitable for precision timing

The C0G figure surprises designers who assume the most temperature-stable ceramic must also be the best behaved in a hold circuit. Temperature stability and dielectric absorption are separate properties, and a Class 1 ceramic is excellent at the first while showing roughly thirty times the absorption of polypropylene. The gap between C0G and X7R, by contrast, is small: absorption does not separate the ceramic classes the way temperature coefficient and DC bias do.

Three mechanisms are easily conflated, and keeping them apart matters when reading a datasheet. Dielectric absorption is reversible charge storage: slow polarization returns charge after the terminals are released, and it clears once the part has rested. Leakage is a steady current through the dielectric that continues for as long as voltage is applied and rises with voltage and temperature. Aging is the permanent, logarithmic loss of capacitance in a ferroelectric Class 2 or Class 3 ceramic, which resting does not reverse and only heating above the Curie point resets. A dielectric may be good at one and poor at another: polypropylene has very low absorption and very low leakage and, not being ferroelectric, does not age at all, while C0G does not age yet absorbs thirty times as much as polypropylene.

In sample-and-hold circuits, integrators, and other precision applications, dielectric absorption causes errors by allowing charge from previous states to influence current measurements. Low-absorption capacitors like polypropylene or polystyrene are essential for these applications.

Frequency Dependence of Capacitance

The permittivity of a dielectric falls once the polarization mechanisms that create it can no longer follow the applied field, so the capacitance a part exhibits depends on the frequency at which it is measured. Two distinct effects produce that dependence, and confusing them leads to badly wrong expectations.

The first is genuine dielectric dispersion, a property of the material. It is gradual, and it is why the measurement frequency is part of the specification: Class 2 ceramics above 1000 pF are specified at 1 kHz, while Class 1 ceramics and values of 1000 pF and below are specified at 1 MHz. The second is the impedance rise above self resonance, which an instrument assuming an ideal capacitor reports as a collapse in capacitance even though the dielectric has not changed at all. The abrupt losses seen at tens of megahertz are almost entirely the second effect.

• Class 2 ceramics (X7R, X5R): Measurable dispersion above the 1 kHz reference frequency, steeper in the higher-permittivity formulations
• Y5V and Z5U ceramics: The steepest dispersion of the common dielectrics, consistent with their large temperature and bias sensitivity
• C0G ceramics: Paraelectric rather than ferroelectric, with no significant dispersion; the practical limit is the part's own self resonance, not its dielectric
• Film capacitors: Very low dispersion; the useful range again ends where ESL takes over

Taken together, dispersion and ESL mean that a high-permittivity ceramic may provide far less decoupling at a switching frequency than its low-frequency capacitance suggests. The impedance-versus-frequency curve, not the nominal value, is what a designer should select against.

Aging in Ceramic Capacitors

Class 2 and Class 3 ceramic capacitors exhibit aging, a gradual decrease in capacitance over time due to relaxation of the ferroelectric crystal structure. This aging follows a logarithmic time law, with a fixed percentage decrease per decade of time.

Typical aging rates are:

• X7R ceramics: Aging rate of 1 to 2% per decade of hours
• Y5V ceramics: Aging rate of 3 to 7% per decade of hours
• C0G ceramics: No significant aging

A decade of time means factors of 10 in hours: 1 hour, 10 hours, 100 hours, 1000 hours, and so on. A capacitor with 2% per decade aging rate that has aged for 10,000 hours, a little over a year, will have lost approximately 8% of its initial capacitance (four decades from 1 hour to 10,000 hours).

Heating the capacitor above its Curie point resets the aging process, temporarily restoring capacitance. This occurs during soldering and explains why capacitors should be allowed to age for 24 to 48 hours after soldering before final measurement.

Core Saturation Effects

Inductors using magnetic cores (ferrite, iron powder, or laminated steel) exhibit saturation when the magnetic flux density exceeds the core material's capability. As the core approaches saturation, permeability decreases, causing inductance to drop. At full saturation, the inductance approaches that of an air-core inductor, often a factor of 100 or more below nominal.

Saturation characteristics vary by core material:

• Ferrite materials: Saturation flux density of 0.3-0.5 Tesla; sharp saturation knee
• Iron powder cores: Saturation flux density of 1-1.5 Tesla; gradual saturation characteristic
• Laminated steel: Saturation flux density of 1.5-2 Tesla; used for power frequencies
• Permalloy and similar: Very high permeability but low saturation; easily saturated

The gradual saturation characteristic of iron powder cores is often preferred in power applications because it provides soft saturation, maintaining some inductance even when overdriven. Ferrite cores, with their sharp saturation knee, can cause sudden loss of inductance and potential circuit malfunction if ratings are exceeded.

For DC bias applications, manufacturers specify inductance versus DC current curves showing how inductance decreases with bias current. Manufacturers define the saturation current rating at a stated inductance drop, commonly 10, 20, or 30 percent, so two parts carrying the same rated current are not necessarily comparable until the criterion behind each rating is known. A separate rating, the RMS or heating current, is set by temperature rise rather than by saturation, and in a switching converter both must be satisfied: the peak current against the saturation rating and the RMS current against the thermal one.

DC Resistance and Copper Losses

The wire used to wind an inductor has resistance that dissipates power when current flows. This DC resistance (DCR) creates several effects:

• Power loss: P = I² DCR causes heating and reduces efficiency, with the AC resistance replacing DCR once skin and proximity effects matter
• Voltage drop: DC current creates a voltage drop across the DCR
• Q factor limitation: DCR limits the quality factor at low frequencies

DCR is minimized by using larger wire gauges, but this increases inductor size and cost. Power inductors for DC-DC converters are optimized for low DCR to maximize efficiency, while high-inductance coils may have higher DCR due to the many turns of fine wire required.

At low frequency, where the winding resistance is essentially its DC value, the quality factor follows:

Q = 2πfL/DCR

That expression is a low-frequency limit, not a general result. As frequency rises, skin effect, proximity effect, and core loss add to the series resistance, so the correct denominator is the total AC resistance rather than DCR. Substituting DCR alone makes Q grow without bound: a 10 microhenry part with 50 milliohms of DCR would show a Q of 1257 at 1 MHz and 12,570 at 10 MHz, whereas a real inductor's Q peaks and then falls. The expression also loses meaning at and above the self-resonant frequency described below, where the winding's distributed capacitance rather than its inductance controls the impedance.

Core Losses

Magnetic cores dissipate energy through hysteresis losses and eddy current losses. These losses increase with frequency and flux density, limiting inductor performance in high-frequency applications.

Hysteresis loss arises from the energy required to reverse magnetic domains as the core is magnetized in alternating directions. The loss per cycle is proportional to the area of the B-H hysteresis loop:

• Power ferrites: Optimized for low hysteresis loss at power frequencies
• RF ferrites: Lower permeability but lower losses at high frequencies
• Iron powder: Distributed air gaps reduce hysteresis but also reduce permeability

Eddy current loss occurs when the changing magnetic field induces currents in the conductive core material. These circulating currents dissipate energy through resistive heating. Eddy current loss is minimized by:

• Laminated cores: Thin insulated laminations increase resistance to eddy current flow
• Ferrite materials: High resistivity limits eddy current magnitude
• Iron powder cores: Distributed particles with insulating binder limit eddy currents

Core loss is typically specified as a loss factor or by curves showing loss versus frequency and flux density. Exceeding the manufacturer's recommended operating conditions can result in thermal runaway as heating increases losses, which increase heating further.

Skin Effect and Proximity Effect

At high frequencies, current distribution in the inductor winding becomes non-uniform, increasing effective resistance above the DC value. Skin effect forces current to flow near the conductor surface, while proximity effect causes current to concentrate on the sides of adjacent conductors facing each other.

The skin depth, the distance over which current density falls to 1/e, or about 37%, of its surface value, is:

δ = √(ρ/πfμ)

Here ρ is the resistivity, f the frequency, and μ the absolute permeability of the conductor, μ = μ_rμ₀, not the relative permeability on its own. Copper has a relative permeability of essentially 1, so μ reduces to μ₀. A magnetic conductor is a different matter: nickel, widely used as an underplate beneath gold, is ferromagnetic, and its skin depth at a given frequency is a small fraction of copper's, which is one reason plating metallurgy affects high-frequency conductor loss.

For copper at room temperature, taking ρ = 1.72 × 10⁻⁸ ohm-meters:

• At 60 Hz: Skin depth is approximately 8.5 mm
• At 10 kHz: Skin depth is approximately 0.66 mm
• At 1 MHz: Skin depth is approximately 66 micrometers
• At 100 MHz: Skin depth is approximately 6.6 micrometers
• At 1 GHz: Skin depth is approximately 2.1 micrometers

Because the depth falls as the inverse square root of frequency, it shrinks by a factor of about 3.16 for every decade. A conductor much thicker than a few skin depths buys no further reduction in AC resistance, which is why hollow tubing serves as well as solid bar for RF current and why copper foil thicker than the skin depth wastes weight.

Proximity effect is particularly severe in multilayer windings where conductors in adjacent layers carry current in the same direction. The AC to DC resistance ratio can exceed 10:1 in poorly designed windings at high frequencies.

Mitigation techniques include using Litz wire (multiple insulated strands woven to equalize current distribution), flat ribbon conductors, or single-layer windings. For high-frequency applications, these techniques are essential for achieving acceptable Q factors and efficiency.

Distributed Capacitance and Self-Resonance

Capacitance exists between adjacent turns of an inductor winding, creating a distributed parasitic capacitance. This capacitance can be modeled as an equivalent parallel capacitance that resonates with the inductance at the self-resonant frequency (SRF).

The resonant frequency follows the same relation as for a capacitor, f = 1/(2π√(LC)), with C the winding's distributed capacitance. The difference is the topology: that capacitance sits in parallel with the inductance rather than in series with it, so the resonance is a parallel one. The impedance therefore reaches a maximum at the SRF, the mirror image of the minimum a capacitor shows at its own self resonance. Above SRF, the component behaves as a capacitor rather than an inductor. The useful frequency range of an inductor is typically limited to well below its SRF, often to one-third or one-quarter of the SRF value.

Self-resonant frequency depends on construction:

• Multi-layer wound inductors: Lower SRF due to inter-layer capacitance
• Single-layer inductors: Higher SRF but limited inductance per unit size
• Bank-wound or progressive-wound: Techniques to maximize SRF
• Chip inductors: Small size limits parasitic capacitance, achieving high SRF

For RF applications, component selection must consider SRF. Using an inductor too close to its SRF reduces effective inductance and increases loss.

Temperature Effects on Inductors

Inductor characteristics change with temperature through several mechanisms:

• Winding resistance: Copper resistance increases approximately 0.4% per degree Celsius
• Core permeability: Ferrite permeability varies with temperature, often with a peak near room temperature
• Saturation: Core saturation flux density decreases at elevated temperatures
• Core losses: Often exhibit a minimum at some temperature, increasing at both lower and higher temperatures

Temperature-compensated inductors use materials chosen to minimize inductance change over temperature. For precision applications, inductors may need temperature-controlled mounting or compensation circuits.

In power applications, self-heating must be considered. Core losses and winding losses both generate heat, raising the inductor temperature above ambient. If the temperature rise causes increased losses (as with copper resistance), a thermal runaway condition can occur. Proper thermal design ensures that operating temperature remains within safe limits.

Thermal Time Constants

Components do not change temperature instantaneously; they have thermal mass and thermal resistance to their environment. The thermal time constant describes how quickly a component's temperature responds to changes in power dissipation or ambient temperature.

Typical thermal time constants range from milliseconds for small chip components to minutes for large power components. During power-up or when operating conditions change, component parameters drift as temperatures equilibrate. This can cause circuit behavior to vary during warm-up periods.

In precision circuits, warm-up time specifications account for thermal stabilization. Keeping components at constant temperature through controlled environments or allowing adequate warm-up time can significantly improve performance.

Thermal Matching

When circuit accuracy depends on component ratios rather than absolute values, matching temperature coefficients becomes critical. If two resistors in a voltage divider have identical TCR, their ratio remains constant as temperature changes even though both resistance values drift.

Techniques for achieving thermal matching include:

• Using matched component pairs: Manufacturers offer matched resistor and capacitor pairs with correlated temperature coefficients
• Integrated component networks: Components on a common substrate track temperature together
• Thermal coupling: Mounting matched components in close physical proximity and on the same thermal mass ensures they experience the same temperature
• Avoiding thermal gradients: Keeping heat-dissipating components away from precision components

Low-Temperature Effects

At low temperatures, several component issues emerge:

• Electrolytic capacitors: ESR increases dramatically; capacitance decreases
• Film capacitors: Some dielectrics become brittle and crack
• Ferrite inductors: Permeability changes; may exhibit different saturation characteristics
• Resistors: Carbon composition types may change value permanently

For military and aerospace applications operating at extended temperature ranges, special component grades and testing are required. The conventional grades are commercial, ordinarily 0 to 70 degrees Celsius; industrial, ordinarily −40 to +85 degrees Celsius; automotive passives qualified under AEC-Q200, whose Grade 1 spans −40 to +125 degrees Celsius; and military, −55 to +125 degrees Celsius. Outside its stated range a part is not merely out of specification but uncharacterized, which is the more serious problem: the datasheet offers no basis for predicting what it will do.

High-Temperature Effects

Elevated temperatures accelerate many degradation mechanisms and may cause immediate parameter shifts:

• Electrolytic capacitors: Electrolyte evaporation accelerates, shortening life
• Plastic film capacitors: Dielectric softening may cause shorts
• Ferrite cores: Approaching Curie temperature causes permeability collapse
• Resistors: May exhibit permanent resistance change after exposure to high temperature

Operating components at elevated temperatures requires derating. Manufacturers provide curves showing permissible voltage or current versus temperature. Exceeding these limits dramatically reduces component reliability and life.

Impedance Versus Frequency

Each passive component type shows characteristic impedance behavior across frequency:

• Resistors: Impedance equals nominal resistance at DC. As frequency increases, parasitic inductance causes impedance to rise slightly, then parasitic capacitance causes impedance to fall. The frequency at which a resistor begins to deviate from its nominal value depends on its construction; chip resistors maintain their nominal value to higher frequencies than leaded types.
• Capacitors: Impedance falls with increasing frequency (following 1/(2πfC)) until ESL causes the impedance to reach a minimum at the self-resonant frequency. Above resonance, impedance rises with frequency as the component becomes inductive.
• Inductors: Impedance rises with frequency (following 2πfL) until distributed capacitance causes resonance. Above the self-resonant frequency, the component becomes capacitive and impedance falls.

Quality Factor Frequency Dependence

Quality factor (Q) measures the ratio of energy stored to energy dissipated per cycle and varies strongly with frequency:

• Capacitors: Q is the reciprocal of the dissipation factor, Q = X_C/ESR, and it falls with frequency because the reactance drops faster than the ESR. The dissipation factors quoted earlier set the scale: 0.1% corresponds to a Q of 1000 and 1% to a Q of 100, and the best film dielectrics reach a few parts in ten thousand at low frequency, a Q of several thousand. Near self resonance Q falls to the order of tens, and at resonance the reactance cancels entirely and the quantity loses its meaning.
• Inductors: Q starts low at DC (where only DCR matters), rises with frequency as inductive reactance increases relative to resistance, then falls at high frequencies as core losses and skin effect dominate. The frequency of maximum Q depends on construction and may range from kilohertz to hundreds of megahertz.

For resonant circuits and filters, component Q determines the achievable selectivity and insertion loss. Using components with inadequate Q results in poor filter performance.

High-Frequency Component Models

At high frequencies, simple models become inadequate and more complete equivalent circuits are needed:

• Resistor model: A resistor becomes a resistance in series with parasitic inductance, all in parallel with parasitic capacitance. For chip resistors at microwave frequencies, distributed models may be necessary.
• Capacitor model: A capacitor becomes a capacitance in series with ESR and ESL, with a parallel leakage resistance. Multiple resonances may occur in large electrolytic capacitors due to their physical structure.
• Inductor model: An inductor becomes an inductance in series with frequency-dependent resistance (accounting for core loss, skin effect, and proximity effect), all in parallel with distributed capacitance.

Component manufacturers often provide S-parameter data or SPICE models that capture high-frequency behavior more accurately than simple L-R-C models.

Lead Inductance

Component leads and traces connecting components to the circuit add series inductance. Each millimeter of wire or trace contributes approximately 1 nH of inductance. For a typical leaded component with 10 mm total lead length:

• Inductance: Approximately 10 nH
• Reactance at 10 MHz: Approximately 0.6 ohms
• Reactance at 100 MHz: Approximately 6 ohms

This lead inductance can dominate the impedance of small-value capacitors at high frequencies, explaining why surface-mount components with minimal lead length are essential for RF and high-speed digital applications.

Stray Capacitance

Capacitance exists between any two conductors separated by a dielectric. In electronic circuits, stray capacitance appears between:

• Adjacent component leads: Typically 0.1 to 1 pF
• PCB traces: A 50-ohm microstrip on FR-4 holds roughly 0.1 pF per mm to its reference plane; coupling between two parallel traces is a fraction of that and falls quickly as the spacing grows
• Component body to ground plane: Can be several picofarads for leaded components

In high-impedance circuits, stray capacitance loads the circuit and can reduce bandwidth. In oscillators and filters, stray capacitance shifts resonant frequencies. Layout techniques to minimize stray capacitance include using ground plane cutouts, minimizing parallel trace runs, and separating high-impedance nodes from other conductors.

Mutual Inductance

Current flowing through one conductor creates a magnetic field that can couple into nearby conductors, inducing voltages. This mutual inductance can cause crosstalk between circuit paths and create feedback paths in amplifiers.

Mutual inductance is minimized by:

• Increasing separation: Coupling falls rapidly with distance
• Perpendicular orientation: Conductors at right angles have minimal coupling
• Shielding: Conductive shields block magnetic coupling (but shields must be continuous)
• Twisted pairs: Twisting signal and return conductors cancels external magnetic pickup

Parasitic Resonances

Combinations of parasitic inductance and capacitance create resonances that can cause unexpected circuit behavior. For example:

• Decoupling anti-resonance: Two capacitors of unequal value in parallel resonate against one another, and between their individual self-resonant frequencies the pair presents an impedance peak rather than the low impedance either provides alone. A single capacitor's own series resonance is the opposite case, the frequency of its lowest impedance
• Feedback loop resonances: Stray capacitance and inductance in amplifier feedback networks can cause oscillation
• Ground plane resonances: At high frequencies, ground planes can exhibit standing wave patterns

Identifying and controlling parasitic resonances requires understanding the complete physical structure of the circuit, not just the schematic. Electromagnetic simulation tools can help identify potential resonances before physical prototyping.

Resistance Drift

Resistors exhibit long-term drift through several mechanisms:

• Oxidation: Metal film resistors can oxidize, gradually increasing resistance
• Electromigration: Current flow can physically move material, changing resistance
• Stress relaxation: Mechanical stresses from manufacturing relax over time, changing resistance
• Moisture absorption: Humidity changes can affect thick-film and carbon composition resistors

High-stability resistors are specifically designed to minimize drift. Hermetic sealing prevents moisture effects, and controlled manufacturing processes minimize initial stresses. Such resistors can achieve drift rates below 25 ppm per year.

For less critical applications, standard metal film resistors typically drift by well under 0.1% per year in normal service, accumulating to a fraction of a percent over a long product life.

Capacitor Aging

Beyond the logarithmic aging of Class 2 and 3 ceramics, capacitors can degrade through:

• Electrolyte dry-out: Aluminum electrolytic capacitors lose electrolyte through the seal, reducing capacitance and increasing ESR
• Film degradation: Plastic film capacitors can suffer dielectric breakdown of weak spots, causing gradual capacitance loss
• Terminal corrosion: Environmental exposure can corrode terminations, increasing ESR

Electrolytic capacitor life depends strongly on temperature. A common rule of thumb states that life doubles for every 10 degrees Celsius of reduction in operating temperature. A capacitor rated for 2000 hours at 105 degrees Celsius might achieve 8000 hours at 85 degrees Celsius and even longer at lower temperatures.

Inductor Aging

Inductors are generally more stable than capacitors, but can still age:

• Core relaxation: Magnetic materials can slowly change permeability over time
• Winding insulation degradation: Thermal cycling can crack enamel insulation, potentially causing shorts
• Core cracking: Ferrite cores are brittle and can develop cracks from thermal or mechanical stress

Environmental Stress Effects

Environmental factors can accelerate aging:

• Thermal cycling: Repeated temperature changes stress materials and connections
• Humidity: Moisture penetration affects many component types
• Vibration: Mechanical stress can fatigue leads and internal connections
• Chemical exposure: Flux residues, cleaning solvents, and atmospheric contaminants can attack components

For long-life applications, environmental protection through conformal coating, hermetic packaging, or controlled atmosphere enclosures can significantly extend component life.

Piezoelectric and Microphonic Effects

Ceramic capacitors, particularly multilayer types using high-K dielectrics, exhibit significant piezoelectric behavior. Mechanical stress creates electrical charge, and conversely, applied voltage causes mechanical deformation.

This piezoelectric effect creates several issues:

• Microphonics: Vibration generates noise signals; ceramic capacitors on printed circuit boards can pick up acoustic noise
• Audible singing: Voltage variations cause mechanical vibration; capacitors in switching power supplies can emit audible noise
• Noise injection: Board flexure during handling can inject transient signals into sensitive circuits

Mitigation techniques include using C0G dielectrics (which have minimal piezoelectric effect), avoiding board flex during operation, and using polymer end-terminated capacitors that mechanically decouple the ceramic from the board.

Stress-Induced Parameter Shifts

Mechanical stress from soldering, board flexure, and thermal expansion can permanently shift component parameters:

• Resistors: Trimming-induced stress in precision resistors can cause drift until the stress relaxes
• Ceramic capacitors: Cracking from board flex can cause parameter shifts or complete failure
• Inductors: Core stress from mounting pressure can change permeability and inductance

For precision applications, components should be allowed to stabilize after soldering before final calibration. Temperature cycling after assembly can help relax stresses.

Thermal Expansion Mismatches

Different materials expand at different rates when heated, creating stress at interfaces. The coefficient of thermal expansion (CTE) mismatch between components and circuit boards is a major reliability concern:

• Ceramic components: CTE around 6-8 ppm per degree Celsius
• FR4 PCB material: CTE around 14-17 ppm per degree Celsius in-plane
• Copper: CTE around 17 ppm per degree Celsius

The CTE mismatch between ceramics and FR4 stresses solder joints and can crack ceramic bodies during thermal cycling. Larger components and wider temperature ranges increase stress. Using board materials with lower CTE or flexible termination capacitors can improve reliability.

Vibration and Shock Effects

Mechanical vibration and shock can damage components and create intermittent connections:

• Lead fatigue: Repeated flexing of component leads can cause fracture
• Solder joint failure: Vibration can fatigue solder joints, especially for heavy components
• Core movement: Inductor cores can shift within their structures
• Wire bond failure: Internal connections in chip components can break

For high-vibration environments, strain relief, conformal coating, and mechanical securing of heavy components are essential. Component selection should consider vibration ratings, and vibration testing should be included in qualification.

Derating Guidelines

Operating components below their maximum ratings improves reliability and reduces parameter variations:

• Voltage derating: Operating at 50 to 75% of rated voltage reduces stress and, for ceramics, improves capacitance stability
• Power derating: Keeping power dissipation well below ratings reduces self-heating and improves stability
• Temperature derating: Avoiding operation near maximum temperature limits extends life and reduces parameter drift
• Current derating: For inductors, staying below the saturation current limit maintains inductance

Component Selection Strategy

Selecting components requires balancing multiple factors:

• Identify critical parameters: Determine which non-idealities most affect circuit performance
• Match requirements to technology: Choose component technologies that minimize the critical non-idealities
• Consider the complete operating envelope: Account for temperature range, frequency range, and voltage range
• Include aging margins: Allow for parameter drift over the intended product life
• Verify with manufacturer data: Use complete specifications, not just headline values

Simulation and Modeling

Accurate simulation requires component models that include relevant non-idealities:

• Use manufacturer-provided models: Many component manufacturers provide SPICE models that include key parasitics
• Add parasitic elements: Include lead inductance, stray capacitance, and ESR in circuit models
• Perform corner analysis: Simulate with component values at their tolerance limits and over temperature
• Include frequency-dependent effects: For RF circuits, use S-parameter models rather than simple L-R-C models

Test and Verification

Physical testing complements simulation by revealing effects that may not be captured in models:

• Temperature testing: Verify performance across the full operating temperature range
• Frequency sweep: Measure impedances across the frequency range of interest
• Life testing: Accelerated aging tests predict long-term behavior
• Environmental stress screening: Thermal cycling and vibration testing reveal mechanical sensitivities