Resonance & Q Factor#

Resonance is what happens when capacitive and inductive reactance cancel at a specific frequency. Energy sloshes back and forth between the electric field of the capacitor and the magnetic field of the inductor, and the circuit’s behavior at that frequency is dramatically different from its behavior at any other frequency. Resonance is the basis of filters, oscillators, and tuned circuits — and it’s also the mechanism behind parasitic ringing and unexpected frequency peaks in circuits that weren’t meant to resonate at all.

All resonance and Q formulas on this page are collected in the Formula Reference appendix for quick lookup.

LC Resonance#

At one specific frequency, an inductor’s rising reactance and a capacitor’s falling reactance meet and exactly cancel. All that remains is resistance — the reactive parts add to zero, and the circuit behaves as though the L and C aren’t there. That frequency is the resonant frequency.

The expression comes from setting X_L = X_C — that is, 2πfL = 1/(2πfC) — and solving for f:

f₀ = 1 / (2π√(LC))

At f₀, the two reactances are equal and opposite. What happens next depends on whether the L and C are in series or parallel.

Series Resonance#

Series RLC circuit — at resonance, impedance drops to R

In a series LC circuit, the impedance at resonance drops to its minimum — ideally zero (just the series resistance of the real components). At f₀, the series combination of L and C looks like a short circuit to the driving source (limited only by the resistance in the loop).

Below f₀: the circuit is capacitive (X_C > X_L, net negative reactance)
At f₀: purely resistive (reactances cancel, Z = R_series)
Above f₀: the circuit is inductive (X_L > X_C, net positive reactance)

Series resonance creates an impedance minimum. It’s used in series-tuned filters (bandpass) and is the mechanism behind crystal oscillator operation at the series resonant frequency.

Parallel Resonance#

Parallel RLC circuit — at resonance, impedance rises to R

In a parallel LC circuit, the impedance at resonance rises to its maximum. At f₀, the parallel combination of L and C looks like an open circuit (ideally infinite impedance, limited by losses).

Below f₀: the circuit is inductive (inductor has lower impedance, dominates the parallel combination)
At f₀: purely resistive (reactances cancel, Z = R_parallel, which can be very high)
Above f₀: the circuit is capacitive (capacitor has lower impedance, dominates)

Parallel resonance creates an impedance maximum. It’s the basis of tank circuits in RF amplifiers and oscillators, and it’s what causes the impedance peak at an inductor’s self-resonant frequency.

The Symmetry#

Series resonance: impedance minimum, current maximum. Parallel resonance: impedance maximum, current minimum (from the external source — internally, large circulating currents flow between L and C). Both occur at the same frequency f₀ = 1/(2π√LC) for the same L and C values.

Quality Factor (Q)#

Q measures how “sharp” the resonance is — the ratio of energy stored to energy lost per cycle. A high-Q circuit resonates strongly at f₀ and attenuates quickly away from it. A low-Q circuit has a broad, gentle resonance.

Q = 2π × (energy stored per cycle) / (energy dissipated per cycle)

The key to understanding Q in circuits is where the resistance sits relative to the energy circulation path between L and C. In a series RLC circuit, the resistance is in the circulation path — current flowing between L and C must pass through R, so lower R means less loss and higher Q. In a parallel RLC circuit, the resistance shunts energy out of the tank — current leaks through R instead of circulating between L and C, so higher R means less leakage and higher Q. This inversion is why the series and parallel Q formulas look reciprocal:

For a series RLC circuit:

Q = (1/R) × √(L/C) = X_L/R = X_C/R (all evaluated at f₀)

For a parallel RLC circuit:

Q = R × √(C/L) = R/X_L = R/X_C (all evaluated at f₀)

Q and Bandwidth#

A high-Q circuit is picky about frequency — it responds strongly right at resonance and barely at all a little ways away. The bandwidth of a resonant circuit — the range of frequencies where the response is within 3 dB of the peak — quantifies exactly how picky:

BW = f₀ / Q

A resonant circuit at 1 MHz with Q = 100 has a 3 dB bandwidth of 10 kHz. The higher the Q, the narrower the bandwidth, and the sharper the frequency selectivity.

What Q Values Mean in Practice#

Q < 1 — Overdamped. No real resonance, just a gentle roll-off. This is what many power supply filters look like intentionally
Q = 0.5 to 1 — Critically damped to slightly underdamped. Often the design target for maximally flat (Butterworth) filter responses
Q = 5 to 50 — Moderate Q. Typical of practical LC filters, tuned amplifier stages, and IF transformers
Q = 100 to 1000+ — High Q. Quartz crystals, ceramic resonators, cavity resonators. Very narrow bandwidth and sharp frequency selection
Q > 10,000 — Extremely high Q. Found in high-quality quartz crystals used for frequency standards and precision oscillators

Loaded vs. Unloaded Q#

The unloaded Q (Q_U) of a resonant circuit is its Q with no external load — determined only by the internal losses of the L and C (mostly the inductor’s resistance and core losses).

The loaded Q (Q_L) includes the effect of whatever is connected to the resonant circuit. External loads provide an additional path for energy to leave the resonant circuit, which always reduces Q:

1/Q_L = 1/Q_U + 1/Q_external

This matters a lot in practice. A high-Q inductor in a tank circuit might have Q_U of 200, but once coupled to a low-impedance load, Q_L drops to 20. The selectivity that looked great on the bench with nothing connected degrades when the circuit is put to use.

Where Resonance Shows Up#

Intentional Resonance#

Bandpass filters — Series or parallel LC circuits tuned to pass a desired frequency range. The Q determines the bandwidth
Oscillators — LC tank circuits, crystal oscillators, ceramic resonator circuits. The resonant element determines the oscillation frequency and Q determines frequency stability
Impedance matching networks — L-networks, pi-networks, and T-networks use resonant or near-resonant structures to transform impedance at a specific frequency
Tuned amplifiers — RF amplifiers with tank circuit loads that amplify only near the resonant frequency

Unintentional Resonance#

Parasitic ringing — Stray inductance (trace, lead, wiring) resonating with stray or intentional capacitance. Shows up as damped oscillation after switching edges
Component self-resonance — Every real capacitor and inductor has an SRF where it stops being what it’s supposed to be. This is LC resonance between the component’s intended element and its dominant parasitic
PCB resonances — Power plane pairs, long traces, and connector pin inductances form resonant structures at frequencies determined by their geometry
Cable resonances — Unterminated cables are transmission line stubs that resonate at frequencies related to their length

Crystal Operation#

Quartz crystals are mechanical resonators with extremely high Q (typically 10,000 to 100,000+). They have both a series and parallel resonant frequency very close together. The series resonant frequency is where the crystal impedance is minimum (acts like a very low resistance). The parallel resonant frequency — slightly higher — is where the impedance is maximum. Crystal oscillator circuits are designed to operate at one or the other, and the crystal specification (series vs. parallel cut) must match the circuit type.

Why Every Real Circuit Has Unintended Resonances#

Any time there’s inductance and capacitance in proximity, there’s a potential resonance. In real circuits:

Every wire and PCB trace has inductance
Every pair of conductors has capacitance
Every component has parasitic L and C

The result: unintended LC combinations exist everywhere. Most are at frequencies far from the operating range and cause no trouble. But when a parasitic resonance lands near a clock frequency, switching frequency, or signal bandwidth, it can amplify noise, cause ringing, or create unexpected frequency peaks.

Identifying and damping unwanted resonances is a recurring theme in EMC work, power integrity, and high-speed digital design.

Tips#

Use f₀ = 1/(2π√LC) as a diagnostic tool — When you see ringing on a scope, measure the frequency. Then work backwards to figure out what L and C are resonating. The frequency often points directly to the culprit
Damping resistors kill unwanted resonance — A small resistor in series with an inductor (or parallel with a capacitor) lowers Q and suppresses ringing. The tradeoff is reduced filter performance or increased losses, but in many cases a controlled low-Q response is better than an unpredictable high-Q one
Check component SRF against operating frequency — If a capacitor’s SRF is below the operating frequency, it’s an inductor at that frequency. If an inductor’s SRF is below the operating frequency, it’s a capacitor. Neither is doing its intended job
Higher Q means more sensitivity to component tolerance — A high-Q filter needs tighter component values. A 5% tolerance on L or C shifts f₀ by up to 2.5%, which might be a large fraction of the bandwidth if Q is high

Caveats#

Resonance can amplify voltage or current beyond supply levels — In a series RLC circuit at resonance, the voltage across L or C individually can be Q times the source voltage. A Q of 50 means the cap might see 50× the input voltage. This is real and can damage components
Parallel resonant impedance is not infinite — It’s limited by losses (mostly inductor resistance and core losses). In practice, the impedance peak at parallel resonance is Q × X_L, which can be very high but is always finite
Q is frequency-dependent in real components — Inductor Q varies with frequency due to skin effect, proximity effect, and core losses. The Q at the designed resonant frequency might be quite different from the Q measured at a test frequency. Always check Q at the actual operating frequency
Series and parallel resonant frequencies aren’t exactly equal in real circuits — Losses and parasitic elements shift them slightly apart. For most practical purposes they’re close enough to treat as equal, but in high-Q crystal circuits the difference matters and is specified separately

In Practice#

Damped ringing after a switching edge is the signature of a parasitic resonance. The ring frequency identifies the LC product, and the decay rate indicates Q. More cycles before decay = higher Q = less damping
An LC filter with a sharp peak in its response near the cutoff frequency has higher Q than intended. Adding series resistance to the inductor (or choosing an inductor with higher DCR) flattens the response at the cost of insertion loss
A circuit that oscillates at a frequency unrelated to any clock or signal is likely hitting a parasitic resonance. Check for resonances between bypass cap ESL and nearby capacitance, or between trace inductance and load capacitance
Crystal oscillator that starts up slowly or not at all may have excessive loading that kills the Q below what the oscillator circuit needs to sustain oscillation. Check that the load capacitance matches the crystal specification