LC filter resonance is the specific frequency where an inductor's reactance and a capacitor's reactance perfectly cancel each other out, causing the circuit's impedance to drop to near-zero in a series configuration or spike to near-infinity in a parallel configuration. When you build a filter to smooth a power supply or shape an audio signal, this resonant point dictates whether your circuit behaves like a well-damped shock absorber or an uncontrollable ringing bell. What changes in a real circuit at this frequency is the impedance profile and the phase margin; an undamped LC filter will exhibit massive voltage peaking or severe phase shifts that can push a closed-loop feedback system into violent oscillation.

The Core Takeaway: Resonance isn't just a theoretical peak on a Bode plot. In a physical installation, an unmanaged LC resonant peak will amplify switching noise instead of filtering it, or cause a voltage regulator to latch off due to overvoltage transients during load steps.

The Core Physics and the Resonance Formula

Inductors and capacitors store energy in fundamentally different ways. An inductor resists changes in current by storing energy in a magnetic field, while a capacitor resists changes in voltage by storing energy in an electric field. Because of this, their reactances move in opposite directions as frequency changes. Inductive reactance ($X_L$) increases with frequency, while capacitive reactance ($X_C$) decreases.

At one precise frequency, these two reactances become equal in magnitude but opposite in phase (180 degrees apart). This is the resonant frequency ($f_r$). According to All About Circuits, the mathematical derivation for this point is straightforward:

$f_r = \frac{1}{2\pi\sqrt{LC}}$

Where:
$f_r$ = Resonant frequency in Hertz (Hz)
$L$ = Inductance in Henrys (H)
$C$ = Capacitance in Farads (F)

In a purely theoretical, lossless circuit, the impedance at this exact point would be zero (series) or infinite (parallel). In reality, parasitic resistances—specifically the Equivalent Series Resistance (ESR) of the capacitor and the DC resistance (DCR) of the inductor wire—limit the peak. This limitation is defined by the Quality factor (Q). A high Q means a sharp, dangerous peak; a low Q means a broad, manageable response.

Worked Numeric Example: Buck Converter Output Filter

Let's look at a real-world scenario: designing the output filter for a 5V, 3A buck converter. You need to filter out the switching ripple (typically 500 kHz), so you select a standard power inductor and a bank of ceramic capacitors.

  • Inductor (L): 10 µH (Shielded ferrite, DCR = 15 mΩ)
  • Capacitor (C): 100 µF (Multilayer Ceramic Capacitor / MLCC, ESR = 3 mΩ)

Plugging these into the resonance formula:

$f_r = \frac{1}{2\pi\sqrt{10 \times 10^{-6} \times 100 \times 10^{-6}}}$

$f_r = \frac{1}{2\pi\sqrt{1 \times 10^{-9}}}$

$f_r = \frac{1}{2\pi \times 3.162 \times 10^{-5}} \approx 5,032 \text{ Hz}$

Your LC filter resonance sits at roughly 5.03 kHz. Because you used MLCCs with an extremely low ESR (3 mΩ), the Q factor of this filter is massive. If your load suddenly drops from 3A to 0A, the energy trapped in the 10 µH inductor has nowhere to dissipate quickly. It will ring into the 100 µF capacitor at 5.03 kHz, causing the output voltage to overshoot well past the 5V nominal rail, potentially tripping the converter's overvoltage protection (OVP) or damaging the load.

Where You Meet LC Filter Resonance in Practice

You will rarely see a bare LC circuit on a modern PCB without some form of damping. Here is where resonance actively fights you in practical designs:

Switch-Mode Power Supply (SMPS) Output Filters

As shown in the numeric example, the output LC filter of a buck or boost converter creates a double pole in the control loop transfer function at the resonant frequency. This double pole introduces up to 180 degrees of phase shift. If the converter's compensation network isn't designed to roll off the gain before this phase shift destroys the phase margin, the power supply will audibly squeal and oscillate.

EMI Input Filters and the Middlebrook Criteria

When you place an LC filter on the input of a switching regulator to meet EMI standards, you create a hidden trap. A switching regulator acts as a constant power load, which means it exhibits negative incremental input impedance. If the output impedance of your LC input filter peaks at resonance and exceeds the magnitude of the regulator's negative input impedance, the system will latch into instability. This is known as the Middlebrook criteria, and it is the number one reason prototype boards fail when long, inductive input cables are added.

Audio Crossovers and RF Tuning

In audio, LC resonance is intentionally used in bandpass filters to route specific frequencies to tweeters or woofers. In RF, a parallel LC 'tank' circuit is used to tune an antenna to a specific carrier frequency, relying on the massive impedance spike at resonance to reject all other signals.

Series vs. Parallel: The Most Common Confusion

What people most commonly confuse regarding LC filter resonance is the behavioral difference between series and parallel topologies, as well as the difference between the resonant frequency and the cutoff frequency.

Characteristic Series LC Resonance Parallel LC Resonance (Tank)
Impedance at $f_r$ Minimum (Approaches zero, limited by ESR/DCR) Maximum (Approaches infinity, limited by parallel leakage)
Current at $f_r$ Maximum (Acts as a short circuit) Minimum (Acts as an open circuit)
Primary Use Case Bandpass filters, notch (trap) filters RF oscillators, impedance matching, notch filters
Phase Shift Transitions from capacitive to inductive Transitions from inductive to capacitive

Resonance vs. Cutoff: In a second-order low-pass LC filter, the -3dB cutoff frequency ($f_c$) is mathematically related to the resonant frequency ($f_r$). However, if the filter is underdamped (high Q), the actual response will peak above 0dB right at $f_r$ before rolling off. Designers often mistakenly assume the signal is attenuated at $f_r$, only to find it is actually being amplified.

Taming the Peak: Practical Damping Techniques

If your oscilloscope shows ringing at the LC resonant frequency during a load transient, you must lower the Q factor. Here are the three standard bench fixes, ranked by efficiency:

  1. Swap the Capacitor Chemistry: Replace a portion of your low-ESR MLCC bank with a polymer or electrolytic capacitor. A 100 µF polymer cap might have an ESR of 15 mΩ, which provides natural, lossy damping without requiring extra components.
  2. Add an RC Snubber: Place a resistor and capacitor in series, then put that network in parallel with the main filter capacitor. The optimal snubber resistor is roughly $R_{snub} = \sqrt{\frac{L}{C}}$, and the snubber capacitor should be about $2 \times C$. This absorbs the high-frequency ringing energy and dissipates it as heat in the resistor.
  3. Add a Series Damping Resistor: Placing a small resistor (e.g., 0.5 Ω) in series with the inductor will kill the resonance peak dead. However, this ruins your power efficiency and creates a DC voltage drop, making it viable only for low-current signal filters, not power rails.

Frequently Asked Questions

How does capacitor ESR affect LC filter resonance peaking?

Capacitor ESR acts as a built-in damping resistor. In a low-pass LC filter, a higher ESR reduces the Q factor, flattening the impedance peak at the resonant frequency and preventing voltage overshoot during load transients. While power supply designers usually want the lowest possible ESR for ripple reduction, an ESR that is too low (like in pure ceramic MLCCs) creates an underdamped system that rings violently. This is why many reference designs intentionally mix high-ESR electrolytic caps with low-ESR ceramics.

Why does my LC input filter cause my switching regulator to oscillate?

This happens due to impedance interaction. A switching regulator draws constant power, meaning if input voltage drops, it draws more current, exhibiting a negative incremental input impedance. If the output impedance of your LC input filter peaks at its resonant frequency and exceeds the magnitude of the regulator's negative input impedance, the phase margin collapses. To fix this, you must heavily damp the LC input filter's resonant peak using an electrolytic capacitor or an RC snubber so that the filter's peak impedance remains well below the regulator's input impedance across all frequencies.

What is the difference between LC resonant frequency and cutoff frequency?

The resonant frequency ($f_r$) is the exact point where inductive and capacitive reactances cancel out. The cutoff frequency ($f_c$) is the point where the filter's output power drops by half (-3dB). In a perfectly damped second-order Butterworth LC low-pass filter, $f_r$ and $f_c$ are identical. However, in underdamped circuits (high Q), the signal actually peaks and amplifies at $f_r$ before eventually rolling off, meaning the true -3dB cutoff frequency is pushed much higher than the resonant frequency. Always simulate the Bode plot rather than relying solely on the $f_r$ formula for filter bandwidth.