NewAI Part GeneratorFree to get started
FREE TOOL

LC filter calculator

Given L, C and the load, this LC filter calculator finds the cutoff, natural frequency, Q and peaking of a second-order low-pass or high-pass filter. It also works the other way, giving L and C values for a Butterworth, Bessel or critically damped response. Source resistance and the response at any frequency are included.

Solve for
Examples
Filter type
H
F
Ω
Ω

Optional: driver output resistance (plus the inductor DCR for a low-pass).

Hz

Optional: gain and phase at this frequency.

Cutoff frequency (−3 dB)

40.92 kHz

Natural frequency f0
41.15 kHz 1/(2π√LC)
Characteristic impedance √(L/C)
5.688 Ω
Quality factor Q
0.703 ζ = 0.711
Gain at f0 (vs passband)
-3.06 dB
Peak
none Q ≤ 0.707
−40 dB at
411.5 kHz

The inductor’s winding resistance (DCR) is in series with the source: add it to the source resistance. It lowers the passband gain and damps any peaking.

The load is taken as a pure resistance. A speaker, a motor or a regulator input is not, so check the filter at the extremes of its impedance.

Frequency response (Bode plot)
-80-60-40-200-150-100-5001k10k100k1MFrequency (Hz)Gain (dB)Phase (°)−3 dBfc 40.9 kHz
Gain (dB)Phase (°)

How do you calculate the LC filter cutoff frequency and Q?

The natural frequency is f0 = 1 ÷ (2π√(LC)) and, with a low-impedance source and a load RL across the capacitor (low-pass) or inductor (high-pass), Q = RL × √(C/L). For 10 µH, 1.36 µF and 2 Ω (TI’s class-D example) f0 = 43.2 kHz and Q = 0.738, so the −3 dB cutoff is 44.9 kHz. For a flat (Butterworth) response choose Q = 0.707: L = √2·RL ÷ (2πfc) and C = 1 ÷ (√2·2πfc·RL).

LC filter formula: cutoff frequency, Q and transfer function

A low-pass LC filter puts the inductor in series and the capacitor across the load; the high-pass swaps them. Each is a second-order section, so every LC filter equation below comes down to a natural frequency and a Q.

f0 = 1 ÷ (2π√(LC))     Q = RL × √(C ÷ L) = RL ÷ Z0
f0
natural frequency in Hz
L, C
inductance in H, capacitance in F
RL
load resistance in Ω (across C for a low-pass, across L for a high-pass)
Z0
characteristic impedance √(L/C) in Ω

With a source resistance Rs the LC filter transfer function for the low-pass becomes H(s) = RL ÷ (s²LC·RL + s(L + Rs·RL·C) + Rs + RL), so:

ω0² = (Rs + RL) ÷ (L·C·RL)
Q = (Rs + RL) ÷ (ω0 (L + Rs·RL·C))
passband gain = RL ÷ (Rs + RL)

For the high-pass, ω0² = RL ÷ (L·C·(Rs + RL)) and Q = RL ÷ (ω0 (L + Rs·RL·C)). The −3 dB point and the peak of the low-pass follow from Q; for the high-pass, f0 ÷ f3dB takes the same value and the peak sits at f0 divided by the same factor:

(f3dB ÷ f0)² = (1 − 1/(2Q²)) + √((1 − 1/(2Q²))² + 1)
peak = Q ÷ √(1 − 1/(4Q²)) at f0 × √(1 − 1/(2Q²))

LC filter design for a Butterworth response (Q = 0.707)

With no source resistance, setting Q = 1/√2 and ω0 = 2πfc gives TI’s design equations for a flat response:

L = √2 × RL ÷ ω0     C = 1 ÷ (√2 × ω0 × RL)

With a source resistance the calculator solves the quadratic for L exactly. A solution exists only when Rs ÷ (Rs + RL) ≤ 1 ÷ (4Q²), so equal source and load resistances top out at Q = 0.707.

How to calculate LC filter values: worked examples

Example 1: TI’s class-D output filter (10 µH, 1.36 µF, 2 Ω)

TI’s LC filter note reduces a 4 Ω bridge-tied speaker, with 10 µH in each output and 0.68 µF across the speaker, to a single-ended filter of 10 µH, 1.36 µF and 2 Ω. f0 = 1 ÷ (2π√(10 µH × 1.36 µF)) = 43.2 kHz and Q = 2 × √(1.36 ÷ 10) = 0.738 (TI truncates this to 0.737 and so prints −2.65 dB), so the level at f0 is -2.64 dB, the −3 dB cutoff is 44.9 kHz, and a 400 kHz switching frequency is attenuated by 38.7 dB.

Example 2: Butterworth low-pass, 40 kHz into 4 Ω

ω0 = 2π × 40 kHz. L = √2 × 4 ÷ ω0 = 22.5 µH and C = 1 ÷ (√2 × ω0 × 4) = 703 nF. The nearest E12 parts, 22 µH and 680 nF, give a cutoff of 40.9 kHz with Q = 0.703.

Example 3: a supply filter that peaks at light load

10 µH and 22 µF have f0 = 10.7 kHz and √(L/C) = 674 mΩ. Into 1 Ω, Q = 1.48 and the response peaks by 3.9 dB at 9.43 kHz. At a 10 Ω load Q rises to 14.8 and the peak to 23.4 dB: ripple near f0 is amplified, not filtered. This is why input and output LC filters get a damping network.

Butterworth LC low-pass filter values for common loads

L and C for a flat (Q = 0.707) second-order low-pass with no source resistance, for common loads. The high-pass uses the same values with L and C swapped in position.

Cutoff4 Ω load: L / C8 Ω load: L / C50 Ω load: L / C
1 kHz900 µH / 28.1 µF1.8 mH / 14.1 µF11.3 mH / 2.25 µF
3 kHz300 µH / 9.38 µF600 µH / 4.69 µF3.75 mH / 750 nF
10 kHz90 µH / 2.81 µF180 µH / 1.41 µF1.13 mH / 225 nF
40 kHz22.5 µH / 703 nF45 µH / 352 nF281 µH / 56.3 nF
100 kHz9 µH / 281 nF18 µH / 141 nF113 µH / 22.5 nF

Low-pass response versus Q

Gain relative to the passband at multiples of f0 for four values of Q. For a high-pass read the rows as f0 / f.

f / f0Q = 0.5Q = 0.707Q = 1Q = 2
0.1×-0.09 dB0.00 dB0.04 dB0.08 dB
0.5×-1.94 dB-0.26 dB0.90 dB2.04 dB
1×-6.02 dB-3.01 dB0.00 dB6.02 dB
2×-13.98 dB-12.30 dB-11.14 dB-10.00 dB
5×-28.30 dB-27.97 dB-27.79 dB-27.65 dB
10×-40.09 dB-40.00 dB-39.96 dB-39.92 dB

Q only shapes the response within about an octave of f0. A decade above it every second-order low-pass is close to −40 dB whatever its Q (-40.09 dB at Q = 0.5, -39.92 dB at Q = 2), so a high Q buys a sharper knee, not more stopband attenuation.

Real inductors, capacitors and loads

The inductor’s winding resistance (DCR) adds to the source resistance in a low-pass, which costs passband gain and damps any peaking; enter it in the source field. Above its saturation current the inductance drops, so size the inductor for the peak current, not the average. Every inductor and capacitor also has a self-resonant frequency, above which the filter stops attenuating; check it against the highest frequency you need to reject. Class 2 ceramic capacitors lose capacitance under DC bias, which raises f0.

The calculator treats the load as a resistor. A loudspeaker’s impedance varies with frequency, and a switching regulator’s input looks like a negative resistance at low frequency, which can turn a high-Q input filter into an oscillator. Design for the lightest load you expect, and simulate the circuit when the load is not resistive.

Glossary

Natural frequency (f0)
The pole frequency of the second-order response, 1/(2π√(LC)) with no source resistance.
Characteristic impedance
√(L/C), the impedance level of the LC pair. Q with a parallel load is RL divided by it.
Q (quality factor)
How lightly the filter is damped. 0.707 is maximally flat (Butterworth); higher values peak and ring.
Damping ratio (ζ)
1/(2Q). ζ = 1 (Q = 0.5) is critical damping: the fastest step response with no overshoot.
Butterworth response
The flattest passband possible for a given order, with the −3 dB point exactly at f0.
Doubly terminated
A filter driven from a source resistance and loaded by an equal resistance, as in 50 Ω RF systems.

Related guides

Frequently asked questions

What is the LC filter cutoff frequency formula?

The natural frequency is f0 = 1 / (2π√(LC)). The −3 dB cutoff equals f0 only for a Butterworth response (Q = 0.707); in a low-pass a higher Q puts the −3 dB point above f0 with a peak just below f0, and a lower Q puts it below f0 (a high-pass mirrors this). The calculator gives both.

How do I calculate L and C for a Butterworth LC low-pass filter?

With a low-impedance source and a load RL: L = √2 × RL / (2π fc) and C = 1 / (√2 × 2π fc × RL). For 40 kHz into 4 Ω that is 22.5 µH and 703 nF.

What is the LC filter Q factor with a resistive load?

Q = RL × √(C/L), the load resistance divided by the characteristic impedance √(L/C). A heavier load (smaller RL) damps the filter; a light load lets it peak. 10 µH and 22 µF have √(L/C) = 674 mΩ, so Q = 1.48 at 1 Ω and 14.8 at 10 Ω.

Why does my LC filter have a peak at the cutoff?

The load is too light for the L/C ratio, so Q is above 0.707. The peak is Q / √(1 − 1/(4Q²)) at f0 × √(1 − 1/(2Q²)): about +23.4 dB for Q = 14.8. Damp it with a series RC across the capacitor or design for the lightest load you expect.

What sets LC filter damping?

The damping ratio is ζ = 1/(2Q), so the L/C ratio, the load and any source resistance set it: a heavier load damps the filter, and series resistance damps any peaking. ζ = 1 (Q = 0.5) is critical damping, the fastest step response with no overshoot. If the load alone is too light, add a series RC across the capacitor of a low-pass or a resistor across the inductor of a high-pass.

How steep is an LC filter compared with an RC filter?

An LC filter is second order, so it falls 40 dB per decade beyond the cutoff, against 20 dB per decade for an RC filter. A Butterworth LC low-pass is 40 dB down at ten times its cutoff.

Does the source resistance change an LC filter?

Yes. In a low-pass it adds in series with L, lowering the passband gain to RL/(Rs + RL) and damping any peaking. With equal source and load resistances, as in a 50 Ω system, the highest Q possible is 0.707, and the Butterworth values become L = √2·R/ω0 and C = √2/(R·ω0): 1.13 µH and 450 pF at 10 MHz.

How do I use this for a class-D amplifier output filter?

Analyze the single-ended equivalent. TI’s LC filter note converts a bridge-tied (BTL) filter with one capacitor across the speaker into a single-ended filter with RL = R_speaker / 2 and C = 2 × C_BTL. Design for the lowest speaker impedance you support, and check the response at the highest one.

More free PCB tools

See all free electronics tools →

Sources

Design the whole board, not one number

ProtoFlow is free AI PCB design software: describe a circuit and get a schematic with real parts, PCB layout with an autorouter, SPICE simulation and Gerber files. Native KiCad, Altium and EAGLE import and export. Free on macOS, Windows and Linux.

Get for Mac

Free to start. No card required.