A second order low pass filter is an electronic circuit that allows low-frequency signals to pass while attenuating frequencies above a specific cutoff point at a steep rate of -40 dB per decade, utilizing two independent energy-storing components like capacitors or inductors.
The Core Mechanics: Understanding the -40 dB/Decade Roll-Off
While a basic first-order RC filter drops off at -20 dB per decade (halving the voltage for every doubling of frequency past the cutoff), a second order low pass filter doubles that slope to -40 dB per decade (or -12 dB per octave). This steeper attenuation is critical when you need to aggressively strip away high-frequency noise without eating into your desired baseband signal.
In practical circuit design, achieving this second-order response requires two reactive components. This can be done passively using an inductor and a capacitor (LC filter) or two cascaded RC stages buffered by an op-amp. However, simply cascading two passive RC filters without a buffer results in loading effects that ruin the transfer function. This is why active topologies, specifically the Sallen-Key architecture, are the industry standard for precision analog work.
Standard Component Values for Sallen-Key Active Filters
Designing a Sallen-Key filter from scratch requires solving complex quadratic equations for the Q-factor (damping). To save you the math, here is a data-dense reference table for a Unity-Gain Butterworth response (maximally flat passband, Q = 0.707). This configuration uses equal resistors ($R1 = R2$) and a capacitor ratio of roughly 2:1 ($C1 \approx 2 \times C2$).
| Target Cutoff ($f_c$) | R1 & R2 (E24) | C1 (E12) | C2 (E12) | Recommended Dielectric |
|---|---|---|---|---|
| 100 Hz | 11 kΩ | 220 nF | 100 nF | X7R Ceramic or Film |
| 1 kHz | 11 kΩ | 22 nF | 10 nF | C0G/NP0 Ceramic |
| 10 kHz | 11 kΩ | 2.2 nF | 1 nF | C0G/NP0 Ceramic |
| 100 kHz | 11 kΩ | 220 pF | 100 pF | C0G/NP0 Ceramic |
Critical Component Note: For cutoff frequencies above 1 kHz, you must use C0G/NP0 ceramic capacitors or polypropylene film. Standard X7R or Y5V ceramics exhibit severe voltage coefficients and microphonic effects, meaning your filter's cutoff frequency will physically shift as the signal voltage changes, introducing nasty harmonic distortion into audio or precision ADC paths.
Worked Numeric Example: Filtering a 25 kHz PWM Signal
Let's look at a common maker scenario: you are using a microcontroller's 25 kHz PWM output to generate a smooth 0-3.3V DC analog signal to drive an industrial 0-10V module (via a subsequent gain stage). A first-order RC filter would leave too much 25 kHz ripple, causing the industrial module to read erratic values. We need a passive second order LC filter.
The Design Parameters:
- PWM Frequency ($f_{pwm}$): 25,000 Hz
- Target Cutoff ($f_c$): 2,500 Hz (Rule of thumb: set $f_c$ to 1/10th of the PWM frequency for >90% ripple attenuation)
- Chosen Inductor (L): 10 mH (e.g., Bourns 78FR10K-RC radial choke)
The Calculation:
The resonant frequency formula for an LC circuit is $f_c = \frac{1}{2 \pi \sqrt{LC}}$. Rearranging to solve for C:
$C = \frac{1}{(2 \pi f_c)^2 \times L}$
$C = \frac{1}{(2 \pi \times 2500)^2 \times 0.01}$
$C = \frac{1}{246,740,110 \times 0.01} \approx 405 \text{ nF}$
The Real-World Selection:
405 nF is not a standard value. The closest E12 standard value is 390 nF. Let's use a WIMA MKS2 390 nF polyester film capacitor. If we plug 390 nF back into the formula, our actual cutoff shifts slightly to 2,549 Hz, which is perfectly acceptable for this application. To prevent the LC tank from ringing (peaking at the cutoff frequency due to high Q), ensure the load impedance of your downstream circuit is relatively low (e.g., < 10 kΩ), or add a small series damping resistor (10-47 Ω) in series with the inductor.
Where You Meet This In Practice (And What It Changes)
In a real circuit or installation, a second order low pass filter changes the signal integrity from "noisy and unusable" to "clean and precise." Here is where you will actively deploy them on the bench:
- Anti-Aliasing for ADCs: Before feeding a sensor signal into a 16-bit ADC like the ADS1115, a 2nd order active filter removes high-frequency RF interference that would otherwise alias into your digital readings, causing unexplainable jitter.
- Audio Crossovers: In DIY speaker builds, a 2nd order passive LC filter protects woofers from high-frequency distortion and ensures a smooth acoustic handoff to the tweeter at the crossover point.
- Switching Power Supply Ripple: Buck converters switching at 1 MHz often leave high-frequency spikes on the DC rail. A 2nd order LC pi-filter cleans this rail before it reaches sensitive analog sensors.
What People Commonly Confuse It With
The most frequent mistake hobbyists make is confusing the cutoff frequency with the point of total attenuation. At the -3dB cutoff frequency ($f_c$), the signal is only reduced by about 30% (0.707V/V). It does not disappear. The -40 dB/decade slope means the signal drops rapidly after this point. Another common error is assuming an active filter can drive a heavy load; op-amps like the LM358 max out around 20-30 mA. If your filter needs to drive a 50 Ω transmission line, you must add a dedicated buffer stage.
Quick Reference FAQ
Q: Can I just cascade two first-order RC filters to get a second-order response?
A: Not without a buffer. The second RC stage will load the first stage, altering the transfer function and resulting in a sluggish, poorly damped response that never truly achieves a -40 dB/decade slope. Use an op-amp unity-gain buffer between them, or use a proper Sallen-Key topology.
Q: Why choose Butterworth over Bessel or Chebyshev?
A: Butterworth provides a maximally flat passband, making it ideal for general audio and DC measurement. Choose Bessel if you are filtering digital data pulses and need to preserve the step-response (minimal phase shift/ringing), or Chebyshev if you need an even steeper roll-off and can tolerate passband ripple.






