2nd Order Low Pass Filter: Design, Math, and Sallen-Key Example

A 2nd order low pass filter is an electronic circuit that attenuates frequencies above a specific cutoff point at a steepness of -40 dB per decade using two reactive energy-storage components. Unlike a basic first-order RC filter that lazily rolls off high-frequency noise at -20 dB/decade, a second-order topology aggressively crushes unwanted AC signals, making it the absolute minimum requirement for clean audio crossovers, precision ADC anti-aliasing, and PWM-to-DC smoothing in modern embedded systems.

What this filter changes in a real circuit is the width of the transition band. If you are trying to extract a clean DC voltage from a 5 kHz PWM signal, a 1st order filter with a 500 Hz cutoff will still allow significant 5 kHz ripple to pass through. A 2nd order filter cuts that same ripple by an additional 20 dB, transforming a noisy, unusable waveform into a stable analog reference.

The Roll-Off and What It Changes in Your Circuit

The defining characteristic of any second-order filter is its dual-pole response. Because it utilizes two reactive components (either two capacitors in an active RC design, or an inductor-capacitor pair in a passive LC design), the phase shift approaches -180° and the amplitude drops at -40 dB/decade (or -12 dB/octave) well past the cutoff frequency.

To put that in perspective: one decade past your cutoff frequency, a 1st order filter attenuates the signal by 90% (-20 dB). A 2nd order filter attenuates it by 99% (-40 dB). This steepness is what allows you to place the cutoff frequency closer to your desired signal bandwidth without choking the fundamental frequency, while still completely eliminating high-frequency switching noise or electromagnetic interference (EMI).

Worked Numeric Example: Designing a 500 Hz Butterworth Filter

Let’s design an active 2nd order low pass filter to smooth a 5 kHz PWM output from an ESP32 into a clean analog voltage for a 0-10V industrial motor controller input. We will use the Sallen-Key topology, which is the industry standard for active 2nd order filters due to its simplicity and low component count.

We want a maximally flat passband, known as a Butterworth response (Quality factor Q = 0.707). A common mistake is using equal resistors and equal capacitors, which yields a Q of 0.5 (a sluggish, overdamped response). For a true unity-gain Butterworth response, the capacitor ratio must be 2:1.

The Calculation

• Target Cutoff ($f_c$): 500 Hz
• Topology: Unity-gain Sallen-Key (Op-amp configured as a voltage follower)
• Capacitor Selection: Let $C_2 = 100$ nF (C0G/NP0). Therefore, $C_1 = 200$ nF (achieved by placing two 100 nF C0G caps in parallel).

The formula for the cutoff frequency in this configuration is:

$$f_c = \frac{1}{2 \pi R \sqrt{C_1 C_2}}$$

Solving for R (where $R_1 = R_2 = R$):

$$R = \frac{1}{2 \pi f_c \sqrt{200\text{nF} \times 100\text{nF}}} = \frac{1}{2 \pi \times 500 \times \sqrt{2} \times 100 \times 10^{-9}}$$

$$R \approx 2251 \Omega$$

We will select the closest standard 1% (E96 series) resistor value: 2.26 kΩ. If restricted to 5% (E24) components, use 2.2 kΩ.

Final Bill of Materials (BOM)

With the 2.26 kΩ resistors, the actual cutoff frequency lands at 498 Hz. At the 5 kHz PWM fundamental frequency (one decade above cutoff), the filter will provide roughly -40 dB of attenuation, reducing the PWM ripple to virtually unmeasurable levels on a standard oscilloscope.

Where You Meet This in Practice

You will rarely build a 2nd order low pass filter just for the sake of theory; it is a functional necessity in several specific domains:

• Audio Crossovers: In powered studio monitors, 2nd order (and 4th order Linkwitz-Riley) low pass filters route bass frequencies to the woofer while aggressively blocking mid-range and treble frequencies that would cause intermodulation distortion in the large driver.
• ADC Anti-Aliasing: Before feeding a sensor signal into a high-resolution delta-sigma ADC like the ADS1115, a 2nd order active filter removes high-frequency RF noise. If this noise exceeds the Nyquist frequency of the ADC's sampling rate, it folds back into the baseband as unremovable aliasing errors.
• PWM-to-Analog DACs: Microcontrollers like the ESP32 or STM32 lack true digital-to-analog converters on all pins. A 2nd order low pass filter converts the high-frequency PWM duty cycle into a stable, low-ripple DC voltage for driving analog control loops.

Common Confusions and Pitfalls

The most frequent error hobbyists make with 2nd order filters is confusing the -3 dB cutoff frequency with the resonance peak. In topologies with a Q-factor greater than 0.707 (like Chebyshev filters), the filter actually amplifies signals slightly just before the cutoff frequency. If you are filtering a sensor signal and accidentally design a filter with a Q of 1.5, a noise spike near your cutoff frequency will be amplified, potentially saturating your op-amp.

Another critical pitfall is ignoring the op-amp's Gain-Bandwidth Product (GBP). For a 2nd order filter to perform mathematically as designed, the op-amp's open-loop gain must be at least 40 dB to 60 dB higher than the signal level at the cutoff frequency. If you try to build a 100 kHz 2nd order filter using a legacy LM358 (which has a GBP of roughly 1 MHz), the op-amp will run out of gain, the phase margin will collapse, and your filter will likely oscillate or fail to attenuate high frequencies. For a 100 kHz cutoff, you need an op-amp with a GBP of at least 10 MHz to 20 MHz, such as the OPA2134 or AD823.