A Sallen-Key filter is an active second-order electronic filter topology that uses a single operational amplifier, two resistors, and two capacitors to shape frequency response without the insertion loss of passive designs. In a real circuit, it changes the signal path by providing a sharp -40dB/decade rolloff while buffering the output, allowing you to cascade multiple stages without impedance loading issues dragging down your signal amplitude.
The Core Topology and Common Confusions
Before sizing components, we need to clear up a massive point of confusion that trips up hobbyists and junior engineers alike: the difference between a topology and a response curve.
People routinely confuse the Sallen-Key topology with Butterworth, Chebyshev, or Bessel responses. Sallen-Key is the physical circuit layout—the specific arrangement of the op-amp, resistors, and capacitors. Butterworth, Chebyshev, and Bessel are mathematical tuning targets that dictate the values of those components. You can build a Butterworth filter using a Sallen-Key topology, but you can also build a Chebyshev filter using the exact same Sallen-Key schematic just by changing the resistor and capacitor ratios.
The second common confusion is mixing up Sallen-Key with the Multiple Feedback (MFB) topology. Both are active second-order filters, but Sallen-Key is a non-inverting configuration where the op-amp acts primarily as a buffer (or non-inverting amplifier). MFB is an inverting configuration that feeds the signal back into the inverting input. Sallen-Key is generally preferred for high-impedance sources and unity-gain applications, while MFB is favored when you need high-Q (narrow bandpass) filtering or when the op-amp's open-loop gain limitations would otherwise degrade a Sallen-Key design at high frequencies.
Worked Numeric Example: 1kHz Unity-Gain Low-Pass
Let's design a second-order low-pass Sallen-Key filter with a cutoff frequency ($f_c$) of 1kHz and a Butterworth response ($Q = 0.707$). We will use a unity-gain configuration (gain = 1) to minimize component count and noise.
For a unity-gain Sallen-Key low-pass filter, the standard design equations simplify if we set $R_1 = R_2 = R$. The cutoff frequency and Q factor are defined as:
- $f_c = \frac{1}{2\pi R \sqrt{C_1 C_2}}$
- $Q = \frac{\sqrt{C_1 / C_2}}{2}$
To achieve a Butterworth response ($Q = 0.707$, which is $1/\sqrt{2}$), the ratio of $C_1$ to $C_2$ must be exactly 2:1. Let's select C2 = 10nF and C1 = 20nF. Plugging these into the cutoff equation to solve for R:
$1000 = \frac{1}{2\pi R \sqrt{20 \times 10^{-9} \times 10 \times 10^{-9}}}$
$R \approx 11,253 \Omega$
| Component | Designator | Value / Part Number | Critical Specification |
|---|---|---|---|
| Resistor | R1, R2 | 11kΩ | 1% tolerance, metal film (low noise) |
| Capacitor | C1 | 22nF | C0G/NP0 dielectric (crucial for low THD) |
| Capacitor | C2 | 10nF | C0G/NP0 dielectric |
| Op-Amp | U1 | NE5532 or TL072 | Low noise, adequate slew rate for audio |
A vital design note for audio and precision DC applications: you must use C0G/NP0 ceramic capacitors or polypropylene film capacitors for C1 and C2. If you use X7R or Y5V dielectrics, the piezoelectric effect and severe voltage coefficient of those ceramics will inject massive Total Harmonic Distortion (THD) into your signal as the voltage across the capacitor changes. According to the Analog Devices MT-224 tutorial on active filters, capacitor dielectric absorption and non-linearity are the primary culprits in degrading active filter performance.
Where You Meet This in Practice
You will encounter the Sallen-Key topology repeatedly across three main domains of electronics design:
Active Audio Crossovers
In bi-amped or tri-amped speaker systems, you need to split the audio spectrum before it hits the power amplifiers. A cascaded pair of Sallen-Key low-pass and high-pass filters provides the -40dB/decade (24dB/octave) Linkwitz-Riley or Butterworth slopes required to protect tweeters from low-frequency excursion while keeping the phase response manageable at the crossover point.
DAC Reconstruction (Anti-Imaging) Filters
When a Digital-to-Analog Converter outputs a signal, it leaves high-frequency 'staircase' imaging artifacts centered around the sampling frequency. A Sallen-Key low-pass filter placed immediately after the DAC smooths these steps. For a 44.1kHz audio DAC, a 20kHz Sallen-Key filter removes the ultrasonic noise without touching the audible spectrum.
Sensor Signal Conditioning and Anti-Aliasing
Before an analog sensor signal (like a thermocouple or strain gauge) enters a microcontroller's ADC, it must pass through an anti-aliasing filter to prevent high-frequency EMI from folding back into the baseband. A unity-gain Sallen-Key filter using a rail-to-rail op-amp like the TLV2372 provides a clean, low-impedance drive directly into the ADC's sample-and-hold capacitor.
Sallen-Key Filter FAQ
Why use a Sallen-Key filter instead of a passive RC filter?
A passive first-order RC filter only provides a -20dB/decade rolloff, which is often too gradual to adequately reject noise or separate frequency bands. Furthermore, passive filters suffer from insertion loss and loading effects; if you connect a low-impedance load to a passive RC filter, the cutoff frequency shifts dramatically. The Sallen-Key filter uses an op-amp to provide a -40dB/decade rolloff with zero insertion loss and a buffered output that isolates the filter's internal nodes from whatever load you connect downstream.
How do I calculate Sallen-Key filter component values?
While the manual algebra is straightforward for unity-gain designs, calculating component values for high-Q, high-gain, or Chebyshev responses involves complex quadratic equations. In professional practice, engineers use automated tools. The Texas Instruments Filter Designer or the Analog Devices Analog Filter Wizard allow you to input your desired cutoff frequency, passband ripple, and stopband attenuation, and they will output the exact resistor and capacitor values, complete with standard E24/E96 rounding and simulated Bode plots.
Can a Sallen-Key filter be used for high-pass applications?
Yes. To convert a Sallen-Key low-pass filter into a high-pass filter, you simply swap the positions of the resistors and capacitors in the frequency-determining network. The resistors go to ground, and the capacitors become the series signal path elements. The math for cutoff frequency remains identical, but the circuit will now attenuate low frequencies and pass high frequencies with the same -40dB/decade slope.
What is the difference between Sallen-Key and Multiple Feedback (MFB) topologies?
The Sallen-Key topology is non-inverting and relies on the op-amp's open-loop gain to maintain accuracy at higher frequencies, making it less ideal for very high-Q (narrow) bandpass filters where the required gain approaches the op-amp's physical limits. The Multiple Feedback (MFB) topology is inverting and uses the op-amp in an integrator configuration. MFB is generally preferred for high-Q applications, high-frequency filters, and designs where you need to minimize sensitivity to the op-amp's internal gain-bandwidth product, though it requires more careful management of DC offsets and inverting signal polarity.
