What Is a Notch Filter? Circuit Theory, Design, and Component Picks

A notch filter is a specialized band-stop circuit that severely attenuates a very narrow, specific range of frequencies while allowing all signals below and above that rejected band to pass through unchanged. In a real circuit or installation, a notch filter changes the signal-to-noise ratio by surgically removing a single, persistent interference source—like 60 Hz mains hum or a specific RF carrier—without dulling high-frequency transients or muddying low-end response the way a low-pass filter would.

The Core Mechanics: Q-Factor and Bandwidth

To understand what is notch filter behavior, you have to look at the Quality factor (Q). The Q-factor defines how narrow the rejection band is. It is calculated as the center frequency ($f_c$) divided by the bandwidth ($BW$) at the -3 dB points:

Q = f_c / BW

A standard broad band-stop filter might have a Q of 1 or 2, rejecting a massive swath of the spectrum. A true notch filter operates with a high Q—typically Q > 10, and often Q > 50 in precision audio or RF applications. Think of a notch filter as a surgical scalpel compared to the broadsword of a low-pass filter; it removes exactly the tissue you want gone and leaves the rest alone.

The depth of the notch (how many decibels of attenuation you achieve at $f_c$) is entirely dependent on component matching. In a passive analog design, a 1% mismatch in your resistors or capacitors will limit your maximum attenuation to roughly -40 dB. To achieve -60 dB or deeper, you need 0.1% precision components or an active topology with a tuning trimmer.

Worked Numeric Example: Designing a 60 Hz Mains Hum Killer

Let's design an active Twin-T notch filter to eliminate 60 Hz AC mains hum from an audio preamplifier. The Twin-T topology is the industry standard for analog notch filtering because it uses a single op-amp and provides a deep null when properly buffered.

The formula for the notch frequency is:

f_notch = 1 / (2 * π * R * C)

Step 1: Choose your capacitor.
For audio, you want to avoid electrolytic capacitors due to leakage and tolerance issues. We will select a 100 nF (0.1 µF) C0G/NP0 ceramic or polypropylene film capacitor. Never use X7R ceramics for precision filters; they exhibit voltage coefficients and microphonics that will modulate your signal and ruin the Q-factor.

Step 2: Calculate R.
Rearranging the formula: R = 1 / (2 * π * 60 * 100e-9)
R = 26,525 Ω.

Step 3: Map to the Twin-T Topology.
The Twin-T network requires specific ratios:

• Branch 1 (Series R, Shunt C): Two resistors of value R (26.5 kΩ) in series, with a capacitor of value 2C (200 nF) tied from their midpoint to ground.
• Branch 2 (Series C, Shunt R): Two capacitors of value C (100 nF) in series, with a resistor of value R/2 (13.25 kΩ) tied from their midpoint to ground.

Step 4: Component Selection and Tuning.
Standard 1% resistor values are 26.7 kΩ and 13.3 kΩ, which would shift the notch to 59.6 Hz. To hit exactly 60.0 Hz, use a 25 kΩ multi-turn cermet trimmer in series with a fixed 2.2 kΩ resistor for the 'R' positions, and a 12 kΩ fixed plus a 5 kΩ trimmer for the 'R/2' position. Buffer the output with a low-noise, high-slew-rate op-amp like the Texas Instruments OPA1612, and feed a fraction of the output back to the 'R/2' ground node to boost the Q-factor from a passive ~0.3 up to an active Q of 20 or higher.

Where You Meet Notch Filters in Practice

You will encounter notch filters across several distinct engineering domains, each with different implementation requirements:

• Audio Engineering & Live Sound: Parametric EQs use adjustable notch filters to hunt down and eliminate room resonances or microphone feedback frequencies (often between 1 kHz and 4 kHz) without ruining the overall tonal balance of the mix.
• Biomedical Sensors (ECG/EEG): The human body acts as an antenna for 50/60 Hz mains fields. Electrocardiogram front-ends (like the TI ADS1298) use digital or analog notch filters to strip this out, because the ST-segment of an ECG contains critical diagnostic data right around 10-40 Hz that a low-pass filter would destroy.
• Variable Frequency Drives (VFDs): In industrial power systems, passive LC notch filters (often called 'trap filters') are installed on the output of VFDs to shunt specific high-frequency PWM switching harmonics (like the 3rd or 5th harmonic) to ground, preventing them from overheating motor windings.
• Software-Defined Radio (SDR): Digital notch filters (IIR biquads) are applied in the DSP chain to notch out overwhelmingly strong local FM broadcast carriers that would otherwise desensitize the receiver's front-end LNA and cause intermodulation distortion.

Common Confusions: Notch vs. Low-Pass vs. Broad Band-Stop

When reading datasheets or configuring DSP software, engineers frequently mix up these three filter types. Here is how to keep them straight:

Decision Tree: Choosing Your Notch Filter Implementation

Do not default to a passive RC network; they lack the Q-factor for real-world interference rejection. Use this decision matrix to select the exact architecture and component for your project.

The Default Recommendation: If you are building a bench project, an Arduino/ESP32 sensor frontend, or an audio interface and you just need to kill 60 Hz mains hum, build the Active Twin-T using the OPA1612. It costs under $5 in components, requires no complex software, and provides a deep, clean null that outperforms most basic DSP implementations at a fraction of the development time.

FAQ: Practical Bench Questions

For deeper mathematical modeling of active filter topologies, refer to the Analog Devices MT-222 Tutorial on Notch Filters and the Texas Instruments OPA1612 Datasheet for specific active component performance limits. For foundational theory, Electronics Tutorials provides excellent interactive bandwidth calculators.