Bandstop Filter Design: Topologies, 60Hz Notch Example, and Uses

A bandstop filter (often called a notch or band-reject filter) is an electronic circuit that heavily attenuates a specific, narrow range of frequencies while allowing all frequencies below and above that rejected band to pass through unchanged. In a real circuit or installation, it surgically removes a single interfering tone—like 60 Hz mains hum or a specific RF carrier—without rolling off the high-frequency detail or low-frequency bass of your primary signal. Beginners frequently confuse it with a bandpass filter (which does the exact opposite, passing only the target band) or assume a standard low-pass filter can do the same job. A low-pass filter cannot isolate and remove a mid-band noise spike without also destroying all high-frequency data above your noise floor.

Filter Topologies and Performance Specifications

Before you start soldering, you need to select the right topology for your bandwidth and Q-factor (quality factor) requirements. The Q-factor determines how 'narrow' the rejected band is. A low-Q bandstop filter might attenuate a 2 kHz-wide swath of audio, while a high-Q notch filter targets a single 1 Hz slice. Below is a data-dense comparison of the most common topologies you will encounter on the bench.

According to Analog Devices Tutorial MT-223, achieving a deep null (attenuation > 40 dB) in analog RC networks requires resistor and capacitor matching to at least 0.1%. If you use standard 5% tolerance components in a passive Twin-T, your 'notch' will likely only bottom out at -15 dB, leaving audible hum in sensitive audio chains.

Worked Numeric Example: Designing a 60 Hz Active Notch Filter

Let's design an active Twin-T notch filter to eliminate 60 Hz powerline interference from a high-gain audio preamplifier. We will use the standard Twin-T topology but wrap it in an op-amp feedback loop to artificially boost the Q-factor and deepen the null.

The center frequency (f_c) of a Twin-T network is determined by the formula:

f_c = 1 / (2 π R C)

Step 1: Choose a standard resistor value.
Let's select R = 10 kΩ. This is a practical value that keeps impedance low enough to minimize thermal noise but high enough to avoid loading the op-amp.

Step 2: Calculate the required capacitance.
Rearranging the formula for C:
C = 1 / (2 π × R × f_c)
C = 1 / (2 π × 10,000 Ω × 60 Hz)
C ≈ 265.25 nF

Step 3: Map to real-world E12/E24 component values.
You cannot buy a 265.25 nF capacitor off the shelf. The closest standard E12 value is 270 nF. Let's see how this shifts our center frequency:
f_c(actual) = 1 / (2 π × 10,000 Ω × 270 × 10^-9 F) = 58.9 Hz.

Step 4: The Q-Boost Feedback Loop.
A bare passive Twin-T has a Q of roughly 0.25, meaning the attenuation slope is very gentle. To make it a true 'notch' filter, we tap the output of the filter and feed a portion of it back to the common ground node of the Twin-T via a buffer op-amp (like the low-noise TI OPA1612). By adjusting a 10 kΩ potentiometer in this feedback path, you control the Q. Turn the pot to increase feedback, and the null depth drops from -15 dB down to -50 dB or lower, effectively erasing the 60 Hz tone.

Where You Meet Bandstop Filters in Practice

You will rarely see a bandstop filter used for general tone shaping; they are almost exclusively deployed as problem-solvers for specific interference.

1. Biomedical Instrumentation (ECG/EEG)

Electrocardiogram (ECG) signals are incredibly weak, typically between 1 mV and 2 mV peak-to-peak. The human body acts as an antenna, picking up massive amounts of 50 Hz or 60 Hz electromagnetic interference from room wiring. If this noise isn't removed, it completely obscures the P-wave and QRS complex. Biomedical front-ends use high-Q active biquad bandstop filters (often integrated into chips like the Texas Instruments ADS1298) to notch out the powerline frequency without distorting the critical 1-40 Hz biological data.

2. RF Front-Ends and Software Defined Radios (SDR)

If you live near a high-power commercial FM broadcast tower, the sheer RF energy can overload the front-end low-noise amplifier (LNA) of your SDR, causing intermodulation distortion that ghosts across the entire spectrum. RF engineers use LC bandstop filters—often implemented as physical coaxial cavity resonators—to block that specific 98.5 MHz carrier. At VHF/UHF frequencies, the skin effect dictates that the physical plating and geometry of the cavity determine the Q-factor, allowing for incredibly deep rejection of the local tower while passing adjacent frequencies.

3. Digital Audio and DSP (ESP32 / STM32)

In modern embedded systems, analog notch filters are being replaced by Digital Signal Processing. If you are routing I2S audio through an ESP32, you can implement an Infinite Impulse Response (IIR) biquad bandstop filter in software. By calculating the exact digital coefficients for a 60 Hz notch and loading them into the ESP32's DSP pipeline, you achieve a mathematically perfect null with zero component drift, zero thermal noise, and infinite tunability via a simple variable update in your C++ code.

Bandstop vs. Bandpass: Clearing Up the Confusion

Because the names are similar and the schematics look like mirror images to the untrained eye, these two filters are constantly mixed up in parts ordering and circuit debugging. Here is the definitive breakdown.

For a deeper dive into the underlying AC theory and impedance math that governs these networks, the All About Circuits textbook chapter on band-stop filters provides excellent foundational schematics.

Frequently Asked Questions