A bandreject filter (also known as a notch or band-stop filter) is a circuit that attenuates a specific, narrow range of frequencies while allowing frequencies both above and below that rejected band to pass through unchanged. When you have a clean signal corrupted by a single, stubborn interference frequency—like 60Hz mains hum in an audio preamp or a 2.4GHz Wi-Fi spike in a software-defined radio—a bandreject filter surgically removes the noise without dulling the high-frequency content of your actual signal. What it changes in a real circuit is the amplitude and phase of a highly specific slice of the spectrum, leaving the rest of the bandwidth intact. Hobbyists and students frequently confuse it with a bandpass filter (which does the exact opposite, keeping only the target band) or a low-pass filter (which indiscriminately chops off everything above a cutoff frequency).

The Core Concept: What a Bandreject Filter Actually Does

Think of a bandreject filter like a multi-lane highway where one specific exit ramp is blocked by concrete barriers. Traffic on the main highway (frequencies below the notch) and traffic on the overpass (frequencies above the notch) flow freely, but cars trying to take that exact exit (the noise frequency) are stopped dead. In the frequency domain, this looks like a deep, narrow 'V' or 'U' shape carved out of an otherwise flat Bode plot.

The Phase Shift Penalty: While a bandreject filter preserves amplitude outside the notch, it inherently introduces a phase shift around the rejected frequency. If the Q-factor (quality factor) of the filter is set too high to create an ultra-narrow notch, this rapid phase shift can cause 'ringing' or smearing in the time domain. This is why you must balance notch depth and width against your signal's transient requirements.

Unlike a simple RC low-pass filter that rolls off at -20dB per decade and destroys your high-frequency data, a well-designed bandreject filter drops 40dB to 60dB exactly at the interference frequency, then immediately recovers to unity gain (0dB) just a few hertz away.

The Math and the Bench: A 60Hz Notch Filter Worked Example

Let's design a passive Twin-T network to kill 60Hz hum, a classic bench exercise. The center frequency ($f_c$) of a Twin-T filter is determined by the formula:

$f_c = \frac{1}{2 \pi R C}$

We want $f_c = 60$ Hz. Let's pick a standard, easy-to-source film capacitor value of $C = 0.1 \mu F$.

Calculated R: $R = \frac{1}{2 \pi \times 60 \times 0.1 \times 10^{-6}} = 26,525 \Omega$

The closest standard 1% E96 resistor value is 26.7 kΩ. However, the Twin-T topology requires specific ratios: two resistors of $R$, one of $R/2$ (13.35 kΩ), two capacitors of $C$, and one of $2C$ (0.2 µF).

Here is where bench reality bites theory. If you build this with standard 5% carbon film resistors and 10% ceramic capacitors, your actual notch frequency might land at 54Hz or 67Hz. Worse, the mismatched component ratios will destroy the destructive interference that creates the notch, degrading your attenuation from a theoretical -60dB down to a useless -12dB. To achieve a deep null, you must use 1% or 0.1% tolerance metal film resistors and tightly matched polypropylene or polystyrene film capacitors. For the $2C$ capacitor, parallel two matched 0.1 µF film caps. For the $R/2$ resistor, use a 13.3 kΩ 1% resistor or series a 10k and 3.3k.

Where You Meet Bandreject Filters in Practice

You will encounter bandreject topologies across three primary domains in electrical engineering and DIY electronics:

  • Audio Engineering & Hi-Fi: Removing 50Hz (EU/UK) or 60Hz (US) ground loop hum from microphone preamps and phono stages without killing the bass response of the music.
  • Biomedical Sensors: ECG and EEG amplifiers measure millivolt-level biological signals. The human body acts as an antenna for mains wiring, so aggressive, high-Q active bandreject filters are mandatory at the front end to strip out 60Hz interference before the ADC clips.
  • RF Communications: In software-defined radio (SDR) or ham repeater sites, cavity notch filters are used to block a nearby high-power commercial FM broadcast tower from desensitizing the receiver front-end.

Decision Tree: Choosing the Right Bandreject Topology

Selecting the right filter architecture depends entirely on your frequency range, signal type, and tuning requirements. Use this decision path to pick your topology and components.

If your application is... And you need... Choose this Topology Concrete Part / Implementation Pick
Analog Audio / Instrumentation (10Hz - 20kHz) Deep null, tunable Q, low noise Active State-Variable or MFB Notch Texas Instruments OPA2134 (Low THD, FET input)
Analog Audio (Budget / Hobby) Simple 60Hz kill, low cost Active Twin-T with buffer TI TL072 (Classic JFET dual op-amp)
RF / Microwave (VHF to GHz) High power handling, extreme Q Coaxial Cavity Notch Filter Mini-Circuits BPN-50N+ (or tunable cavity)
Digital Audio / DSP (Post-ADC) Perfect tuning, zero component drift Biquad IIR Digital Notch Teensy 4.1 Audio Library (AudioFilterBiquad)
The Default Recommendation: For 90% of bench, DIY, and student analog sensor projects, build an active state-variable notch filter using the OPA2134 or TL072 op-amp. Unlike the passive Twin-T, the active state-variable topology allows you to tune the notch frequency and the Q-factor independently using trimmer potentiometers, completely eliminating the headache of matching 1% component ratios.

Common Mistakes and Troubleshooting

When your bandreject filter fails to null the target frequency, or introduces new distortion, check these common failure modes:

  1. Op-Amp Slew Rate Limiting: If you are filtering a high-frequency RF signal or a fast transient, a slow op-amp like the LM741 (Slew rate: 0.5 V/µs) will distort the signal outside the notch. Ensure your op-amp's Gain Bandwidth Product (GBW) is at least 100x higher than the highest frequency you intend to pass.
  2. Ground Loop Injection: If your active filter is powered by a noisy linear regulator or shares a ground return path with a digital microcontroller, the power rail noise will bypass the filter entirely. Use a dedicated star-ground point for the analog filter stage.
  3. Capacitor Dielectric Absorption: Using cheap X7R or Y5V ceramic capacitors in the feedback loop of an active notch filter introduces microphonic noise and dielectric absorption, which smears the notch. Always use C0G/NP0 ceramics or polypropylene film caps for filter timing components.

For a deeper dive into active filter design equations and Sallen-Key variations, the Electronics Tutorials band-stop guide and All About Circuits provide excellent foundational math. For op-amp selection, always verify the noise density and GBW in the manufacturer datasheet before finalizing your BOM.

Frequently Asked Questions

Q: Can I just use a low-pass filter with a 100Hz cutoff to kill 60Hz hum?
A: No. A standard 1st-order low-pass filter at 100Hz will only attenuate 60Hz by a few decibels. To get -40dB of attenuation at 60Hz using a low-pass filter, you would need to set the cutoff frequency around 6Hz, which would completely destroy the bass response of an audio signal or the transient response of a sensor. A bandreject filter targets only the 60Hz spike.

Q: What is the Q-factor in a notch filter, and how do I calculate it?
A: The Q-factor (Quality factor) defines the width of the notch relative to its center frequency ($Q = f_c / \text{Bandwidth}$). A high Q (e.g., Q=20) creates a very narrow, surgical notch that removes exactly 60Hz and leaves 55Hz and 65Hz untouched. A low Q (e.g., Q=1) creates a wide, shallow dip that removes a broad swath of the midrange. In biomedical ECG designs, a high Q is preferred to preserve the low-frequency ST-segment data.

Q: Why does my passive Twin-T filter only attenuate by 15dB instead of the theoretical 60dB?
A: Component mismatch. The Twin-T relies on perfect destructive interference. If your R and 2R resistors, or C and 2C capacitors, are off by even 2%, the two signal paths will not perfectly cancel each other out at the summing junction. Switch to 0.1% tolerance components or move to an active state-variable topology to fix this.