A band reject filter is an electronic circuit that allows all signal frequencies to pass through unattenuated except for a specific, targeted range of frequencies that it heavily attenuates or blocks entirely. In a real circuit or installation, it changes the signal profile by surgically removing a narrow band of noise—like 60 Hz mains hum or a specific RF carrier—without degrading the bass or treble of the surrounding audio or data. Beginners most commonly confuse it with a band-pass filter (which does the exact opposite, blocking everything except the target band) or assume it functions like a simple low-pass filter that just rolls off high frequencies above a single threshold.
Core Topologies and Performance Data
Not all band reject filters are created equal. The topology you choose dictates the steepness of the rejection (the Q-factor), the component count, and how easily you can tune it on the bench. While a simple passive LC trap might suffice for rough power supply filtering, precision analog work requires active topologies. Below is a spec-sheet-table comparing the most common architectures used in modern analog design.
| Topology | Typical Q-Factor Range | Component Count (per stage) | Tuning Difficulty | Best Application |
|---|---|---|---|---|
| Passive Twin-T | 0.25 (Fixed) | 6 (3R, 3C) | High (requires matched parts) | Basic audio hum reduction |
| Active Bootstrapped Twin-T | 1.0 to 15.0 | 8 (6 passive + 1 op-amp + 1 feedback R) | Medium (tune via single feedback R) | Precision 50/60 Hz notch filtering |
| Fliege Biquad | 5.0 to 50.0 | 10 (4R, 2C, 2 op-amps) | Low (independent f0 and Q tuning) | Parametric EQ, narrow RF traps |
| State-Variable (Notch Output) | 0.5 to 100+ | 14+ (Multiple R, C, 3+ op-amps) | Low (highly stable, independent controls) | Biomedical ECG/EEG baseline cleanup |
| Passive LC Trap (Parallel) | 10 to 80 (depends on inductor) | 2 (1L, 1C) | High (requires variable inductor/cap) | Power systems, VFD harmonic filtering |
As noted in the Analog Devices Linear Circuit Design Handbook, passive Twin-T networks suffer from a notoriously low intrinsic Q-factor of roughly 0.25, resulting in a wide, shallow 'V' shape rather than a sharp notch. To achieve a deep, narrow null, designers almost universally add an op-amp buffer to bootstrap the ground node, feeding a portion of the output signal back into the Twin-T junction to artificially increase the Q-factor.
The Math in Action: Designing a 60 Hz Notch Filter
Let’s design an active bootstrapped Twin-T network to kill 60 Hz power line hum in a sensitive audio preamplifier. The foundational formula for the center (notch) frequency of a symmetrical Twin-T network is:
fnotch = 1 / (2 × π × R × C)
We need to select standard, readily available component values. Let’s start by fixing the capacitor value. For audio-frequency filtering, we want to avoid electrolytic capacitors due to their high tolerance and leakage. Instead, we will use 0.1 μF (100 nF) polypropylene film capacitors, which offer excellent stability and low dielectric absorption.
Rearranging the formula to solve for R:
R = 1 / (2 × π × 60 Hz × 0.1 × 10-6 F)
R = 1 / (0.000037699)
R ≈ 26,525 Ω
• C1, C2 = 0.1 μF (Film)
• C3 = 0.2 μF (Two 0.1 μF in parallel)
• R1, R2 = 26.7 kΩ (1% tolerance metal film)
• R3 = 13.3 kΩ (1% tolerance metal film)
The depth of the notch (how close to zero the signal gets at 60 Hz) is entirely dependent on the ratio matching of your resistors and capacitors. If you use standard 5% tolerance components, the best null you can hope for is roughly -15 dB. By using 1% metal film resistors and measuring your 0.1 μF capacitors with a bench LCR meter to select pairs matched within 0.5%, you can easily achieve a null depth of -40 dB to -50 dB.
To sharpen this notch, we route the junction of the Twin-T (which would normally go to ground) into the non-inverting input of a unity-gain op-amp buffer (like a low-noise OPA2134). The output of the op-amp is fed back to the junction via a voltage divider. By adjusting the ratio of this feedback divider, you control the Q-factor. For a 60 Hz hum filter, a Q of around 5 to 8 is ideal—sharp enough to kill the fundamental and its immediate sidebands, but not so sharp that it introduces phase-shift ringing on transient audio signals.
Where You Meet This in Practice
Band reject filters are ubiquitous across multiple engineering disciplines, though they often hide under application-specific names. According to fundamental filter theory outlined by All About Circuits, the application dictates the topology:
- Audio Engineering & Studio Gear: Here, they are almost exclusively called 'notch filters.' Parametric equalizers use state-variable band reject topologies to allow sound engineers to sweep across the frequency spectrum, find a resonant room mode or feedback frequency, and notch it out without affecting the rest of the mix.
- Biomedical Instrumentation (ECG/EEG):strong> When measuring microvolt-level bio-potentials from the human body, the body acts as an antenna for 50/60 Hz mains fields. Active Twin-T or Fliege biquad notch filters are hardwired into the front-end of ECG machines to strip this interference before the analog-to-digital converter (ADC) clips the signal.
- Power Electronics & VFDs: Variable Frequency Drives generate massive harmonic currents. In industrial installations, passive LC band reject traps (tuned to the 5th, 7th, or 11th harmonics) are bolted onto busbars to shunt specific harmonic frequencies away from sensitive control transformers.
- Radio Receivers: In software-defined radio (SDR) and superheterodyne front-ends, passive ceramic or SAW (Surface Acoustic Wave) band reject filters are used to block overwhelmingly strong local FM broadcast stations that would otherwise desensitize the receiver's low-noise amplifier (LNA).
Troubleshooting and Common Confusions
When a band reject filter fails to perform on the bench, the issue is rarely the math; it is almost always the physical implementation. Let's clear up the most common confusions and diagnostic paths.
Notch vs. Wide Band-Stop
People frequently use 'notch filter' and 'band reject filter' interchangeably, but in strict analog design, a notch filter is a band reject filter with a very high Q-factor (narrow bandwidth). If you cascade a low-pass filter (cutoff at 400 Hz) and a high-pass filter (cutoff at 4 kHz), you create a wide band reject filter that blocks everything in between. This is useful for removing a wide band of midrange noise, but it will utterly destroy the fundamental frequencies of a human voice. A true notch filter targets a razor-thin slice of the spectrum.
Frequently Asked Troubleshooting Questions
Q: My Twin-T notch filter only attenuates the 60 Hz signal by 6 dB. Why isn't it nulling completely?
A: Your component ratios are mismatched. The Twin-T relies on destructive interference; the signal through the R-C-R path must be exactly equal in amplitude and exactly 180° out of phase with the signal through the C-R-C path at the target frequency. Swap your 5% resistors for 1% or 0.1% tolerance parts, and verify your capacitor values with an LCR meter. Also, ensure you are using C0G/NP0 ceramic or film capacitors; X7R ceramics exhibit severe voltage coefficients and dielectric absorption that will smear the phase response and ruin the null.
Q: I added an op-amp to boost the Q-factor, but now the circuit oscillates at 60 Hz instead of rejecting it.
A: You have pushed the positive feedback too high. In a bootstrapped Twin-T, the feedback ratio determines the Q. If the feedback gain reaches exactly 1.0 at the notch frequency, the circuit turns into a Wien-bridge oscillator. Back off the feedback potentiometer slightly. Aim for a Q of 10 maximum for a stable, non-ringing response.
Q: Can I just use a digital filter (DSP) instead of building this analog circuit?
A: Yes, but only if the interference isn't clipping your ADC. If a massive 60 Hz hum is causing your analog front-end to rail out before the signal even reaches the microcontroller, a digital IIR notch filter won't save you. You must use an analog band reject filter before the ADC to preserve the dynamic range of the desired signal. For deeper insights into mixed-signal filtering, the Electronics Tutorials guide on Band Stop Filters provides excellent crossover schematics.
Mastering the band reject filter requires moving beyond the idealized textbook equations and respecting the physical limitations of real-world components. By selecting low-tolerance, low-dielectric-absorption parts and carefully managing the Q-factor feedback loop, you can build notch circuits that cleanly excise interference while leaving your primary signal pristine.
