How to Design a Passive Band Stop Filter for Mains Hum Rejection

A passive band stop filter is an electronic circuit built exclusively from resistors, capacitors, and inductors that attenuates a specific band of frequencies while allowing all lower and higher frequencies to pass unimpeded. In a real circuit, it changes the signal profile by surgically removing a narrow interference spike—like 60 Hz mains hum from an audio line or a specific switching harmonic from a sensor feed—without requiring an external DC power supply or introducing op-amp thermal noise. Think of it like a physical sieve that lets fine sand and large rocks through, but traps medium-sized gravel. Beginners frequently confuse it with a band-pass filter (which blocks everything except the target band) or assume all notch filters are active (using op-amps to achieve a sharper Q factor).

While active filters dominate modern high-precision designs, passive topologies remain essential in high-voltage environments, RF front-ends, and ultra-low-noise audio paths where op-amp power supply rejection ratio (PSRR) limitations or active device noise floors are unacceptable. Below, we break down the most practical passive topology, provide exact component values, and highlight the impedance-matching traps that ruin notch depth on the bench.

The Twin-T RC Topology and Component Selection

For audio and sub-audio frequencies (10 Hz to 20 kHz), the Twin-T RC network is the gold standard for passive band stop filtering. It uses two T-shaped networks: one low-pass (two resistors, one capacitor) and one high-pass (two capacitors, one resistor), wired in parallel. At the target notch frequency ($f_c$), the phase shifts of the two paths perfectly cancel each other out, theoretically driving the output to zero.

The governing equation for the center notch frequency is:

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

Unlike active filters, a passive Twin-T requires exact component ratios to maintain the null. The shunt arm requires a resistor of exactly $R/2$, and the series arm requires a capacitor of exactly $2C$. Below is a data-dense reference table for the most common interference frequencies you will encounter in the lab or field.

Worked Numeric Example: Building a 60 Hz Hum Killer

Let us design a 60 Hz passive band stop filter for a high-impedance guitar preamp input that is picking up severe 60 Hz magnetic field interference from a nearby dimmer switch. We will use the 60 Hz row from our table as a starting point, but we must adapt it to real-world E24/E96 standard component values.

Step 1: Selecting Standard Values

The ideal math calls for $R = 26.52 k\Omega$. The closest standard 1% metal film resistor (E96 series) is 26.7 kΩ. Using 26.7 kΩ and our 100 nF C0G capacitors, the actual notch frequency shifts slightly:

$f_{actual} = \frac{1}{2 \pi \times 26700 \times 100 \times 10^{-9}} = 59.6 \text{ Hz}$

This 0.4 Hz shift is entirely acceptable for rejecting a 60 Hz fundamental, as the notch will still cover the 59.5 Hz to 60.5 Hz variance of the grid.

Step 2: Constructing the Ratio Arms

The Twin-T demands an $R/2$ shunt resistor and a $2C$ series capacitor.

• For R/2: We need 13.35 kΩ. The closest 1% value is 13.3 kΩ. Alternatively, wire two 26.7 kΩ 1% resistors in parallel to get an exact 13.35 kΩ match, which drastically improves the notch depth.
• For 2C: We need 200 nF. Simply wire two 100 nF C0G capacitors in parallel. Do not use a single 200 nF capacitor unless you can verify its exact tolerance, as parallelizing averages out minor manufacturing variances.

Step 3: The Impedance Loading Trap

This is where most hobbyists fail. A theoretical passive Twin-T provides infinite attenuation (a perfect null) at $f_c$. However, this assumes an ideal zero-ohm source impedance and infinite load impedance. If you drive this 26.7 kΩ filter from a 10 kΩ source and load it with a 10 kΩ amplifier input, the impedance loading will completely destroy the phase cancellation. Your -60 dB theoretical notch will collapse to a useless -3 dB dip.

The Bench Rule: The load impedance must be at least 100 times greater than the filter's R value, and the source impedance must be less than 1/100th of R. For our 26.7 kΩ filter, your source must be under 267 Ω, and your load must be above 2.67 MΩ. If your circuit cannot meet these boundaries, you must buffer the filter with unity-gain op-amps (transitioning to an active design) or use an inductor-based topology.

Where You Meet This in Practice

While active filters are common in mixed-signal ICs, passive band stop filters are explicitly chosen in several demanding real-world installations:

• High-Voltage Sensor Telemetry: When monitoring current shunts or voltage dividers on 480V AC motor drives, you cannot easily route high-voltage transients into the sensitive inputs of an active op-amp filter. A passive LC or RC band stop filter safely rejects the 50/60 Hz fundamental before the signal hits the isolation amplifier.
• RF Front-Ends and Mixers: In software-defined radio (SDR) or superheterodyne receivers, passive LC band stop (notch) filters are used to block extremely strong local broadcast stations (like a nearby 100 kW FM transmitter) that would otherwise cause intermodulation distortion in the first low-noise amplifier (LNA). Active filters cannot handle the RF power levels present at the antenna terminal.
• Power Line Communication (PLC): In smart grid and home-automation PLC modems, passive LC band stop filters are wired across the mains input to prevent the high-frequency data carriers (typically 20 kHz to 500 kHz) from being shorted out by the low impedance of the mains transformer or local power supplies.

For authoritative deep-dives into filter topology selection, the Electronics Tutorials guide on Band Stop Filters provides excellent AC theory breakdowns, while All About Circuits offers practical Bode plot visualizations for these networks.

Passive vs. Active Band Stop Filters

Deciding between a passive and active topology dictates your bill of materials (BOM), power budget, and noise floor. Below is a direct comparison to guide your design choice.

Choose Passive When: You are operating at RF frequencies, dealing with signals that exceed op-amp voltage rails, requiring zero-power operation for remote sensor nodes, or designing an ultra-low-noise audio path where op-amp current noise is unacceptable. For more on the noise trade-offs, Analog Devices' classic Ask The Applications Engineer column details how passive components avoid active device noise injection.

Choose Active When: You need a sharp, high-Q notch (like removing exactly 60.0 Hz without touching 55 Hz or 65 Hz), your source and load impedances are fixed and unfavorable, or you need to recover the insertion loss with voltage gain.