Designing an Active Circuit of Band Pass Filter: MFB Topology Guide

When you need to isolate a specific frequency band—like pulling a 1 kHz pilot tone out of a noisy audio feed or stripping DC and high-frequency switching noise from a sensor signal—a passive RC network rarely cuts it. The circuit of band pass filter that dominates professional audio and instrumentation is the Multiple Feedback (MFB) active topology. Unlike passive filters that suffer from insertion loss and loading effects, the MFB configuration uses an operational amplifier to provide gain, sharp roll-off, and a low-impedance output.

Below is a complete bench-to-breadboard guide to the MFB band pass filter, including node-level topology, real-world E24 component math, and the exact failure modes you will encounter when a solder joint goes bad.

The Multiple Feedback (MFB) Band Pass Topology

The MFB band pass filter relies on a single op-amp, two capacitors, and three resistors. It is an inverting topology, meaning the output signal is 180° out of phase with the input at the center frequency. Here is the node-by-node breakdown using a standard DIP-8 op-amp (like the TL072 or NE5532):

• Vin (Input): Feeds through resistor R1 into the summing junction.
• Node A (Summing Junction): Connects to the op-amp's inverting input (Pin 2). This is the critical mixing node where R1, R2, R3, C1, and C2 all meet.
• Node B (Vout): The op-amp output (Pin 6). R2 and C1 route feedback from here back to Node A.
• Node C (Non-Inverting Input): Pin 3. In a dual-supply system (±15V), this ties directly to system ground. In a single-supply system, it ties to a VCC/2 bias voltage.

To design effectively, you must understand how each component pulls the filter's center frequency ($f_c$), Quality factor ($Q$), and Gain ($A_v$). The table below maps component sensitivity and failure modes—keep this on your bench.

Table 1: MFB Component Sensitivity & Behavior Matrix

Component	Primary Function	Effect of +10% Value Shift	Open-Circuit Failure	Short-Circuit Failure
R1 (Input)	Sets input impedance & Gain	Gain drops ~10%; $f_c$ shifts slightly lower	No signal passes; output sits at 0V (or bias)	Gain spikes massively; op-amp rails out or oscillates
R2 (Feedback)	Sets Gain, Q, and $f_c$	Gain and Q increase; $f_c$ drops	DC feedback lost; output saturates to positive rail	Gain drops to zero; acts as a low-pass only
R3 (Ground)	Sets Q and $f_c$	Q drops; $f_c$ shifts higher	DC bias path lost; output saturates to rail	Signal shorts to ground; $f_c$ approaches infinity (no filter)
C1 (Feedback)	High-freq feedback path	$f_c$ drops; Q increases slightly	High-freq roll-off lost; circuit becomes a low-pass	DC feedback shorted; op-amp output pins to rail
C2 (Ground)	Low-freq shunt path	$f_c$ drops; Q increases slightly	Low-freq roll-off lost; circuit becomes a high-pass	Node A shorted to ground; no signal reaches op-amp

Why MFB Over Sallen-Key or Passive RLC?

If you are wondering why we choose the MFB circuit of band pass filter over the popular Sallen-Key or a passive inductor-capacitor (RLC) tank, it comes down to Q-factor stability and component sensitivity.

The Bench Reality: Sallen-Key filters are great for low-pass and high-pass designs, but when you push a Sallen-Key band pass above a Q of 3, the required component spread becomes absurd, and the filter becomes hyper-sensitive to the op-amp's Gain-Bandwidth Product (GBW). The MFB topology comfortably handles Q values up to 15 using standard, cheap op-amps.

Furthermore, passive RLC filters require physical inductors. At 1 kHz, a 10mH inductor is a bulky, expensive coil that picks up 60Hz mains hum like an antenna. The MFB active filter synthesizes the inductor's behavior using capacitors and resistors, keeping the footprint tiny and immune to magnetic interference.

Design Walkthrough: 1 kHz Audio BPF with Real E24 Values

Let's design a filter for a 1 kHz center frequency, a Q of 5 (yielding a bandwidth of 200 Hz), and a voltage gain of 10 (20 dB). We will use a TL072 op-amp running on ±15V rails.

Step 1: Choose the Capacitors
Select a standard capacitor value to anchor the math. Let's use C1 = C2 = 10 nF.
Critical Spec: Use C0G/NP0 dielectric ceramic capacitors. Do not use X7R or Y5V. X7R capacitors exhibit microphonics (they act as piezoelectric microphones) and their capacitance drops by up to 30% under DC bias, which will detune your filter on the bench.

Step 2: Calculate R2 (Feedback Resistor)
Formula: $R_2 = \frac{Q}{\pi \cdot f_c \cdot C}$
$R_2 = \frac{5}{\pi \cdot 1000 \cdot 10 \times 10^{-9}} = 159,154 \Omega$
Real E24 Value: 160 kΩ

Step 3: Calculate R1 (Input Resistor)
Formula: $R_1 = \frac{R_2}{2 \cdot A_v}$
$R_1 = \frac{160,000}{2 \cdot 10} = 8,000 \Omega$
Real E24 Value: 8.2 kΩ (Yields an actual gain of ~9.75)

Step 4: Calculate R3 (Ground Resistor)
To find R3, we use the center frequency equation rearranged for the parallel equivalent of the input network:
$R_{eq} = \frac{R_1 \cdot R_2}{R_1 + R_2} = \frac{8.2k \cdot 160k}{168.2k} = 7,800 \Omega$
$R_3 = \frac{1}{(2 \pi \cdot f_c \cdot C)^2 \cdot R_{eq}}$
$R_3 = \frac{1}{(2 \pi \cdot 1000 \cdot 10nF)^2 \cdot 7800} = 32,461 \Omega$
Real E24 Value: 33 kΩ

Failure Modes: What Breaks at the Extremes?

When you power up a newly breadboarded MFB filter and the oscilloscope shows a flatline or a railed-out DC voltage, you need a diagnostic decision tree. Here is what breaks when components fail or solder joints go cold.

1. Output Saturated to +14V or -14V (Rail-Out)

Cause: Loss of DC feedback path. In the MFB topology, R2 provides the critical DC negative feedback. If R2 is open, or if you accidentally swapped R2 with a capacitor, the op-amp runs open-loop at DC. The input bias current (pin 2) will charge the stray capacitance until the output slams into the positive or negative supply rail.
Fix: Power down. Measure resistance from Pin 6 (Vout) to Pin 2 (Inverting Input) with your multimeter. It should read exactly your R2 value (160 kΩ). If it reads OL (open), reseat R2.

2. Output is a Clipped, Distorted Mess

Cause: Input signal is too hot, or R1 is shorted. Remember that the MFB is an inverting amplifier with a gain of ~10. If your function generator is pushing 2V peak-to-peak, the filter will try to output 20V p-p. On ±15V rails, the TL072 will clip around ±13V.
Fix: Drop the function generator output to 500 mV p-p. If it's still clipping, check R1 for solder bridges to ground.

3. Severe 60Hz / 50Hz Hum on the Output

Cause: Missing decoupling capacitors or floating non-inverting input. If Pin 3 (Node C) is left floating instead of being tied to a solid ground plane, it acts as an antenna for mains hum.
Fix: Ensure Pin 3 is tied directly to the ground bus. Add 100 nF ceramic decoupling capacitors physically adjacent to Pin 7 (V+) and Pin 4 (V-) of the op-amp, routing them to ground with the shortest possible leads.

Step-by-Step Breadboard Testing Protocol

Do not just wire it up and blindly turn on the power. Follow this sequence to verify the circuit of band pass filter without risking your op-amp or your test gear.

1. Power and Decouple: Wire ±15V to Pins 7 and 4. Place 100 nF decoupling caps from each power pin to ground. Do not apply the audio signal yet.
2. DC Bias Verification: Power the board. Set your multimeter to DC Volts. Probe Pin 6 (Vout). It should read between -5mV and +5mV (the op-amp's input offset voltage). If it reads ±14V, you have an open feedback loop (check R2 and R3).
3. Inject the Sweep: Connect your function generator to Vin. Set it to a logarithmic sine sweep from 10 Hz to 10 kHz, with an amplitude of 500 mV p-p.
4. Scope the Output: Connect your oscilloscope probe to Vout. Set the scope to AC coupling, 2V/div, and trigger on the input signal.
5. Find the Peak: Pause the sweep when the output amplitude is at its maximum. Measure the exact frequency (should be ~992 Hz) and the peak-to-peak voltage (should be ~4.8V, confirming the 9.75x gain).
6. Verify the -3dB Points: Calculate the -3dB voltage ($4.8V \times 0.707 = 3.39V$). Manually tune the function generator up and down in frequency until the scope reads 3.39V p-p. Note these two frequencies ($f_L$ and $f_H$). The difference between them is your bandwidth. Divide $f_c$ by the bandwidth to verify your Q-factor.

For deeper analysis of active filter topologies and op-amp limitations at higher frequencies, consult the Analog Devices Active Filter Design Guide and the All About Circuits tutorial on active band pass filters. If you are pushing $f_c$ above 50 kHz, you will also need to review Texas Instruments SLOA093 to account for the op-amp's Gain-Bandwidth Product (GBW) derating your Q-factor.