An op amp bandpass filter isolates a specific frequency band by combining high-pass and low-pass characteristics within an active feedback loop. Unlike passive RC filters, an active design provides voltage gain, prevents the load from dragging down the filter's Q-factor, and eliminates insertion loss. For ±15V dual-supply audio applications, the safest default part is the TL072 (JFET input, low noise, 3MHz GBW, ±22V max supply). For single-supply, low-speed DC sensor filtering (3V to 32V), the LM358 (BJT input, 1MHz GBW) remains the industry workhorse.

This guide skips the abstract physics and goes straight to the bench: pinouts, biasing networks, a fully calculated 1 kHz Multiple-Feedback (MFB) circuit, and how to test a suspect chip with a multimeter.

Anatomy and Pinout of the Operational Amplifier

Before wiring the feedback network, you must understand the physical package and the schematic symbol. The standard 8-pin Dual In-Line Package (DIP) or SOIC dual op-amp (like the TL072 or LM358) contains two independent amplifiers sharing a common power supply.

Standard 8-Pin Dual Op-Amp Pinout (Top View, Notch Left)
Pin Number Name Function in Bandpass Filter
1 OUT A Filter output signal (low impedance).
2 IN- A Inverting input; receives the feedback network (R2) and ground resistor (R3).
3 IN+ A Non-inverting input; tied to signal ground or VCC/2 bias voltage.
4 V- (or GND) Negative supply rail (e.g., -15V) or 0V ground in single-supply setups.
5 IN+ B Non-inverting input for the second amplifier channel.
6 IN- B Inverting input for the second amplifier channel.
7 OUT B Output for the second amplifier channel.
8 V+ Positive supply rail (e.g., +15V or +5V).

In the schematic symbol, the op-amp is drawn as a triangle. The non-inverting input (+) is on top or bottom depending on the designer's routing preference, but the inverting input (-) always feeds the negative feedback loop required for linear filter operation.

Operating Regions and Biasing for Filter Applications

To function as a linear filter, the op-amp must remain in its active (linear) region. If the output hits the power supply rails, the amplifier saturates, clipping your waveform and destroying the filter's frequency response. Furthermore, how you bias the non-inverting input (Pin 3) dictates whether the circuit can handle AC signals on a single power supply.

Op-Amp Operating Regions (Assuming ±15V Supply)
Region Output Voltage Input Differential (V+ minus V-) State in Filter Circuit
Linear (Active) -13V to +13V ~0V (Virtual Short) Normal operation; passing and amplifying the target frequency band.
Positive Saturation ~+13.5V Positive (> 0V) Fault state or massive overload; output is clipped at the positive rail.
Negative Saturation ~-13.5V Negative (< 0V) Fault state or massive overload; output is clipped at the negative rail.
Biasing Rule of Thumb: If using a dual supply (±15V), tie Pin 3 directly to 0V ground. If using a single supply (e.g., +12V), you must create an artificial mid-rail bias (VCC/2, or +6V) using a resistive voltage divider (e.g., two 10kΩ resistors) buffered by a 10µF bypass capacitor to ground. Tie Pin 3 to this VCC/2 node, and AC-couple your input and output signals with series capacitors.

Step-by-Step: Building a 1 kHz Audio Bandpass Filter

The Multiple-Feedback (MFB) topology is the standard for narrow to medium Q-factor bandpass filters. It uses a single op-amp, two capacitors, and three resistors. We will design a filter with a center frequency ($f_c$) of 1 kHz, a Gain of 10 (20 dB), and a Q-factor of 5 (yielding a bandwidth of 200 Hz).

For precise filter responses, you must use 1% metal film resistors and C0G/NP0 dielectric ceramic capacitors. Avoid X7R or Y5V capacitors; their capacitance shifts with applied voltage and temperature, which will detune your center frequency.

  1. Select the Capacitors: Choose $C_1$ and $C_2$ to be equal standard values. Let $C_1 = C_2 = 10\text{ nF}$ (0.01 µF).
  2. Calculate R1 (Input Resistor): Using the MFB equation $R_1 = \frac{Q}{2 \pi f_c C \cdot Gain}$, we get $R_1 = \frac{5}{2 \pi (1000) (10 \times 10^{-9}) (10)} \approx 7957\Omega$. Select the nearest 1% value: 8.06 kΩ.
  3. Calculate R2 (Feedback Resistor): Using $R_2 = \frac{Q}{\pi f_c C}$, we get $R_2 = \frac{5}{\pi (1000) (10 \times 10^{-9})} \approx 159154\Omega$. Select the nearest 1% value: 158 kΩ.
  4. Calculate R3 (Ground Resistor): Using $R_3 = \frac{Q}{2 \pi f_c C (2Q^2 - Gain)}$, we get $R_3 = \frac{5}{2 \pi (1000) (10 \times 10^{-9}) (50 - 10)} \approx 1989\Omega$. Select the nearest 1% value: 2.00 kΩ.
  5. Wire the Input: Connect your AC signal source to $R_1$. The other end of $R_1$ connects to the inverting input (Pin 2).
  6. Wire the Ground Node: Connect $R_3$ between the inverting input (Pin 2) and signal ground.
  7. Wire the Feedback Loop: Connect $R_2$ and $C_2$ in parallel between the inverting input (Pin 2) and the output (Pin 1).
  8. Wire the Input Capacitor: Connect $C_1$ between the junction of $R_1$/Pin 2 and signal ground.
  9. Verify Power and Bias: Apply ±15V to Pins 8 and 4. Place a 100nF bypass capacitor across the power pins as close to the IC body as possible. Tie Pin 3 to 0V ground.

For automated calculations and topology verification, the Analog Devices MT-223 Active Filter Tutorial provides exhaustive derivations for MFB and Sallen-Key variants.

Failure Modes and Multimeter Testing

Op-amps rarely fail without a cause. The most common killer in filter circuits is latch-up from input voltages exceeding the supply rails, or thermal runaway from driving a low-impedance load (like an 8Ω speaker) directly from the output pin.

Here is how to test a suspected dead op-amp using a standard digital multimeter (DMM):

  • In-Circuit Virtual Short Test (Powered): With the circuit powered and an AC signal applied, set your DMM to DC millivolts. Measure between Pin 2 (IN-) and Pin 3 (IN+). In the linear region, this differential voltage should be virtually zero (< 5mV). If you read several volts, the internal differential stage is blown, or the feedback loop is broken.
  • Power Pin Short Test (Unpowered): Set the DMM to resistance mode. Measure between Pin 8 (V+) and Pin 4 (V-). You should read a high resistance (typically >10kΩ). If it reads near 0Ω, the internal power rail bonding wires have melted together. Desolder and replace.
  • Input Protection Diode Test (Unpowered): Set the DMM to diode test mode. Place the red probe on Pin 4 (V-) and the black probe on Pin 2 (IN-). You should read a standard silicon diode drop (0.6V to 0.8V). Reverse the probes; it should read 'OL' (open loop). A short or open here indicates fried input protection diodes, usually caused by electrostatic discharge (ESD) or hot-plugging signal cables.

Frequently Asked Questions

What is the best op amp part number for an audio bandpass filter?

For high-fidelity audio (20 Hz to 20 kHz), the TL072 (Texas Instruments) and NE5532 (ON Semiconductor) are the undisputed safe defaults. The TL072 features JFET inputs, yielding incredibly low input bias currents and low voltage noise, making it ideal for high-impedance sensor inputs. The NE5532 uses BJT inputs, offering superior current noise performance and higher output drive capability (capable of driving 600Ω loads), making it the standard for line-level audio mixing consoles. Both are rated for ±22V maximum supplies and cost under $1.00 in single quantities.

Why is my op amp bandpass filter outputting a massive DC offset?

A large DC offset at the output usually stems from three issues. First, if you are using a single-supply configuration, you likely forgot to bias the non-inverting input (Pin 3) to VCC/2, forcing the output to slam against the ground rail to compensate for input bias currents. Second, if using a BJT-input op-amp like the LM358, the input bias current flowing through a high-value feedback resistor ($R_2$) creates an unwanted voltage drop; always ensure the DC resistance seen by both inputs is roughly matched. Third, check your capacitors. If you accidentally used an electrolytic capacitor instead of a ceramic film for $C_1$ or $C_2$, internal leakage current will charge the capacitor and shift the DC operating point.

Can I cascade two op amp bandpass filters to get a steeper rolloff?

Yes, cascading two identical 2nd-order MFB bandpass filters creates a 4th-order filter, steepening the rolloff from 12 dB/octave to 24 dB/octave outside the passband. However, you must account for component tolerance stacking. Even with 1% resistors, cascading two stages can cause the Q-factors to interact, creating an uneven 'ripple' in the passband rather than a flat response. For cascaded designs, it is highly recommended to use an active filter design tool or consult the All About Circuits Active Filters Chapter to calculate the specific damping coefficients (like Butterworth or Chebyshev alignments) required for each individual stage to maintain a flat overall response.

How do I calculate the bandwidth of an active bandpass filter?

The bandwidth (BW) is strictly defined by the center frequency ($f_c$) and the Quality factor (Q). The formula is simply $BW = \frac{f_c}{Q}$. For example, if your filter is centered at 1,000 Hz and has a Q of 5, the -3dB bandwidth is 200 Hz (spanning from roughly 900 Hz to 1,100 Hz). If you need a wider bandwidth (e.g., 1,000 Hz), you must lower the Q to 1. Lowering the Q in an MFB topology requires reducing the ratio of the feedback resistor ($R_2$) to the input resistor ($R_1$), which consequently lowers the voltage gain of the filter stage.