A pass band filter circuit (more commonly referred to as a band-pass filter) allows a specific range of frequencies to pass while attenuating signals above and below that band. For audio, sensor conditioning, and RF applications, an active multiple-feedback (MFB) topology using an operational amplifier is the standard choice. It provides signal gain, avoids the need for bulky inductors, and maintains a high quality factor (Q) without the loading effects that plague passive designs.
The Active Pass Band Filter Circuit Topology
When designing an active pass band filter circuit, you generally choose between the Sallen-Key and the Multiple Feedback (MFB) topologies. We use the MFB topology here. While the Sallen-Key is excellent for low-pass and high-pass stages, its band-pass variant struggles with stability at high Q-factors (Q > 10) and requires tight component matching. The MFB topology, by contrast, is highly stable, relies on standard capacitor values, and provides excellent phase response.
Topology Description and Node Labels
The MFB band-pass filter uses one op-amp, two capacitors, and three resistors. Here is the node mapping for the standard inverting MFB configuration:
- Vin (Input Node): The AC signal enters through resistor R1.
- Node A (Summing Junction): The inverting input (-) of the op-amp. This is a virtual ground where R1, R2, R3, C1, and C2 converge.
- Node B (Reference Node): The non-inverting input (+) of the op-amp, tied to circuit ground (or a Vcc/2 bias in single-supply designs).
- Vout (Output Node): The op-amp output, which feeds back through C2 and R2 to Node A.
Design Walkthrough: Picking Real Component Values
Let us design a pass band filter circuit for an audio sensor application with the following target specifications:
- Center Frequency ($f_c$): 1 kHz
- Gain ($H_0$): 2 (6 dB)
- Quality Factor ($Q$): 5 (Yielding a Bandwidth of 200 Hz)
We will use a TL072 op-amp, which offers low noise and high slew rate for audio frequencies. According to Analog Devices' active filter design guidelines, we start by selecting a standard capacitor value to anchor our resistor calculations. Let us choose $C_1 = C_2 = C = 10 \text{ nF}$.
Using the standard MFB design equations:
- Calculate R2: $R_2 = \frac{Q}{\pi \cdot f_c \cdot C} = \frac{5}{\pi \cdot 1000 \cdot 10 \times 10^{-9}} \approx 159.15 \text{ k}\Omega$. We will use 158 k\Omega (a standard 1% E96 series resistor).
- Calculate R1: $R_1 = \frac{R_2}{2 \cdot H_0} = \frac{158 \text{ k}}{4} = 39.5 \text{ k}\Omega$. We will use 39.2 k\Omega (E96 series).
- Calculate R3: $R_3 = \frac{R_2}{4Q^2 - 2H_0} = \frac{158 \text{ k}}{4(25) - 4} = \frac{158 \text{ k}}{96} \approx 1.645 \text{ k}\Omega$. We will use 1.65 k\Omega (E96 series).
By selecting 1% tolerance resistors and 5% C0G/NP0 dielectric capacitors, the physical circuit will land within 1.5% of the 1 kHz target center frequency, avoiding the severe drift seen when using X7R ceramics or 5% carbon film resistors.
Element Behavior and Failure Modes
Understanding how each component shapes the transfer function is critical for debugging. The table below details what happens when element values drift, followed by a look at catastrophic failure modes.
| Component | Primary Parameter Affected | If Value Increases | If Value Decreases |
|---|---|---|---|
| R1 (Input) | Gain ($H_0$) | Gain drops; $f_c$ shifts slightly lower | Gain increases; risk of op-amp clipping |
| R2 (Feedback) | Center Freq ($f_c$) & Q | $f_c$ drops; Q increases (narrower band) | $f_c$ rises; Q decreases (wider band) |
| R3 (Ground Ref) | Q-Factor | Q drops (band widens, less selective) | Q spikes (circuit may oscillate) |
| C1, C2 (Caps) | Center Freq ($f_c$) | $f_c$ drops proportionally | $f_c$ rises proportionally |
Failure Mode Contrast: Series vs. Parallel Networks
In a passive series RLC band-pass filter, the components form a single signal path. If the capacitor fails open, impedance becomes infinite, and the signal path is completely dead (0V out). However, in our active MFB parallel-feedback network, the feedback components (R2, C2) sit in parallel with the op-amp's output and inverting input. If R2 fails open, the DC negative feedback path is broken. The op-amp enters open-loop mode, acting as a comparator, and the output will instantly slam into the positive or negative supply rail, saturating the circuit rather than just silently killing the signal.
Breadboard Testing Step-by-Step
Before committing this pass band filter circuit to a printed circuit board, validate it on a breadboard. Parasitic capacitance on a breadboard (typically 2-5 pF per node) will slightly shift a 1 kHz filter, but it is negligible for audio prototyping. Referencing Texas Instruments' active filter application notes, follow this sequence to avoid chasing phantom oscillations.
- Establish Clean Power and Bias: Power the TL072 with a dual supply (+/- 9V). Place 100 nF bypass capacitors directly across the V+ and V- pins to ground. If using a single supply, create a Vcc/2 virtual ground using a buffered voltage divider and tie Node B to it.
- Verify DC Operating Point: With no AC input, use a multimeter to measure Node A and Vout. Both should read 0V (or Vcc/2 in single supply). If Vout is pegged to a rail, you have a wiring error in the feedback loop (likely an open R2 or swapped op-amp inputs).
- Inject and Sweep the AC Signal: Connect a function generator to Vin. Set it to a 1 kHz sine wave at 100 mVpp. Connect an oscilloscope to Vout. You should see a 200 mVpp sine wave (Gain = 2). Sweep the frequency from 100 Hz to 10 kHz. The amplitude should peak at 1 kHz and roll off at -20 dB/decade on the low end and -20 dB/decade on the high end (total -40 dB/decade band-pass response).
- Measure the -3 dB Points: Calculate the target -3 dB voltage: $200 \text{ mV} \times 0.707 = 141 \text{ mV}$. Sweep down in frequency until the scope reads 141 mV (should be ~900 Hz). Sweep up until it hits 141 mV again (should be ~1100 Hz). The difference is your bandwidth, confirming your Q-factor.
Pass Band Filter Circuit FAQ
Why is my pass band filter circuit oscillating at high frequencies?
High-frequency oscillation (usually in the MHz range) in an active pass band filter circuit is almost always caused by poor power supply decoupling or capacitive loading on the output. The TL072 and similar op-amps can become unstable if they are driving a capacitive load (like a long coaxial cable or a scope probe without a series isolation resistor). Add a 47 \Omega to 100 \Omega resistor directly in series with the op-amp's output pin (before the feedback network tap and the output connector) to isolate the capacitive load. Additionally, ensure your breadboard power rails have 100 nF ceramic capacitors placed as close to the op-amp VCC pins as physically possible.
How do I cascade two pass band filter circuits for a steeper roll-off?
To achieve a steeper attenuation slope (e.g., -80 dB/decade instead of -40 dB/decade), you cascade two identical MFB stages. However, you cannot simply wire them back-to-back without considering loading. Because the MFB topology has a low output impedance but a relatively complex input impedance, you should buffer the first stage or ensure the second stage's input resistor (R1) is at least 10 times larger than the first stage's output impedance. A better approach is to use a dual op-amp package (like the TL072), wire the output of the first op-amp directly into the R1 input of the second op-amp, and stagger the center frequencies slightly if you need a wider, flatter peak (a Butterworth or Chebyshev alignment).
Can I use a single-supply op-amp for an AC pass band filter circuit?
Yes, but you must bias the circuit. Op-amps like the LM358 or MCP6002 can run off a single 5V or 3.3V rail, but they cannot output negative voltages. To pass an AC signal, you must create an artificial mid-supply reference (e.g., 2.5V on a 5V system) using a precision voltage divider buffered by another op-amp stage. You then tie Node B (the non-inverting input) to this 2.5V reference instead of ground. The AC signal will swing above and below this 2.5V DC offset. Remember to AC-couple your input (via a series capacitor) and your output to block this DC bias from affecting the rest of your system.
