Designing an Active Gain High Pass Filter: Theory, Formulas, and Bench Testing

A gain high pass filter is an active electronic circuit that blocks low-frequency signals and DC while simultaneously amplifying the remaining high-frequency AC signals above a specific cutoff threshold. Unlike a simple passive RC network that only attenuates unwanted low frequencies (and suffers insertion loss on the passband), an active topology uses an operational amplifier to strip away DC offsets and boost the AC signal amplitude in a single stage. This is a critical building block when conditioning real-world sensor data or audio signals for microcontrollers and analog-to-digital converters (ADCs).

The Core Mechanics: What It Changes in a Real Circuit

In a standard passive high-pass filter (just a capacitor in series with a resistor to ground), the maximum theoretical voltage gain is 1 (0 dB). In reality, due to source impedance and load interactions, you always lose a bit of signal. What an active gain high pass filter changes in a real installation is the ability to shift the signal baseline and scale the amplitude simultaneously.

When you feed a raw sensor signal into an active filter, the circuit performs two jobs:

1. DC Blocking: It removes static charge, thermal drift, or DC bias voltages that would otherwise offset your signal and eat up your ADC's dynamic range.
2. Amplification: It multiplies the remaining AC waveform so that millivolt-level sensor spikes become robust volt-level signals that a microcontroller can read without external pre-amplification.

For a deeper look at op-amp topologies, the All About Circuits guide on active filters provides excellent baseline schematics for Sallen-Key and multiple-feedback layouts.

Worked Numeric Example: Sizing Components for a 50Hz Cutoff

Let’s design a first-order, non-inverting active high-pass filter. We want a cutoff frequency ($f_c$) of 50Hz to block 60Hz mains hum (wait, 50Hz passes 60Hz. Let's target a 100Hz cutoff to block 50/60Hz hum and low-frequency mechanical vibrations) and a voltage gain ($A_v$) of 10 (20 dB).

Step 1: Calculate the RC High-Pass Network

The formula for the cutoff frequency is $f_c = \frac{1}{2 \pi R C}$. We need to pick standard component values. Let’s select a standard ceramic capacitor $C = 100\text{nF}$ ($0.1\mu\text{F}$). Solving for $R$:

$R = \frac{1}{2 \pi \times 100 \times 100 \times 10^{-9}} \approx 15.9\text{k}\Omega$

The closest standard E12 resistor value is 15kΩ, which shifts our actual cutoff slightly to roughly 106Hz. This is perfectly acceptable for most hobbyist and industrial sensor applications.

Step 2: Set the Non-Inverting Gain

For a non-inverting op-amp configuration, the gain formula is $A_v = 1 + \frac{R_f}{R_g}$. To get a gain of 10:

$10 = 1 + \frac{R_f}{R_g} \implies \frac{R_f}{R_g} = 9$

If we choose $R_g = 1\text{k}\Omega$, then $R_f$ must be $9\text{k}\Omega$. The closest standard value is 9.1kΩ, yielding a real-world gain of 10.1 (20.08 dB).

Where You Meet This in Practice

You will rarely see a gain high pass filter used in high-fidelity studio audio (where passive AC coupling followed by discrete transistor gain stages is preferred for noise reasons). Instead, this circuit dominates in embedded systems and sensor conditioning:

Real-World Scenario Walkthrough: The ESP32 Piezo Sensor Failure

To understand why this topology matters, let’s look at a common bench failure involving an ESP32 and a piezo disc.

The Setup

A maker is building a glass-break detector using a 27mm piezo disc wired to an ESP32 DevKit v1. The goal is to detect high-frequency acoustic shocks. The ESP32 ADC operates on a 3.3V reference with 12-bit resolution (4096 steps, roughly 0.8mV per step).

The Numbers (First Attempt - Passive)

The maker uses a simple passive AC-coupling network: a 1µF capacitor in series, followed by a 1MΩ pulldown resistor to ground, feeding directly into GPIO 34. When tapped, the piezo generates a 50mV AC spike, but it also accumulates a wandering 1.5V DC static charge from handling.

The Outcome

The ESP32 serial monitor shows random, erratic ADC readings. Genuine glass-break taps are completely missed.

What Went Wrong

The passive capacitor successfully blocked the 1.5V DC static charge. However, because the ESP32 ADC has a notoriously high input impedance mismatch and a noise floor of roughly ±50mV, the raw 50mV piezo spike was buried in the noise. The ADC was only seeing a swing of about 60 steps (50mV / 0.8mV), which the software debounce algorithm filtered out as jitter.

The Fix: Active Gain High Pass Filter

The maker swapped the passive network for an LM358 op-amp configured as an active gain high pass filter. They set the cutoff to 200Hz (using a 10nF cap and 82kΩ resistor) to ignore low-frequency table bumps, and set the gain to 50 (using a 1kΩ $R_g$ and 49.9kΩ $R_f$).

Now, the 50mV AC spike is multiplied by 50, becoming a 2.5V pulse. This spans over 3,000 ADC steps, blowing past the ESP32’s noise floor and resulting in flawless, repeatable knock detection. For more on sensor signal conditioning, refer to the Texas Instruments Op-Amp Design Resources.

Bench Debugging: Numbered Steps for Common Failures

When your active filter output looks wrong on the oscilloscope, follow this decision path before blaming the op-amp.

1. Check the Virtual Ground (Single Supply Trap): If you are powering your LM358 from a single 5V supply, you cannot reference your input to 0V. The op-amp cannot output negative voltages. You must create a $V_{CC}/2$ virtual ground (using a resistor divider and a bypass capacitor) and bias both the input and the non-inverting terminal to 2.5V. If your output is pinned at 0V or 5V, you likely forgot the bias network.
2. Verify the Gain-Bandwidth Product (GBWP): The LM358 has a GBWP of about 1MHz. If you set a gain of 100, your maximum usable frequency drops to 10kHz. If you are trying to filter and amplify a 50kHz ultrasonic sensor signal with an LM358, the op-amp will heavily attenuate it. Upgrade to a TL072 or OPA2134 for higher frequency audio/ultrasonic work.
3. Measure the Slew Rate: If your high-frequency spikes look like "triangles" instead of sharp edges on the scope, you are hitting the slew rate limit. The LM358 slews at roughly 0.5V/µs. A 2V spike takes 4µs to rise. If your application demands nanosecond edge preservation, you need a high-speed op-amp like the LM318.
4. Check for Rail Clipping: Ensure your amplified signal does not exceed $V_{CC} - 1.5V$ (for standard op-amps) or the supply rails (for rail-to-rail output types). If your 3.3V circuit outputs a waveform that flattens at 2.8V, your gain is too high for the input signal amplitude.

FAQ: Active Gain High Pass Filter Nuances