When dealing with inductance in an AC circuit, the core principle is simple: inductors resist changes in current. In an alternating current environment, this resistance manifests as inductive reactance ($X_L$), which scales linearly with frequency ($X_L = 2\pi fL$). This predictable frequency-dependent behavior makes the inductor the ultimate tool for choking high-frequency AC ripple while allowing DC or low-frequency signals to pass with minimal loss.
In this guide, we will break down the series RL (Resistor-Inductor) low-pass filter topology. We will map the nodes, analyze the math with real component values, contrast it against RC alternatives, and walk through a bench-test procedure you can replicate on your workbench today.
The Series RL Low-Pass Topology: Nodes and Behavior
The most common configuration for leveraging inductance in an AC circuit for signal or power conditioning is the series RL low-pass filter. The topology is straightforward:
- Node IN ($V_{in}$): The AC or mixed-signal source connects to the first lead of the inductor (L1).
- Node OUT ($V_{out}$): The junction where the second lead of L1 meets the first lead of the resistor (R1). This is your filtered output.
- Node GND: The second lead of R1 connects to the common ground return.
In this arrangement, the inductor is the series impedance element, and the resistor is the shunt element. As frequency increases, $X_L$ increases, dropping more voltage across L1 and leaving less at Node OUT.
Component Behavior Matrix
Understanding how tweaking one variable affects the whole system is critical for iterative design. Here is what happens when you change a single parameter in a series RL low-pass topology:
| Parameter Changed | Effect on Cutoff Freq ($f_c$) | Effect on DC Pass-Through | Effect on High-Freq Attenuation |
|---|---|---|---|
| Increase Inductance (L) | Decreases (rolls off earlier) | Unchanged (ignoring DCR) | Increases (steeper drop) |
| Decrease Inductance (L) | Increases (rolls off later) | Unchanged | Decreases |
| Increase Resistance (R) | Increases (rolls off later) | Decreases (more voltage drop) | Decreases |
| Decrease Resistance (R) | Decreases (rolls off earlier) | Increases (less voltage drop) | Increases |
Real-World Frequency Sweep Data
Abstract formulas are useless without real numbers. Below is the calculated behavior of an RL circuit using a 10mH inductor and a 100Ω resistor driven by a 1V peak-to-peak AC source. The cutoff frequency ($f_c = R / 2\pi L$) is approximately 1,591 Hz.
| Frequency (Hz) | Inductive Reactance ($X_L$) | Total Impedance ($Z$) | $V_{out} / V_{in}$ Ratio | Attenuation (dB) |
|---|---|---|---|---|
| 10 Hz | 0.63 Ω | 100.0 Ω | 0.999 | -0.01 dB |
| 100 Hz | 6.28 Ω | 100.2 Ω | 0.998 | -0.02 dB |
| 1,591 Hz ($f_c$) | 100.0 Ω | 141.4 Ω | 0.707 | -3.00 dB |
| 10,000 Hz | 628.3 Ω | 636.2 Ω | 0.157 | -16.1 dB |
| 100,000 Hz | 6,283 Ω | 6,284 Ω | 0.016 | -36.0 dB |
Why an RL Topology Over an RC Alternative?
A common question on the bench is why we would use an inductor (RL) instead of a capacitor (RC) for low-pass filtering, especially since capacitors are cheaper and smaller. The answer lies in current handling and DC resistance.
In an RC low-pass filter, the resistor is in series with the signal path. If you are filtering a power rail or a high-current PWM motor drive, that series resistor will dissipate massive amounts of heat ($I^2R$ losses) and cause an unacceptable DC voltage drop.
An inductor, conversely, passes DC with near-zero resistance—limited only by its internal DC Resistance (DCR), which is typically under 1Ω for power-grade components. When dealing with power inductors from manufacturers like Coilcraft, you can filter out 50kHz switching ripple from a 5A buck converter while losing only millivolts of DC potential. Furthermore, inductors do not suffer from the dielectric absorption or catastrophic short-circuit failure modes inherent to electrolytic capacitors.
Extreme Failure Modes: Series vs. Parallel Contrast
Understanding what breaks at the extremes is vital for troubleshooting and designing failsafe circuits. The failure behavior of an RL circuit changes drastically depending on whether the components are in series or parallel.
Series RL Failure Modes (L in series, R to ground)
- Inductor Opens (Most Common): Usually caused by exceeding the RMS current rating, melting the internal winding. Result: Node OUT drops to 0V. The circuit is completely dead.
- Inductor Shorts: Rare, but can happen if insulation breaks down and windings short to a metallic core. Result: If it shorts to ground, it blows the source fuse. If it experiences an internal turn-to-turn short, inductance drops severely, the cutoff frequency spikes, and high-frequency noise floods Node OUT.
- Resistor Opens: Caused by overvoltage transients. Result: Node OUT floats. The filter loses its shunt path, and high-frequency AC couples through the inductor's parasitic capacitance to the output.
Parallel RL Failure Modes (L and R in parallel across the load)
- Inductor Opens: Result: The circuit degrades to a purely resistive path. You lose the phase-shift and frequency-dependent filtering, but DC/low-frequency signals still pass. This is a "fail-safe" degradation.
- Resistor Opens: Result: The circuit becomes purely inductive. If the source is interrupted, the inductor will generate a massive flyback voltage spike ($V = -L \frac{di}{dt}$) that can destroy upstream semiconductors. This is a highly dangerous failure mode.
Design Walkthrough: Sizing Real Components for PWM Smoothing
Let’s design a practical circuit. We need to smooth a 3.3V, 25kHz PWM signal from an ESP32 GPIO pin into a clean, pseudo-DC analog reference voltage.
Step 1: Define the Cutoff Frequency
To effectively smooth a 25kHz square wave, the filter's cutoff frequency ($f_c$) should be at least one decade (10x) below the fundamental frequency. Let’s target $f_c = 2,500\text{ Hz}$.
Step 2: Select the Resistor
An ESP32 GPIO can safely source about 12mA to 20mA. To keep the load light and prevent the GPIO from browning out, we’ll limit the maximum current. Using Ohm’s law ($R = V/I$), a 330Ω resistor limits the current to exactly 10mA ($3.3\text{V} / 330\Omega$). This is a standard E12 value.
Step 3: Calculate the Inductance
Using the standard RL cutoff formula $f_c = \frac{R}{2\pi L}$, we rearrange to solve for L:
$L = \frac{R}{2\pi f_c} = \frac{330}{2 \times \pi \times 2500} = 0.0210\text{ H}$ (or 21mH).
Step 4: Pick a Real Part
The closest standard value is 22mH. We’ll select the Bourns 78FR22K-RC, a shielded radial inductor. Shielding is critical here to prevent the inductor's magnetic field from inducing noise into adjacent high-impedance analog traces on the PCB.
Step 5: Verify the Actual Cutoff
Recalculating with 22mH: $f_c = \frac{330}{2 \times \pi \times 0.022} = 2,387\text{ Hz}$. This is perfectly within our target margin. The DCR of this specific Bourns part is about 14Ω, which will cause a negligible DC voltage drop of $14\Omega \times 0.01\text{A} = 0.14\text{V}$ at maximum duty cycle.
Step-by-Step Breadboard Testing
Do not trust the simulation until you have verified it on the bench. Here is how to test your RL filter using a standard function generator (e.g., Siglent SDG1032X) and an oscilloscope (e.g., Rigol DS1054Z).
- Wire the Topology: Insert the 22mH inductor and 330Ω resistor into the breadboard in series. Connect the function generator’s BNC-to-alligator clip to Node IN and GND. Connect the oscilloscope probe (set to 1X attenuation) to Node OUT, clipping the probe ground to the breadboard ground rail.
- Verify the DC Baseline: Set the function generator to a 3.3V DC offset (or use a bench power supply). Measure Node OUT with a multimeter. You should read approximately 3.16V (accounting for the inductor's 14Ω DCR). If you read 0V, check for an open inductor winding.
- Inject the Fundamental Sine Wave: Switch the generator to a 25kHz sine wave, 3.3Vpp, 0V offset. On the scope, set the timebase to 10μs/div and use AC coupling. You should see heavy attenuation; the 25kHz signal should be reduced by roughly -20dB (about 1/10th of the input amplitude).
- Sweep for the Cutoff: Slowly dial the function generator frequency down. Watch the scope's peak-to-peak measurement. When the output amplitude hits exactly 0.707 of your input amplitude (approx 2.33Vpp), stop. Read the frequency on the generator display. It should read close to 2.38kHz, confirming your $f_c$ calculation.
- Test with Real PWM: Switch the generator to a 25kHz square wave (simulating the ESP32). Set the scope to DC coupling and 5ms/div. You should see a flat, steady DC line at roughly 1.65V (assuming a 50% duty cycle), with only a few millivolts of high-frequency sawtooth ripple visible when you zoom in the timebase.
Mastering inductance in an AC circuit requires moving beyond textbook reactance formulas and grappling with real-world parasitics, physical failure modes, and component selection. By understanding the RL topology deeply, you can design robust filters that handle both delicate signals and heavy power rails with equal precision.
