A series inductor opposes changes in current, making it the definitive component for low-pass filtering, EMI choking, and current smoothing. Unlike a series resistor, which dissipates energy as heat to achieve filtering, a series inductor stores energy in a magnetic field and releases it, allowing DC or low-frequency signals to pass with minimal power loss. If you need to smooth a 20 kHz PWM signal into a clean DC voltage or block high-frequency switching noise from a sensitive analog sensor, the series inductor topology is your baseline architecture.
The Series Inductor Topology: Nodes, Behavior, and Basics
The most common configuration for signal and power smoothing is the LC low-pass filter, which places an inductor in series with the signal path and a capacitor in parallel (shunt) to ground. Here is the standard node topology:
- Node A (Input): Receives the raw signal (e.g., 5V PWM from a microcontroller GPIO or a switching regulator).
- L1 (Series Inductor): Connected between Node A and Node B. It presents high impedance to high-frequency harmonics ($Z_L = j\omega L$) while passing DC with only its parasitic DC resistance (DCR).
- Node B (Output): The filtered output node, delivering smoothed DC to the load.
- C1 (Shunt Capacitor): Connected between Node B and Ground. It provides a low-impedance path to ground for the high-frequency noise that L1 blocks.
Behavior Matrix: What Changes When Elements Shift?
| Variable Changed | Effect on Cutoff Frequency ($f_c$) | Effect on Transient Response | Effect on DC Output |
|---|---|---|---|
| Inductance (L) Increases | Drops (filters lower frequencies) | Slower (more ringing potential) | Unchanged (ideal) / Slight drop (real DCR) |
| Capacitance (C) Increases | Drops (filters lower frequencies) | Slower (damps ringing) | Unchanged |
| Input Frequency Increases | N/A | N/A | Output ripple decreases (more attenuation) |
| Load Current Increases | N/A | Voltage droop at Node B | Drops due to $I \times DCR$ loss |
Why Choose a Series Inductor Over a Series Resistor?
When designing a low-pass filter, you can use an RC (resistor-capacitor) or an LC (inductor-capacitor) topology. For low-current signal lines, RC is cheaper and takes less board space. But for power delivery, motor drives, or high-current LED strings, the series inductor wins decisively.
| Criteria | Series Resistor (RC Filter) | Series Inductor (LC Filter) |
|---|---|---|
| DC Power Dissipation | High ($I^2R$ loss). A 10Ω resistor at 500mA burns 2.5W. | Negligible. A 1mH choke with 0.1Ω DCR at 500mA burns only 0.025W. |
| DC Voltage Drop | Significant. Steals voltage headroom from the load. | Minimal. Only drops voltage based on wire resistance (DCR). |
| Filter Order | First-order (-20 dB/decade roll-off). | Second-order (-40 dB/decade roll-off) when paired with a shunt cap. |
| Component Size/Cost | Very small, extremely cheap (0402 resistor). | Larger, more expensive (shielded power inductor). |
Design Walkthrough: 20 kHz PWM Smoothing Filter
Let’s design a filter to smooth a 5V, 20 kHz PWM signal (50% duty cycle) from an ESP32 into a clean 2.5V DC reference for an analog circuit. We need a second-order LC filter to adequately crush the 20 kHz fundamental and its odd harmonics.
1. Select the Cutoff Frequency ($f_c$)
We want the cutoff frequency to be at least one decade below the PWM frequency to ensure deep attenuation. Let’s target $f_c \approx 1.5$ kHz.
2. Pick Real Component Values
Using the formula $f_c = \frac{1}{2 \pi \sqrt{L \times C}}$, we can select standard off-the-shelf parts:
- L1: 1 mH. We will use the Bourns 78FR01K-RC (radial molded choke, 1 mH, DCR = 1.2 Ω, $I_{sat}$ = 330 mA). Cost: ~$0.45.
- C1: 10 µF. We will use the Murata GRM21BR61E106KA73 (0805 X5R MLCC, 25V rating). Cost: ~$0.10.
3. Verify the Math
$f_c = \frac{1}{2 \pi \sqrt{0.001 \times 0.00001}} = \frac{1}{2 \pi \sqrt{10^{-8}}} = \frac{1}{2 \pi \times 0.0001} \approx 1591$ Hz.
At 20 kHz (12.5x the cutoff frequency), a second-order filter provides roughly -42 dB of attenuation. The 2.5V peak-to-peak ripple at Node A will be reduced to approximately 20 mV of ripple at Node B—well within the noise floor for most analog ADCs.
4. Check the DC Current Limits
If the analog circuit draws 50 mA of DC current, the voltage drop across L1 is $V_{drop} = I \times DCR = 0.050A \times 1.2\Omega = 60$ mV. Your 2.5V output will sit at 2.44V. Furthermore, the 50 mA load is well below the Bourns choke's 330 mA saturation current limit, ensuring the inductance holds steady at 1 mH. For a deep dive on how core materials affect this limit, refer to Coilcraft's guide on inductor saturation.
Failure Modes: What Breaks at the Extremes?
Understanding how a series inductor fails is critical for troubleshooting and designing protective circuits. Here is the failure-mode contrast for the LC topology:
- L1 Opens (Wire breaks inside the choke): The DC path is severed. Node B drops to 0V (or floats to leakage voltage). The load loses power entirely. This is a safe failure mode.
- L1 Shorts (Internal winding insulation melts): The inductor becomes a piece of wire. The filter reverts to a first-order capacitive filter. The 20 kHz PWM square wave hits C1 directly. Because a capacitor acts as a near-short to high frequencies, the microcontroller GPIO or PWM driver will attempt to source massive instantaneous current to charge C1 on every rising edge. This can instantly fry the driving MOSFET or microcontroller pin.
- C1 Shorts (Dielectric breakdown): Node B is pulled directly to ground. L1 now acts as a simple series current limiter. The driver will see a dead short, limited only by L1's DCR and the driver's internal resistance, likely triggering overcurrent protection or burning out the driver.
- C1 Opens (Capacitor cracks/loses connection): The circuit becomes a pure series inductor feeding the load. If the load is highly resistive, it forms an RL filter with a much higher cutoff frequency, allowing severe PWM ripple to reach the load.
Step-by-Step Breadboard Testing Guide
Do not trust SPICE simulations blindly; parasitic breadboard capacitance and inductor ESR will alter your results. Follow this bench procedure to validate your series inductor filter.
- Configure the Source: Set your function generator to output a 0–5V square wave at 20 kHz with a 50% duty cycle. Connect the generator's ground to your breadboard's ground rail.
- Place the Components: Insert the Bourns 1 mH choke so one leg is on the input rail (Node A) and the other is on an isolated middle rail (Node B). Place the 10 µF MLCC between Node B and the ground rail.
- Probe the Input (Node A): Connect your oscilloscope probe to Node A. Verify you see a clean 5Vpp square wave. If you see massive ringing on the edges, your function generator's output impedance is mismatched; add a 50 Ω series terminator at the source.
- Probe the Output (Node B): Move the probe to Node B. You should see a flat line at ~2.5V. Switch the scope to AC coupling and increase the vertical scale to 10 mV/div to measure the residual ripple.
- Load Testing: Connect a 100 Ω power resistor between Node B and ground to simulate a 25 mA load. Observe the DC voltage drop. It should dip slightly due to the inductor's DCR, but the ripple amplitude should remain largely unchanged.
For further reading on how series RL and LC circuits behave under alternating current, All About Circuits provides an excellent breakdown of the phasor math involved.
Series Inductor FAQ
How do you calculate the voltage drop across a series inductor in a DC circuit?
In a pure DC steady-state circuit, an ideal inductor acts as a short circuit, meaning the voltage drop is exactly 0V. However, real-world inductors are made of physical copper wire, which has a specified DC Resistance (DCR). To calculate the real voltage drop, simply use Ohm's Law: $V_{drop} = I_{load} \times DCR$. For example, if your load draws 200 mA and your inductor has a DCR of 0.5 Ω, the DC voltage drop is 0.1V. Always check the datasheet for the DCR value before finalizing your power budget.
What happens when a series inductor reaches core saturation?
Inductors rely on a magnetic core (ferrite, powdered iron, etc.) to multiply their inductance. If the DC current passing through the coil exceeds the component's saturation current rating ($I_{sat}$), the core's magnetic domains align completely and can no longer store additional magnetic energy. When this happens, the inductance plummets—often dropping to just the air-core value (a fraction of a microhenry). In a filter, this means your cutoff frequency skyrockets, and high-frequency noise passes straight through. In a switching regulator, core saturation causes massive current spikes that can destroy the switching MOSFET.
Can I put a series inductor on the ground path instead of the supply line?
Electrically, placing an inductor on the ground return path will filter the current loop just as it would on the high-side supply line. However, from an EMC (Electromagnetic Compatibility) and signal integrity perspective, this is a terrible idea. Placing an inductor on the ground path creates a "floating ground" for the load at high frequencies. The load's ground reference will bounce relative to the system ground, turning the entire load circuit into an antenna that radiates common-mode EMI. Always place your series filtering inductors on the high-side (supply) line, keeping the ground plane solid and continuous.
