When designing LC filters or DC-DC output stages, choose series inductance when you need to multiply total inductance while maintaining the lowest current rating of the chain, and choose parallel inductance when you need to divide inductance while multiplying the saturation current (Isat) handling. For high-current buck converters (>5A), parallel is the default; for high-impedance RF chokes or low-current Pi-filters, series wins. This guide breaks down the exact node behavior, failure modes, and a concrete design walkthrough so you can pick the right topology and part numbers without guessing.

Topology Breakdown and Node Behavior

Inductors in series and parallel follow the exact same mathematical rules as resistors, but with a critical physical caveat: mutual coupling. If unshielded inductors are placed too close together, their magnetic fields interact, invalidating the standard formulas. Always assume ideal math only applies to shielded parts or widely spaced unshielded parts.

Series Topology

Node Path: Vin → [L1] → Node A → [L2] → Vout
Formula: Ltotal = L1 + L2
In a series configuration, the same current flows through both components. The total DC resistance (DCR) adds up, which increases copper losses (I²R heating). The total inductance increases, which lowers the cutoff frequency of your filter but also slows down transient response.

Parallel Topology

Node Path: Vin splits to Node A1 (L1) and Node A2 (L2), recombining at Vout
Formula: Ltotal = (L1 × L2) / (L1 + L2) [for two equal inductors: L/2]
In parallel, the current divides between the branches. This effectively halves the DCR and doubles the current handling capability, making it ideal for high-power rails where a single inductor would physically saturate or overheat.

Behavior Matrix: What Changes When One Element Shifts
Configuration Parameter Changed Effect on Total L Effect on Current Handling Effect on Total DCR
Series Increase L1 value Increases proportionally Unchanged (limited by lowest Isat) Increases
Series Decrease L1 DCR No change Unchanged Decreases slightly
Parallel Increase L1 value Increases (approaches L2 value) Current shifts toward L2 Increases slightly
Parallel Decrease L1 DCR No change Current shifts toward L1 Decreases

Failure Modes at the Extremes: Open and Short Scenarios

Understanding how series and parallel inductance behaves when a component fails is critical for safety and reliability, especially in power supplies where a failed filter can pass unfiltered switching noise or short the input rail to the load.

Safety Note: A shorted inductor in a parallel configuration on a DC-DC converter output will typically result in a direct short from Vin to Vout (or Vout to GND depending on topology), potentially destroying the load if the upstream overcurrent protection does not trip in microseconds.
  • Series Open (L1 fails open): The circuit breaks completely. Vout drops to zero. The load loses power, but downstream components are generally protected from overvoltage.
  • Series Short (L1 shorts internally): Total inductance drops to just L2. The filter cutoff frequency spikes, and ripple current increases dramatically. The remaining inductor (L2) may now saturate due to the higher peak-to-peak ripple, leading to thermal runaway.
  • Parallel Open (L1 fails open): All current is forced through L2. Total inductance increases to L2. If L2 cannot handle the full load current, it will saturate, overheat, and eventually fail short or open.
  • Parallel Short (L1 shorts internally): This creates a dead short across the parallel node. Total inductance drops to zero. In a buck converter, this shorts the switching node to the output, causing catastrophic overvoltage and overcurrent unless the controller has fast OCP (Overcurrent Protection).

Design Walkthrough: Sizing Inductors for a 10A Buck Converter

Let’s design the output filter for a 12V to 3.3V synchronous buck converter delivering a continuous 10A load. Our controller targets a switching frequency of 500 kHz, and the math dictates we need 4.7µH total inductance with a minimum saturation current (Isat) of 15A (accounting for a 30% ripple margin).

Attempt 1: Series Configuration

We try to build 4.7µH by placing two 2.2µH inductors in series. We select the Wurth Elektronik 74437346022 (2.2µH, shielded).
The Math: 2.2µH + 2.2µH = 4.4µH (Close enough to 4.7µH).
The Failure: The Isat of this specific Wurth part is only 6.5A. Because current is identical in a series chain, the moment our 10A load turns on, both inductors saturate instantly. The inductance collapses, ripple current spikes, and the MOSFETs blow. Series fails here due to current limitations.

Attempt 2: Parallel Configuration (The Winner)

We need to divide the inductance while multiplying the current handling. We target ~10µH per branch to yield ~5µH total. We select the Coilcraft XEL5030-100 (10µH, ultra-low DCR, shielded).
The Math: (10µH × 10µH) / (10µH + 10µH) = 5.0µH total.
The Current: Each XEL5030-100 has an Isat of 11A. In parallel, the 10A load splits roughly 5A/5A. We are well below the 11A saturation threshold per branch, and the total DCR drops from 10.5mΩ to 5.25mΩ, cutting copper losses in half.
Parallel wins decisively for high-current rails.

Layout Tip: Even though the XEL series is shielded, always route the parallel traces symmetrically. If the PCB trace resistance to Node A1 is higher than to Node A2, the current will not split 50/50, and one inductor will run hotter than the other.

Breadboard Testing Protocol

Before committing to a PCB layout, verify your parallel or series inductance on the bench. Parasitic capacitance and mutual coupling can ruin your theoretical math.

  1. De-energize and Isolate: Ensure the circuit is powered off and capacitors are discharged. Lift one leg of the inductor network if it is connected to active silicon (like a MOSFET drain) to prevent the meter from reading parallel semiconductor junctions.
  2. Measure DCR: Use a high-precision multimeter (4-wire Kelvin measurement if possible) to measure the DC resistance across the entire network. Compare this to the sum (series) or parallel equivalent of the datasheet DCR values.
  3. Measure Inductance (LCR Meter): Set your LCR meter to 100 kHz (standard for power inductors) and measure the total L.
    • If Series: Expect L1 + L2. If the reading is higher than the sum, your inductors are mutually coupled in-phase. Move them 2 inches apart and re-measure.
    • If Parallel: Expect L/2. If the reading is erratic, check for ground loops in your test leads.
  4. Thermal Check: Inject a DC current equal to your expected load using a bench power supply with current limiting. Let it run for 5 minutes. Use a thermal camera or thermocouple to ensure neither inductor exceeds 85°C. In parallel setups, verify both inductors are within 5°C of each other.

The Final Decision Tree

Use this decision matrix to lock in your topology and part selection for your next power or filter design.

Design Constraint If True... Topology Choice Concrete Default Pick (2026)
Load Current > 8A Single inductor Isat is too low or DCR causes >2% efficiency loss Parallel 2x Coilcraft XEL5030-100 (10µH each)
Need High Impedance at RF Targeting >100MHz noise rejection with low DC current (<100mA) Series 2x Murata LQH32PN1R0 (1µH each, 0805 footprint)
PCB Height < 2mm Cannot fit a single large high-current drum core Parallel 2x TDK SPM5030 (Low-profile metal alloy)
Need Exact Odd-Value L Standard E12 values don't hit your specific filter cutoff Series Combine 4.7µH + 1.5µH standard shielded parts

The Default Recommendation: For 90% of modern DIY and commercial DC-DC power supply designs operating above 3A, parallel inductance using shielded metal-alloy or ferrite SMD parts is the superior choice. It solves the saturation problem, lowers DCR, and provides a graceful (though degraded) failure mode if one branch opens. Stick to parallel for power, and reserve series for low-current signal filtering.