A mutual inductor—frequently specified as a coupled inductor in modern switch-mode power supply (SMPS) design—consists of two or more windings sharing a single magnetic core. Unlike a traditional transformer, which transfers energy instantaneously from primary to secondary, a mutual inductor in topologies like SEPIC, Cuk, or Zeta converters intentionally stores energy in a gapped core during the switch-on cycle and releases it during the switch-off cycle. The defining metric for these components is the coupling coefficient ($k$), which dictates how tightly the magnetic flux of one winding links to the other.
What Are Mutual Inductors and How Do They Work?
At the bench, you can think of a mutual inductor as two standard inductors forced to share the same magnetic real estate. The mutual inductance ($M$) is calculated as $M = k \sqrt{L_1 L_2}$, where $L_1$ and $L_2$ are the self-inductances of the individual windings and $k$ is the coupling coefficient (ranging from 0 to 1).
In a SEPIC (Single-Ended Primary-Inductor Converter), a high coupling coefficient ($k > 0.95$) is highly desirable. Tight coupling reduces the leakage inductance, which in turn minimizes the voltage spikes across your switching MOSFET when it turns off. If you use a loosely coupled mutual inductor, the un-coupled flux (leakage inductance) has nowhere to go, resulting in high-frequency ringing that can exceed the MOSFET's $V_{DS}$ rating and destroy the silicon.
For a deeper look at the mathematical foundation of flux linkage, the All About Circuits textbook chapter on mutual inductance provides an excellent breakdown of the underlying physics.
Mutual Inductor Types: Selection Matrix and Use Cases
Selecting the right physical construction depends entirely on your switching frequency, EMI constraints, and current requirements. Below is a comparison matrix of the most common mutual inductor constructions used in modern power electronics.
| Construction Type | Example Series | Typical Tolerance | Tempco (Inductance) | Typical Use Case |
|---|---|---|---|---|
| Shielded Ferrite Drum | Coilcraft MSD1260 | ±20% | -2000 ppm/°C | SEPIC/Cuk converters, noise-sensitive IoT boards |
| Unshielded Bobbin | Bourns SRN6045TA | ±20% | -3000 ppm/°C | Low-cost, non-critical filtering, low EMI environments |
| Toroidal (Gapped) | TDK VLF Series | ±15% | Variable (Core dependent) | High-current differential mode filtering, automotive |
| Planar / Flat Wire | Wurth Elektronik WE-FLY | ±10% | Low (Optimized geometry) | High-frequency (>1MHz) GaN/SiC converters |
Which type for which job? If you are designing a battery-powered device with strict radiated emissions limits (like a medical wearable), always choose a shielded ferrite drum core. The magnetic shield prevents the AC flux from coupling into nearby sensitive analog traces. If you are building a high-power, low-frequency (50-100kHz) off-line power supply where board height is not constrained, toroidal mutual inductors offer the lowest DC resistance (DCR) and highest thermal dissipation.
Decoding SMD Markings and the Dot Convention
Reading the markings on a physical mutual inductor is critical, especially for identifying the phase relationship between windings. Getting the phasing wrong in a Cuk converter will effectively short your input to ground through the switching node.
Reading the Inductance Code
Most SMD mutual inductors use a standard three-digit code for the primary inductance (both windings typically have identical turns ratios in SMPS applications):
- 101: 10 × 10^1 μH = 100 μH
- 470: 47 × 10^0 μH = 47 μH
- R47: The 'R' acts as a decimal point = 0.47 μH
The Dot Convention (Pin 1 Identification)
Mutual inductors will have a physical marking—usually a white dot, a painted stripe, or a beveled edge on the bobbin—indicating Pin 1 of the primary winding. Pin 1 of the secondary winding is similarly marked or located diagonally opposite, depending on the footprint.
For comprehensive footprint and phasing guidelines, manufacturers like Coilcraft maintain excellent application libraries on coupled inductor selection that detail exact pad layouts to minimize parasitic capacitance.
Failure Modes and Visual Diagnostics
When a mutual inductor fails on the bench, it rarely does so silently. Here is how to diagnose the three most common failure modes based on visual and multimeter symptoms.
1. Core Saturation and Thermal Runaway
Visual Symptoms: The epoxy coating on the core may turn dark brown or yellow. You might see micro-cracks in the ferrite material, or the solder joints on the pads will look dull, crystalline, or reflowed due to extreme heat. The switching MOSFET is often shorted (drain to source).
The Cause: The peak current exceeded the saturation current ($I_{sat}$) rating. When a ferrite core saturates, its permeability drops to near that of air. The inductance collapses, and the component acts like a low-value resistor, allowing massive current spikes to flow.
2. Inter-Winding Insulation Breakdown
Visual Symptoms: Black carbon tracking or pitting visible on the plastic bobbin between the primary and secondary pins. A multimeter set to continuity or high-resistance mode will show a low-impedance path between Pin 1 (Primary) and Pin 3 (Secondary).
The Cause: High voltage transients (often from switching spikes exceeding the MOSFET's avalanche rating) arc across the thin enamel insulation of the copper wire where the windings cross over each other inside the bobbin.
3. Mechanical Winding Open
Visual Symptoms: The component looks perfectly pristine. No discoloration, no cracks. However, an ohmmeter reads "OL" (Open Loop) across the winding pins. Sometimes, under 10x magnification, you can see the copper wire broken right at the termination point where it solders to the metal pad.
The Cause: Mechanical shock (dropping the PCB) or excessive thermal stress during wave soldering/reflow causing the thin wire to snap at the rigid termination joint.
Safe Substitution Rules When the Exact Part is Obsolete
Supply chain shortages frequently force engineers to substitute mutual inductors. When the exact BOM part is out of stock, follow this hierarchy to substitute safely without redesigning your PCB:
- Match the Inductance ($L$): The primary and secondary inductances must be within ±10% of the original. Changing $L$ alters your control loop crossover frequency and can cause sub-harmonic oscillation.
- Never Compromise on $I_{sat}$: The replacement's saturation current must be equal to or greater than the original. Substituting a part with a lower $I_{sat}$ will result in core saturation and catastrophic switch failure.
- Check $I_{rms}$ (Thermal Limit): $I_{rms}$ dictates how much continuous DC current the wire can handle before overheating. If your replacement has a lower $I_{rms}$, it will act as a heater and eventually open-circuit.
- DCR (DC Resistance): Aim for a replacement with equal or lower DCR. Higher DCR will reduce your overall converter efficiency and increase the component's operating temperature.
- Footprint and Shielding: If you must use a different footprint, you can mount the substitute on its side and use flying wires (magnet wire) for prototyping. If you substitute an unshielded part for a shielded one, expect a 5-10 dB increase in radiated EMI; you may need to add a local copper shield can over the component.
Frequently Asked Questions
Can I use two separate standard inductors instead of a mutual inductor in a SEPIC converter?
Yes, you can. A SEPIC converter will function with two physically separate, uncoupled inductors ($k = 0$). However, doing so requires twice the PCB footprint and doubles the number of components. More importantly, uncoupled inductors mean the ripple current in the two inductors is not perfectly matched, which can increase the RMS current stress on the AC coupling capacitor, reducing its lifespan. Using a true mutual inductor with a 1:1 turns ratio halves the required inductance value per winding and significantly reduces output voltage ripple.
How do I measure the coupling coefficient (k) on the bench?
You need a precision LCR meter. First, measure the self-inductance of the primary ($L_1$) and secondary ($L_2$) with the other winding left open. Next, short the secondary winding completely. Measure the inductance across the primary terminals again; this reading is your leakage inductance ($L_{leak}$). The coupling coefficient is then calculated using the formula: $k = \sqrt{1 - (L_{leak} / L_1)}$. A high-quality SMPS mutual inductor should yield a $k$ value between 0.95 and 0.99.
Why did my replacement mutual inductor overheat when the inductance and DC current ratings matched perfectly?
You likely fell victim to AC core losses and the proximity effect. Datasheet $I_{rms}$ ratings are often based purely on DC resistance and a 40°C temperature rise. In high-frequency switching converters (especially >500kHz), the AC ripple current causes significant hysteresis and eddy current losses in the ferrite core. Furthermore, high-frequency AC current crowds to the outer skin of the wire (skin effect) and concentrates on the sides of adjacent wires (proximity effect), effectively increasing the AC resistance far above the published DC DCR. Always check the manufacturer's core loss vs. frequency graphs, not just the DC current table.
