Magnetism and induction describe how a changing magnetic field generates an electromotive force (voltage) across a conductor, and conversely, how current flowing through a conductor creates a magnetic field that resists changes in that current. In a real circuit, this physical law changes everything about how we handle switching transients, dictating energy storage limits and forcing us to manage voltage spikes when current is interrupted. The most common and destructive confusion hobbyists make is treating inductance (the property, measured in Henries) as the only spec that matters, while completely ignoring magnetic saturation (the physical limit of the core material, measured in Tesla or Amperes), which leads to instantly shorted MOSFETs when the core "runs out" of magnetism.
The Core Mechanics: Fields, Flux, and Flyback
At the bench, inductors and transformers are the physical manifestations of Faraday’s Law of Induction. When you push DC current through a coil of wire, it builds a magnetic field in the core material (usually ferrite or powdered iron). This field stores energy.
The critical parameter for power electronics is not just the inductance value ($L$), but the saturation current ($I_{sat}$). Every magnetic core has a maximum limit to how much magnetic flux it can hold. Once you hit that limit, the core saturates. When a core saturates, the inductance effectively drops to near zero, and the component stops acting like an inductor and starts acting like a dead short (a simple piece of wire). If this happens while a switching regulator's MOSFET is turned on, current ramps up uncontrollably, and the MOSFET will violently fail due to thermal runaway or avalanche breakdown.
Worked Numeric Example: Sizing a Buck Converter Inductor
Let’s apply this theory to a real-world design. You are building a 12V to 5V DC-DC buck converter to power a 2A Raspberry Pi cluster node. You are using a standard 500 kHz switching frequency ($f_{sw}$). What inductor do you actually need?
Step 1: Calculate the Duty Cycle ($D$)
$D = \frac{V_{out}}{V_{in}} = \frac{5V}{12V} = 0.416$
Step 2: Define the Ripple Current ($\Delta I_L$)
Industry standard practice targets a ripple current of 30% of the maximum load current to balance efficiency and transient response.
$\Delta I_L = 0.30 \times 2A = 0.6A$
Step 3: Calculate Required Inductance ($L$)
$L = \frac{V_{out} \times (1 - D)}{f_{sw} \times \Delta I_L}$
$L = \frac{5 \times (1 - 0.416)}{500,000 \times 0.6} = \frac{2.92}{300,000} = 9.73 \mu H$
Step 4: Determine Peak Current and Saturation Limit
$Peak Current = I_{load} + \frac{\Delta I_L}{2} = 2A + 0.3A = 2.3A$
You must select an inductor with a nominal value of 10 µH and a saturation current rating ($I_{sat}$) strictly greater than 2.3A. A safe engineering margin pushes this to at least 3.0A.
Where You Meet Magnetism and Induction in Practice
While switching power supplies are the most math-heavy application, the principles of magnetism and induction govern several other common jobsite and workbench scenarios:
- Relay and Solenoid Flyback Diodes: When you de-energize a mechanical relay coil, the collapsing magnetic field induces a reverse voltage spike that can easily exceed 100V, destroying the driving Arduino GPIO or bipolar transistor. A 1N4148 or 1N4007 diode placed in reverse bias across the coil provides a safe recirculation path for this induced current.
- GFCI Receptacles: A Ground Fault Circuit Interrupter uses a toroidal current transformer. The Line and Neutral wires pass through the center of the magnetic core. Under normal conditions, their magnetic fields cancel out perfectly. If current leaks to ground, the fields no longer cancel, inducing a voltage in the toroid's secondary winding that trips the mechanical latch within milliseconds.
- EMI Chokes on USB Cables: The cylindrical lump of plastic near the end of your monitor or printer cable is a ferrite bead. It relies on magnetic hysteresis to absorb high-frequency RF noise, converting the unwanted induced high-frequency energy into trace amounts of heat.
Decision Tree: Selecting Your Inductor
Choosing the right physical component requires matching your circuit's electrical demands to the core material and physical construction. Use this decision matrix to narrow down your selection.
| Application Constraint | If your circuit requires... | Then choose this core/construction type... |
|---|---|---|
| High Switching Frequency (> 1 MHz) | Low core losses at high frequencies to prevent overheating. | Ferrite core with distributed air gap (e.g., NiZn ferrite). |
| High DC Current (> 5A) | Resistance to early magnetic saturation and low DCR. | Carbonyl iron powder core or molded alloy powder (e.g., Wurth WE-LHMI). |
| Sensitive RF/Analog Circuitry | Minimal radiated magnetic EMI leaking into adjacent traces. | Magnetically shielded construction (enclosed drum core). |
| Low Cost / General Purpose (< 2A, < 500 kHz) | Adequate margin without paying for premium alloy powders. | Standard unshielded drum core ferrite (e.g., Bourns SRN series). |
The Default Pick: If you are building a generic 5V/2A buck converter and want a safe, proven default without over-engineering, buy the Bourns SRN6045TA-100M. It provides 10µH inductance with a 2.5A saturation limit, features a shielded construction to prevent EMI from disrupting nearby microcontrollers, and costs under $0.50 in single quantities from major distributors like Würth Elektronik or Digi-Key.
Frequently Asked Questions
Why does my inductor get hot even when the load is light?
Inductor heating comes from two sources: copper loss ($I^2R$ heating in the wire) and core loss. If the inductor is hot at light loads, you are experiencing core loss. This happens when the core material is poorly suited for your switching frequency, causing the magnetic domains inside the ferrite to physically grind against each other during rapid polarity reversals, generating heat. Check the manufacturer's core loss curves (usually measured in $mW/cm^3$) for your specific switching frequency.
Can I put two 20µH inductors in parallel to get 10µH and double the current rating?
Mathematically, yes ($20 \mu H / 2 = 10 \mu H$). Practically, this is a bad idea unless the inductors are perfectly matched. Slight differences in DC resistance (DCR) and core permeability will cause one inductor to draw more current and hit magnetic saturation before the other. Once the first one saturates, its impedance drops to zero, dumping the entire circuit current through it and causing a catastrophic failure. Always use a single, correctly rated component.
What is the difference between shielded and unshielded inductors?
An unshielded inductor (usually a bobbin/drum style) has an open magnetic path, meaning its magnetic flux lines radiate into the surrounding air. This can induce unwanted noise in nearby high-impedance analog traces or Hall-effect sensors. A shielded inductor encloses the coil in a magnetic material (like a molded powder box), containing the flux lines inside the component. Always use shielded inductors in mixed-signal boards containing ESP32 Wi-Fi antennas or precision ADCs.
