My Cart
Your Shopping Cart is Empty
Inductors: Principles, Types, and Practical Selection
|
0 comments
posted by Vincent Zhang
What inductors do, how core materials and construction affect real-world behaviour, and how to choose the right one for power, RF, and filtering tasks—with an honest look at where the rules of thumb break down.
Contents
1. What Is an Inductor?
An inductor stores energy in a magnetic field when current flows through it. That is the simplest definition—and like most simple definitions in engineering, it conceals a great deal. The component that does this can be a millimetre-scale chip on a smartphone PCB or a laminated-steel choke weighing several kilogrammes in an industrial power supply. Both obey the same physics; neither behaves quite like the textbook model.
Physically, an inductor is a coil of wire, often wound around a magnetic core. The fundamental relationship is Faraday's law of induction: a changing current through the coil produces a changing magnetic flux, which in turn induces a voltage that opposes the change in current. Expressed in circuit terms:
V = L · di/dt
where V is the induced voltage, L is the inductance in henries, and di/dt is the rate of change of current. A larger inductance produces a larger opposing voltage for the same rate of current change—which is why inductors are sometimes described, informally, as "electrical inertia." They resist changes in current the way mass resists changes in velocity.
That analogy has limits. Real inductors carry DC resistance, parasitic capacitance, and core losses that the pure-L model ignores. None of these is negligible in practice, and the art of inductor selection is largely about understanding which non-ideality will hurt you first.
Real Inductor — Equivalent Circuit
Figure 1: A real inductor modelled as a series inductance L, series DC resistance (DCR), and parallel parasitic capacitance Cp. All three elements influence behaviour at different frequencies.
Figure 2: A selection of common inductor types — toroidal wire-wound, shielded SMD power inductors, and a miniature multilayer chip inductor. Each form factor reflects a different balance of inductance, current rating, Q factor, and board footprint.
2. The Parameters That Define an Inductor
A datasheet lists perhaps eight or ten numbers for a given inductor. Five of them do most of the work in selection. Understanding the relationships between them—particularly the ones that pull in opposite directions—is more useful than memorising any single figure.
2a. Inductance (L)
The nominal inductance, measured in henries (H), is the headline specification. For power inductors, it is typically specified at 100 kHz with a small AC excitation. For RF inductors, the test frequency varies with the value—a 100 nH part might be specified at 100 MHz, while a 10 µH part might be measured at 7.9 MHz. The key point: inductance is not a constant. It varies with frequency, DC bias current, and temperature. A 10 µH power inductor at zero bias may measure 7 µH at 2 A of DC current. According to Coilcraft's application data, the effective inductance at 80% of rated saturation current can drop by 30% or more for ferrite-core parts.
2b. DC Resistance (DCR)
DCR is the resistance of the copper winding, measured at DC. It is the simplest parameter to understand and one of the hardest to trade off. Low DCR means thicker wire—which means a physically larger part. In a buck converter output inductor, every milliohm of DCR translates directly to I²R heat. As noted in Sagami's technical literature, terminal welding resistance alone can account for roughly 17% of the total DCR in miniature SMD inductors—a detail easy to miss if you calculate only from wire gauge and length.
The trade-off is straightforward but unforgiving: you can have low DCR, small size, or high inductance. Pick two.
2c. Quality Factor (Q)
The quality factor is the ratio of inductive reactance to effective series resistance at a given frequency:
Q = XL / Reff = 2πfL / Reff
A higher Q means lower energy loss per cycle. In a resonant tank circuit—an oscillator, a narrow bandpass filter—Q directly determines selectivity and phase noise performance. Small fixed inductors for general-purpose use typically achieve Q values around 50. Specialised RF inductors, particularly those using air or ceramic cores, can reach Q values of 70 to 120 in their optimised frequency band, according to Coilcraft application note Doc671.
Q is not flat across frequency. It rises from a low value at DC (where DCR dominates), peaks somewhere in the middle of the usable band, and then collapses as the self-resonant frequency is approached. Above SRF, the component's effective inductive Q is no longer meaningful—it behaves predominantly as a capacitor. Figure 3 shows the characteristic shape.
Figure 3: Representative Q factor versus frequency. At low frequencies, DCR dominates. Q peaks mid-band, then collapses as SRF is approached. Beyond SRF, the component behaves capacitively and Q loses its inductive meaning. (Illustrative only.)
2d. Self-Resonant Frequency (SRF)
The self-resonant frequency is perhaps the most frequently overlooked parameter and the one that causes the most field failures. Every real inductor has parasitic capacitance between adjacent turns of the winding. This capacitance, in parallel with the inductance, forms a resonant tank. At the SRF, the inductive and capacitive reactances cancel, and the component stops being an inductor altogether.
The standard design rule, promoted by Coilcraft and widely adopted, is that the operating frequency should be at least a factor of ten below the SRF for applications where predictable inductive behaviour is required—filters, matching networks, resonant converters. This ensures the effective inductance remains within a few percent of the nominal value.
But there is an important exception. For RF choke applications, where the goal is to block an AC signal while passing DC, the selection rules change. For narrowband choking, selecting an inductor whose impedance peak or SRF lies near the frequency to be attenuated can provide useful isolation—as in a bias tee, where choosing an inductor with SRF near the lowest operating frequency can maximise RF-to-DC port isolation. For broadband choking, however, the impedance must be examined across the entire band. A single inductor’s SRF peak is narrow, and beyond SRF the component turns capacitive, actually reducing isolation at higher frequencies. Broadband chokes often require several inductors in series (staggered SRF values) or lossy ferrite beads to maintain high impedance over a wide range.
The SRF / Inductance Trade-off
Higher inductance values require more turns of wire. More turns mean greater inter-winding capacitance. The result: as inductance goes up, SRF comes down. This is not a manufacturing defect; it is a physical constraint. A 100 µH ferrite inductor might have an SRF of 5 MHz, while a 100 nH air-core coil might resonate at over 1 GHz. When both high inductance and high SRF are needed, designers sometimes place smaller inductors in series to obtain higher total inductance while avoiding the very low SRF of a single high-value part. The resulting SRF still depends on layout, coupling, and total parasitic capacitance, so the assembled network should be verified.
2e. Saturation Current (ISAT)
Saturation current is the DC bias current at which the inductance falls to a specified percentage of its zero-bias value—typically 70% or 80%. When the magnetic core saturates, its permeability collapses toward that of air, and the inductance drops sharply. In a switch-mode power supply, a saturated output inductor loses its ability to limit current ripple, and theresult can be destructive.
How saturation occurs depends on the core material. This is where ferrite and iron powder diverge in a way that shapes real design decisions.
3. Core Materials: Ferrite vs. Iron Powder
The choice of core material is the single most consequential decision in inductor selection. It determines saturation behaviour, AC losses, temperature stability, and size. The two dominant families—ferrite and iron/ alloy powder—are not simply "better" or "worse" versions of each other. They are different tools for different problems.
Figure 4: Representative saturation curves for ferrite and iron/alloy powder cores. Ferrite exhibits a sharp inductance collapse at saturation; powder cores show a gradual, predictable roll-off due to their distributed air gap. (Illustrative only.)
Figure 5: Dark grey ferrite toroidal core (left) versus brownish-yellow iron powder toroidal core (right). The colour difference reflects different material compositions and, consequently, different saturation characteristics, AC loss profiles, and permeability stability under DC bias.
Ferrite (typically MnZn or NiZn) has the lowest AC core losses of any practical magnetic material, making it the default choice for high-frequency transformers and low-power inductors up to several megahertz. Its saturation flux density is modest—roughly 0.3 to 0.5 Tesla at room temperature, declining with heat. The saturation itself is abrupt: inductance holds nearly constant right up to the knee, then collapses. Designers must leave substantial headroom to avoid this cliff, and must derate further for elevated-temperature operation.
Iron powder and alloy powder cores (Sendust/Kool Mµ, MPP, High Flux) saturate at 1.0 to 1.6 Tesla—two to five times higher than ferrite. More important than the absolute number is how they saturate. Each grain of metal powder is insulated from its neighbours, creating a distributed air gap throughout the core volume. As DC current rises, permeability rolls off gradually rather than collapsing at a single point. This gives designers two practical advantages: they can operate into the roll-off to reduce core size, and the circuit has better fault tolerance—a current surge produces only a soft inductance reduction rather than a catastrophic drop.
The trade-off: iron powder cores have higher AC losses than ferrite, particularly above a few hundred kilohertz. Pure iron powder is the lossiest; Sendust and MPP bring losses down but at higher cost. Magnetics Inc.'s design guide notes that Kool Mµ Ultra approaches ferrite-level losses while retaining the soft-saturation characteristic.
| Property | Ferrite (MnZn) | Iron Powder (Pure Fe) | Sendust / Kool Mµ | MPP (FeNiMo) |
|---|---|---|---|---|
| Bsat (T) | 0.3 – 0.5 | 1.0 – 1.6 | ~1.0 | ~0.7 |
| Saturation type | Sharp collapse | Soft roll-off | Soft roll-off | Soft roll-off |
| AC core losses | Very low | High | Low–moderate | Very low |
| Bsat vs. temperature | Declines with heat | Near-constant | Near-constant | Near-constant |
| Relative cost | Low | Low | Moderate | High |
| Best frequency range | 50 kHz – 80+ MHz (NiZn) | <400 kHz (inductive) | 10 kHz – 300 kHz | Wide range |
Then there are air-core inductors. No magnetic core means no saturation and no core losses—at any current, at any frequency. The price is low inductance density: an air-core coil needs far more turns (and therefore occupies more space) to achieve the same inductance as a cored equivalent. In RF circuits above roughly 50 MHz, where core losses become problematic, air-core coils remain the first choice. Classic texts such as the RCA Radiotron Designer's Handbook (RDH4) devote extensive sections to air-core inductance calculations and construction techniques [1].
On Temperature and Saturation
Ferrite's saturation flux density can drop 20–30% between 25°C and 100°C. A design that comfortably avoids saturation on the bench may fail in an enclosed chassis at full load. Iron powder and alloy powder cores are far less affected by this thermal derating effect. When selecting a ferrite-core inductor for a product that runs warm, check the saturation current rating at the worst-case ambient temperature, not at room temperature.
4. Types by Construction and Application
Inductors are classified as much by what they do as by how they are built. The construction determines the parasitic behaviour; the application determines which parameter matters most.
Wire-wound inductors are the general-purpose workhorse. A length of copper wire wound around a core or bobbin. They span the widest range of inductance and current ratings, from sub-microhenry SMD parts to multi-henry chassis-mount chokes. Their Q factors are typically the highest of any construction, making them the preferred choice for resonant circuits. The disadvantage is physical size and, in unshielded variants, susceptibility to magnetic coupling with neighbouring components.
Multilayer chip inductors use alternating layers of ferrite or ceramic and metal electrodes, fabricated using processes similar to MLCC capacitors. They are tiny—some in 0201 (0.6 × 0.3 mm) packages—and their small winding area limits both inductance and Q. But for high-speed digital decoupling, mobile RF front-ends, and any application where board area is the binding constraint, they are the practical default.
Toroidal inductors are wound on a doughnut-shaped core. The closed magnetic path means flux is almost entirely contained within the core, minimising radiated EMI and making the inductor largely self-shielding. This is why toroids appear in audio equipment, medical instruments, and precision measurement circuits—applications where magnetic coupling into nearby signal paths cannot be tolerated. Toroidal winding is, however, more labour-intensive than bobbin winding, which shows up in cost.
Common-mode chokes deserve separate mention. They suppress common-mode noise by exploiting the fact that common-mode currents produce additive flux in a shared core (high impedance), while differential-mode currents produce cancelling flux (low impedance). Nanocrystalline cores are increasingly used for broadband common-mode chokes, offering higher permeability than ferrite at low frequencies. According to Würth Elektronik application data, split-wound chokes provide useful leakage inductance that doubles as free differential-mode filtering, though at the cost of slightly reduced common-mode bandwidth.
Ferrite beads are essentially single-turn chokes that suppress high-frequency noise by converting it into heat rather than reflecting it back into the circuit. A through-wire ferrite bead (or sleeve) slipped over a wire or PCB trace adds virtually no board area. SMD ferrite beads, the more common choice on modern PCBs, occupy a small surface-mount footprint but remain compact and inexpensive—typically a fraction of a cent per unit. Both types are lossy by design and are widely used on power rails and signal lines for broadband EMI suppression.
Planar inductors, fabricated as spiral traces on PCB layers, are an increasingly common sight in high-density DC-DC converter modules. They eliminate a discrete component and its associated assembly cost, but their inductance is limited by the available board area and layer count. The core—typically a planar ferrite slab—adds height but significantly boosts inductance per turn.
5. A Practical Selection Workflow
No single inductor is optimal for every application. The selection process is a series of eliminations. The questions below are ordered by what usually rules out the most candidates fastest.
Decision Tree
1. What is the operating frequency?
< 1 kHz (mains) → Laminated silicon steel or iron core.
1 kHz – 1 MHz (switching power) → Ferrite, iron powder, or metal composite.
> 1 MHz (RF) → Air core, ceramic core, or thin-film chip inductor.
2. How much DC current?
High DC bias (amps) → Powder core preferred (soft saturation, high Bsat).
Low DC bias (mA) → Ferrite is fine.
3. What is the SRF requirement?
For predictable inductance: SRF ≥ 10 × Fop.
For narrowband RF chokes: SRF near Fop (impedance peak at operating frequency).
For broadband RF chokes: verify impedance across the full band; a single SRF peak may be insufficient.
4. Are size or profile constraints dominant?
Yes → Multilayer chip or planar inductor.
No → Wire-wound SMD or through-hole.
5. Is EMI a concern?
Yes → Toroidal (self-shielding) or magnetically shielded SMD.
No → Semi-shielded or unshielded; watch placement.
After narrowing the field by frequency, current, and package, the remaining candidates are compared on secondary parameters: DCR for efficiency, Q for resonant circuits, tolerance for filter alignment, and cost. Datasheet curves—inductance vs. frequency, inductance vs. DC bias, Q vs. frequency—are far more useful than single-point specifications. An inductor rated at 10 µH might only deliver 10 µH under very specific conditions.
Figure 6: A first-pass decision flow for inductor core material selection, based on operating frequency. Subsequent narrowing by DC bias, SRF, size, and EMI requirements follows.
6. Common Pitfalls
Some failures are more common than others. These are the ones worth checking first.
Operating above SRF. The inductor becomes a capacitor. In a DC-DC converter output filter, the consequences can range from excessive ripple to instability. In an RF matching network, the impedance transformation goes wrong. Always check the SRF against the highest frequency component present in the circuit—not just the fundamental, but harmonics too. This is especially important in Class D and Class E amplifiers, where the switching waveform contains significant harmonic energy.
Ignoring DC bias derating. A 10 µH ferrite inductor rated for 3 A might measure 10 µH at 0 A but only 6 µH at 2.5 A of DC bias. If your converter was designed around 10 µH, the ripple current will be 67% higher than expected, potentially tripping overcurrent protection or overheating output capacitors. Manufacturer-provided L vs. IDC curves are essential here—do not rely on the single-point specification.
Magnetic coupling between adjacent inductors. Unshielded inductors placed too close together on a PCB will couple magnetically, creating unintended mutual inductance. Magnetic coupling falls rapidly with distance, but the exact relationship depends on coil geometry, orientation, shielding, and near-field conditions. In many practical PCB layouts, even a few millimetres of spacing and a 90-degree rotation between neighbouring inductors can materially reduce coupling. In a multi-phase buck converter, coupled inductors can actually be beneficial (reducing ripple), but in unrelated circuits, the crosstalk is purely harmful. Use shielded parts or maintain separation. As Morgan Jones notes in Valve Amplifiers, even an inch of separation between power supply chokes can reduce coupling by 20 dB or more [2].
Assuming Q is constant. An inductor with Q = 80 at 10 MHz may have Q = 15 at 100 MHz. If your filter depends on a specific Q for its shape factor, verify the value at the operating frequency, not at the datasheet's test frequency.
Core losses at high ripple current. In a boost converter operating in discontinuous conduction mode, the AC ripple current can be as large as the DC component. Total losses are then I2DC × DCR plus core losses from the AC flux swing. The latter is not negligible above roughly 100 kHz for iron powder cores and should be estimated from the manufacturer's core-loss curves (typically given in mW/cm³ as a function of flux density and frequency).
7. Frequently Asked Questions
What is the difference between an inductor and a choke?
The terms overlap substantially. A choke is an inductor specifically designed to block AC while passing DC. The name comes from its function—it "chokes off" the alternating component. All chokes are inductors; not all inductors are used as chokes. The distinction is functional rather than structural.
Can I use a ferrite bead instead of an inductor for power supply filtering?
Sometimes. Ferrite beads are lossy—they dissipate high-frequency energy as heat rather than storing it in a magnetic field. This makes them effective at suppressing narrowband EMI without creating the resonant peaking that an LC filter can produce. But for ripple smoothing at the switching frequency (typically 100 kHz to 2 MHz), the bead's impedance is usually too low to be useful. Different tools for different frequency ranges.
Why does my inductor get hot?
The most likely cause is I²R loss in the winding (DCR × I2RMS). But core losses become significant at high frequency and high AC flux swing. If the inductor runs hotter than the DCR calculation predicts, check the manufacturer's core-loss data for your operating conditions. A third possibility, less common but worth checking, is that an adjacent component is coupling heat into the inductor through the PCB or shared heatsink.
What happens if I exceed the saturation current?
Inductance drops. In a ferrite-core part, the drop is sudden and large. In a converter, this means ripple current spikes, which can saturate the inductor further in a runaway loop. In a powder-core part, the drop is gradual, giving the circuit some grace. Either way, the component is operating outside its specified range, and the design should be revisited. The 30% inductance-drop threshold used by most manufacturers for the ISAT rating is a convention, not a physical limit—some circuits can tolerate more droop, others less.
Do inductors have a polarity?
Individual inductors do not—they are symmetric two-terminal devices. However, coupled inductors and transformers do have polarity, indicated by dot convention on the schematic. And in layouts, the physical orientation of an unshielded inductor matters because its external flux pattern is not symmetric about the winding axis.
How do I measure inductance without an LCR meter?
There are methods, but they carry caveats. You can build a resonant circuit with a known capacitor and measure the resonant frequency with an oscilloscope and signal generator, then calculate L = 1/(4π²f²C). This is the method described in the ARRL Handbook [3]. It works well for air-core and high-Q inductors. For power inductors with low Q and significant DCR, the resonance will be broad and the measurement less precise. It also gives you the effective inductance only at the resonant frequency, which may differ from the value at your actual operating frequency. A proper LCR meter or impedance analyser remains the tool of choice for serious work.
Find More
- IWISTAO Hollow Inductor Coil Oxygen-Free Copper Subwoofer Crossover HIFI Audio customization1.2MH to 64MH
- IWISTAO Dedicated Inductor for Crossover or Treble Unit Oxygen-free Copper Enameled Wire 0.13 0.18mh
- IWISTAO Dedicated Inductor for Crossover or Treble Unit Oxygen-free Copper Enameled Wire 0.4-1.2mh
References
The following references include classic engineering texts alongside manufacturer application notes and industry resources. Where community forums are cited, they are offered as supplementary rather than primary engineering references.
- Langford-Smith, F. (ed.). Radiotron Designer's Handbook (RDH4), 4th Edition. RCA Manufacturing Company, 1953. Chapters 7 and 10 cover inductance calculations and coil design in detail. Widely regarded as the definitive pre-transistor reference for inductive component design.
- Jones, M. Valve Amplifiers, 4th Edition. Newnes, 2012. Chapter 4 provides practical guidance on choke placement, magnetic coupling, and power supply filtering in audio-frequency circuits.
- ARRL. The ARRL Handbook for Radio Communications. American Radio Relay League, annual editions. Covers practical inductor construction, measurement, and RF choke design.
- Coilcraft. Doc671 — Selecting RF Inductors. Coilcraft Inc. Application note covering Q factor, SRF guidelines, and the 10× SRF rule of thumb. Available at: coilcraft.com
- Magnetics Inc. Inductor Core Material and Shape Choices. Magnetics design guide comparing ferrite, powder, and alloy core materials. Available at: mag-inc.com
- Würth Elektronik. RF Inductors in High-Frequency Design: Selection, Trends, and Challenges. Application note addressing SRF, parasitic capacitance, and high-frequency behaviour. Available at: we-online.com
- Sagami Elec. Main Parameters of the Inductor — Tips for Coil Users. Technical note on DCR measurement, terminal resistance effects, and rated current definitions. Available at: sagami-elec.co.jp
- Williamson, D.T.N. "Design Considerations in High-Fidelity Amplifiers." Wireless World, 1947–1949 (series). Early but still-relevant discussion of choke-input power supply filters and output transformer design.
- Ridley Engineering. Choosing the Inductor for a Buck Converter. Design note covering saturation current, ripple current, and core loss estimation for switch-mode power supplies. Available at: ridleyengineering.com
- Texas Instruments. Practical EMI Considerations for Low-Power AC/DC Supplies: Common-Mode Choke Practicalities. Video application note on winding styles, core materials, and parasitic capacitance trade-offs. Available at: ti.com
0 comments