An inductor stores energy in a magnetic field to resist changes in current. When selecting an inductor for a power supply or RF circuit, the core material dictates its saturation current, frequency response, and physical footprint. For high-current DC-DC switching (buck/boost converters), use shielded ferrite cores. For high-power RF and antenna matching, use powdered iron or air cores. For high-frequency, low-current signal filtering, ceramic or unshielded ferrite cores are the standard.

Choosing the wrong core doesn't just degrade efficiency; it can destroy your switching IC. This guide breaks down how to match core types to specific jobs, decode the cryptic markings on physical parts, and diagnose the exact failure modes you will encounter on the bench.

Core Types and Selection Criteria

The core material determines how much magnetic flux the inductor can store before it saturates, and how it behaves across different frequencies. Below is a direct comparison to help you choose the right type for your specific application.

Core Type Construction Tolerance Tempco (ppm/°C) Typical Use Case
Air Core Wire wound on non-magnetic form (or self-supporting) ±2% to ±5% +50 to +150 VHF/UHF RF circuits, high-frequency crossover networks, where zero core loss is critical.
Ferrite (Shielded) Wire wound on ferrite bobbin, enclosed in a magnetic shield ±10% to ±20% +1000 to +3000 DC-DC buck/boost converters, high-current power filtering. Prevents magnetic flux from coupling into nearby traces.
Ferrite (Unshielded) Wire wound on exposed ferrite bobbin (drum core) ±10% to ±20% +1000 to +3000 Low-cost power supplies, EMI suppression beads, low-current signal filtering where EMI is not a strict concern.
Powdered Iron Iron powder mixed with an insulating binder, pressed into a toroid ±5% to ±10% +50 to +500 Switch-mode power supplies (continuous conduction mode), high-power RF amplifiers, antenna tuners. Handles high DC bias well.
Ceramic Wire wound on a non-magnetic ceramic core ±2% to ±5% +100 to +300 High-frequency RF matching networks, cellular/WiFi antenna traces. Extremely low core loss at GHz frequencies.
Warning: Never use an unshielded drum-core inductor directly adjacent to a sensitive analog trace or a high-gain op-amp input. The stray magnetic flux will induce noise that looks like random high-frequency spikes on your oscilloscope. Always use shielded ferrite (like the Coilcraft MSS or XEL series) in mixed-signal layouts.

Decoding Physical Markings and Color Codes

Unlike resistors, inductor markings are notoriously inconsistent across manufacturers. However, two dominant standards exist for through-hole and surface-mount parts.

SMD Inductor Codes

Surface mount power inductors typically use a three-digit alphanumeric code where the letter 'R' acts as the decimal point, and the unit is always microhenries (µH).

  • 4R7 = 4.7 µH
  • 100 = 10 µH (The third digit is the multiplier: 10 × 10⁰ = 10 µH)
  • 331 = 330 µH (33 × 10¹ = 330 µH)
  • R47 = 0.47 µH

Note: High-frequency ceramic SMD inductors (like the Murata LQG series) often use a different scheme where the unit is nanohenries (nH) and the letter represents tolerance (e.g., 4N7S = 4.7nH ±0.3nH). Always check the manufacturer's datasheet for RF chip inductors.

Axial Color Bands

Through-hole axial inductors use a color band system nearly identical to resistors, but the base unit is microhenries (µH), not ohms. According to standard component identification guides, the bands read as follows:

  1. Band 1 & 2: Significant digits
  2. Band 3: Multiplier (number of zeros)
  3. Band 4: Tolerance (Silver = ±10%, Gold = ±5%)

Example: Brown (1), Black (0), Brown (×10), Silver (±10%) = 100 µH ±10%.

Bench Scenario: The Buck Converter Saturation Disaster

Theory is clean; the bench is messy. Here is a real-world scenario that illustrates why understanding saturation current ($I_{SAT}$) is more critical than just matching the inductance value.

The Setup

I was prototyping a 12V-to-5V buck converter using a classic LM2596 switching regulator to power a 2A servo array. The datasheet called for a 33 µH inductor. I dug through my parts bin and found a generic, unshielded axial choke labeled "33µH 3A". I soldered it in, confident the 3A rating gave me plenty of headroom over my 2A load.

The Numbers

The LM2596 operates at 150 kHz. The peak inductor current ($I_{PEAK}$) in continuous conduction mode is roughly $I_{OUT} + \frac{\Delta I_L}{2}$. For a 2A load with 30% ripple, $I_{PEAK}$ hits about 2.3A. My "3A" inductor should have been perfectly safe.

The Outcome

Upon applying the 2A load, the 5V output immediately drooped to 3.1V. The inductor became too hot to touch within 15 seconds (measured at 92°C with an IR camera). Probing the switching node (pin 2) with my oscilloscope revealed massive, 15V ringing spikes and a distorted, triangular current waveform that looked more like a sawtooth.

What Went Wrong

The "3A" printed on the generic inductor was its thermal RMS current rating ($I_{RMS}$), not its saturation current. The physical core was too small for the magnetic flux density at 2.3A peak. The core saturated, meaning its permeability dropped to near that of air. The inductance collapsed from 33 µH down to about 1 µH. With almost no inductance to limit the $di/dt$, the LM2596 experienced massive current spikes during the switch-on phase, triggering its internal thermal shutdown and causing severe EMI. I swapped the generic part for a Coilcraft MSS1210-333 (a shielded ferrite part with a guaranteed $I_{SAT}$ of 5.2A). The output stabilized at 4.98V, the inductor stayed at 38°C, and the switching node waveform was perfectly clean.

Visual Failure Modes and Diagnosis

Inductors rarely fail silently. When they do fail, they leave physical evidence. Here is what to look for when troubleshooting a dead board.

  • Melted Enamel / Thermal Runaway: Visual Symptom: The copper wire turns dark bronze or black, and the outer heat-shrink or epoxy coating is blistered. You will smell a distinct "burning sugar" odor (the polyurethane enamel burning off). Cause: Exceeding the $I_{RMS}$ rating, causing $I^2R$ (DCR) heating to melt the wire insulation, leading to shorted turns.
  • Ferrite Core Cracking: Visual Symptom: A hairline fracture running through the ferrite drum or shield, sometimes visible only under a 10x loupe. Cause: Severe thermal cycling. Ferrite is brittle; rapid heating and cooling from intermittent high-current loads causes mechanical stress fractures, which alters the magnetic gap and drops the inductance.
  • Pad Lift / Tombstoning (SMD): Visual Symptom: One side of the SMD inductor is lifted off the PCB pad, or the part is standing on its end. Cause: Uneven reflow heating during assembly, or mechanical shock (dropping the enclosure) combined with the heavy mass of a shielded power inductor.
  • Silent Saturation (No Visual Clue): Visual Symptom: The part looks pristine. Diagnosis: You cannot see saturation. You must measure the inductance using an LCR meter equipped with a DC bias current source. If the inductance drops by >20% at your operating DC current, the core is saturating.

Safe Substitution Rules for Missing Parts

When you are out of the exact BOM-specified inductor and need to keep a prototype moving, you can substitute parts, but you must follow a strict hierarchy of parameters. Do not just match the microhenry value and call it a day.

Rule of Thumb: In power circuits, $I_{SAT}$ and $I_{RMS}$ are more important than exact inductance. In RF circuits, exact inductance and Q-factor are more important than current ratings.
  1. Match or Exceed Current Ratings: Your substitute must have an $I_{SAT}$ greater than your peak switching current, and an $I_{RMS}$ greater than your maximum continuous DC load. If the datasheet only lists one current rating, assume it is the lower of the two and derate by 20%.
  2. Inductance Tolerance: For DC-DC converters, you can safely substitute a value within ±20% of the original (e.g., using a 47 µH instead of a 33 µH). A higher inductance will lower your ripple current but may degrade transient response. Do not substitute a lower value; it will increase ripple and risk core saturation.
  3. Check the DCR (DC Resistance): The substitute's DCR must be equal to or lower than the original. A higher DCR will cause excessive voltage drop and thermal heating.
  4. Shielded vs. Unshielded: You can always substitute a shielded inductor for an unshielded one. You should never substitute an unshielded part for a shielded one in a production design, as it will alter your EMI profile and potentially fail FCC/CE radiated emissions testing.
  5. Physical Footprint and Height: Power inductors scale physically with current. If your substitute has the same footprint but claims a 50% higher current rating, verify the manufacturer's testing conditions. Some budget manufacturers rate $I_{SAT}$ at a 40% inductance drop, while premium brands (like Würth or Coilcraft) rate it at a 20% drop.

By treating the inductor as a complex magnetic component rather than a simple wire coil, you eliminate the most common source of switching regulator instability and RF inefficiency. Always verify your $I_{SAT}$ margins on the bench with a current probe before committing a power design to a final PCB spin.