The fundamental formula mutual inductance relies on is M = k × √(L1 × L2). This equation quantifies the magnetic linkage between two discrete coils, dictating how much voltage is induced in a secondary winding when current changes in the primary. Whether you are designing a flyback converter, a SEPIC power supply, or an audio crossover network, calculating $M$ accurately prevents catastrophic voltage spikes and ensures proper energy transfer.

The Core Formula and Symbol Definitions

To use the formula mutual inductance calculations require, you must understand the physical boundaries of the variables. The primary design equation is:

M = k × √(L1 × L2)

Symbol Parameter Standard Unit Physical Meaning & Constraints
M Mutual Inductance Henries (H) The proportionality constant between the rate of current change in one coil and the induced voltage in the other.
k Coupling Coefficient Dimensionless A ratio from 0 (no linkage) to 1 (perfect linkage). Dictated by core geometry, air gaps, and winding proximity.
L1 Primary Self-Inductance Henries (H) The inductance of the first coil measured in isolation, assuming the second coil is open-circuited.
L2 Secondary Self-Inductance Henries (H) The inductance of the second coil measured in isolation, assuming the first coil is open-circuited.

Assumptions and Realistic Magnitudes

This formula applies under three strict assumptions: the magnetic core must be operating in its linear region (below saturation), the coils must be physically stationary (no motional EMF), and the permeability of the medium must remain constant during the current transient. If your core saturates, $L_1$ and $L_2$ collapse, rendering the calculated $M$ invalid.

For context on realistic answer magnitudes: RF coupling networks typically yield $M$ values in the 10 nH to 500 nH range. Audio and intermediate-frequency transformers sit in the 1 mH to 100 mH range. Mains-frequency (50/60 Hz) power transformers exhibit massive mutual inductance, often between 1 H and 10 H, due to their high-permeability silicon steel cores and tight interleaved winding structures (Georgia State University HyperPhysics).

Rearranged Forms for Bench and Design Work

On the workbench, you rarely solve for $M$ directly from physical dimensions. Instead, you measure self-inductances and derive the coupling coefficient, or you use the transient voltage formula to size snubber networks. Here are the critical rearranged forms:

  • Solving for Coupling (k): k = M / √(L1 × L2) (Used when characterizing an unknown transformer on an LCR meter).
  • Solving for Primary Inductance (L1): L1 = M2 / (k2 × L2)
  • Solving for Secondary Inductance (L2): L2 = M2 / (k2 × L1)
  • Transient Voltage Formula: V2 = M × (di1 / dt) (Defines the voltage induced in coil 2 based on the current slew rate in coil 1).
  • Solving for Slew Rate (di/dt): di1 / dt = V2 / M (Used to determine how fast a MOSFET must switch to achieve a target flyback voltage).
Bench Trick for Finding M: If you don't have a dedicated mutual inductance meter, measure $L_1$ and $L_2$ individually. Then, wire the coils in series-opposing (where their magnetic fields cancel) and measure the total inductance $L_{total}$. The mutual inductance is exactly M = (L1 + L2 - Ltotal) / 4. This empirical method bypasses the need to guess $k$ (All About Circuits).

Worked Examples with Strict Unit Tracking

Abstract formulas fail on the bench when unit prefixes are ignored. The following problems track every conversion to base SI units (Henries, Amperes, Seconds, Volts) to prevent calculation errors.

Problem 1: Calculating M for a SEPIC Coupled Inductor

Scenario: You are designing a Single-Ended Primary-Inductor Converter (SEPIC). Your chosen coupled inductor has a primary self-inductance $L_1$ of 4.7 μH, a secondary self-inductance $L_2$ of 12 μH, and a datasheet-specified coupling coefficient $k$ of 0.92. Find $M$.

  1. Convert to base units:
    L1 = 4.7 × 10-6 H
    L2 = 12.0 × 10-6 H
  2. Multiply the inductances:
    L1 × L2 = (4.7 × 10-6) × (12.0 × 10-6) = 56.4 × 10-12 H2
  3. Take the square root:
    √(56.4 × 10-12) = 7.5099 × 10-6 H
  4. Multiply by k:
    M = 0.92 × 7.5099 × 10-6 H = 6.909 × 10-6 H
  5. Final Answer: M = 6.91 μH

Problem 2: Predicting the Flyback Voltage Spike

Scenario: A relay driver circuit uses a coil with a mutual inductance to a nearby sensor trace of $M$ = 2.5 mH. The relay coil is carrying 150 mA. When the driving BJT turns off, the current collapses from 150 mA to 0 A in 50 nanoseconds. What voltage is induced in the sensor trace?

  1. Convert to base units:
    M = 2.5 × 10-3 H
    di = 0.150 A - 0 A = 0.150 A
    dt = 50 × 10-9 s
  2. Calculate the slew rate (di/dt):
    di / dt = 0.150 A / (50 × 10-9 s) = 3,000,000 A/s (or 3 × 106 A/s)
  3. Apply the transient formula:
    V2 = M × (di / dt)
    V2 = (2.5 × 10-3 H) × (3 × 106 A/s)
  4. Final Answer: V2 = 7,500 Volts.

Takeaway: Even a tiny 2.5 mH coupling and a modest 150 mA current can induce a 7.5 kV spike if the switching edge is fast enough. This is why physical separation (lowering $k$) and snubber diodes are non-negotiable in mixed-signal PCB layouts.

Unit Mistakes That Will Break Your Calculations

When the formula mutual inductance math goes wrong, it is almost always due to one of three prefix errors. Avoid these traps:

  • The Microhenry Trap: Multiplying two μH values yields pico-Henries (10-12), not micro-Henries. If you forget to adjust the exponent before taking the square root, your calculated $M$ will be off by a factor of one million.
  • The Timebase Trap in di/dt: Oscilloscope cursors often read time in microseconds (μs) or nanoseconds (ns). If you plug '50' into the $dt$ denominator without converting it to $50 × 10^{-9}$ seconds, your predicted voltage spike will be a billion times smaller than reality, leading to blown MOSFETs on the bench.
  • The Turn-Ratio Confusion: Remember that $L$ scales with the square of the turns ratio ($N^2$), but $M$ scales linearly with the turns ratio if you are reflecting impedances. Do not mix up the geometric mean formula ($M = k√(L_1 L_2)$) with the ideal transformer voltage ratio ($V_1/V_2 = N_1/N_2$). The former accounts for leakage flux via $k$; the latter assumes $k=1$.

Decision Path: Selecting a Coupled Inductor for Power Conversion

Calculating $M$ is only half the battle; sourcing a component that maintains that $M$ under load without saturating is the actual engineering challenge. Use this decision tree to select the right coupled inductor topology for your DC-DC converter.

Operating Condition Core Material & Topology Expected k Value Concrete Part Recommendation
High Frequency (>500 kHz), Low Current (<2A), Tight Space Shielded Ferrite SMD (Drum core with sleeve) 0.85 - 0.95 Bourns SRP1265A Series
Low Frequency (<100 kHz), High Current (>5A), High Ripple Iron Powder or Gapped Ferrite E-Core (Unshielded) 0.90 - 0.98 Coilcraft MSD1560 Series
Isolated Flyback, Wide Input Range, Needs High Leakage for ZVS Gapped Ferrite with split bobbins (intentionally low k) 0.60 - 0.80 Wurth Elektronik 750313735
Standard SEPIC, 12V to 5V, Moderate Current (2A-4A), General Prototyping Ferrite Drum Core, Dual Winding, High Saturation Margin 0.95+ DEFAULT PICK: Coilcraft MSD1278-473

For 90% of hobbyist and mid-level engineering SEPIC or non-isolated coupled-inductor projects operating between 200 kHz and 500 kHz, the Coilcraft MSD1278-473 is the definitive starting point. It provides 47 μH per winding, handles up to 2.5A RMS, and guarantees a coupling coefficient tight enough to minimize leakage-induced ringing on your switch node, while remaining readily available from distributors like Digi-Key and Mouser (Coilcraft Coupled Inductors). Stop guessing your core geometry; start with the MSD1278-473, measure the empirical $M$ on your bench, and tune your compensation network from there.