Leakage Inductance: Design Parameter or Defect, and How to Hold It

Leakage Inductance Is Either a Defect or a Feature

Leakage inductance is the flux that links one winding but not the other. In a distribution transformer it is waste: it causes voltage droop and reactive loss. In a resonant converter, a welding transformer, or a UV ballast, it is the circuit element that makes the topology work. The engineering question is therefore not “how do I minimize it” but “what value does my circuit need, and can I hold it in production.”

This note covers how leakage inductance arises, how it is approximated, and the construction levers that set it.

Where It Comes From

Leakage flux occupies the space between the primary and secondary windings, the winding-to-core gaps, and the ends of the winding where flux fringes. It scales roughly with the square of the turns and inversely with the winding’s effective height, and it rises sharply as the separation between windings increases.

A useful first approximation for a concentric pair, in SI units:

L_lk ≈ μ₀ · N² · π · (d_mean) · (a + b/3) / h

where N is turns, d_mean the mean turn diameter, a the winding separation, b the winding build, and h the winding height. The approximation is crude but it correctly shows the levers: separation in the numerator, height in the denominator, turns squared in front.

Construction Levers

Ways to set leakage inductance
Lever	Effect on leakage	Side effect
Interleaving (split primary)	Reduces sharply, often 3–5×	More layers, higher capacitance, higher cost
Winding separation / added gap	Increases proportionally	Worse coupling, higher DCR if sectioned
Taller, narrower winding	Reduces	Window utilization may suffer
Magnetic shunt	Increases in a controlled way	Ferroresonant behaviour, extra loss
Fewer turns with larger core	Reduces by turns²	Larger footprint and cost

Holding a Target Value in Production

When leakage inductance is a design parameter, its tolerance is as important as its nominal value. A resonant tank designed for 20 µH ±10 % cannot accept parts at 26 µH, because the tank’s operating point moves and the converter loses zero-voltage switching. Practical controls:

State the tolerance as a percentage with the measurement condition (frequency, winding configuration, shorted secondary).
Fix the winding build with a defined layer count and separator thickness; do not leave build to the operator.
Measure on 100 % of units or on a sampling basis with a documented test fixture.
Track the distribution, because a drifting mean signals a material or process change before units go out of tolerance.

When Leakage Is Waste

For a distribution or isolation transformer, the goal is to keep leakage low enough that voltage regulation and short-circuit impedance meet the standard. Interleaving is the main tool, balanced against the capacitance it adds, which matters if the transformer also carries high dV/dt. A common compromise: split the primary into two halves with a single secondary between them, which typically cuts leakage by 60–70 % at modest cost.

What to Put in the Drawing

Leakage inductance nominal value and tolerance, with the test method (secondary shorted, measurement frequency).
Winding arrangement (layer order and separators), stated as a controlled feature.
Winding height and build limits.
Whether a shunt is used, and its geometry tolerance.
Acceptance: 100 % test or AQL sampling, and the fixture description.