What Is a Planar Transformer? Definition, Construction, and How It Works

A planar transformer is a magnetic component whose primary and secondary windings are flat copper layers — etched traces on a multi-layer PCB or stamped lead frames — clamped between two halves of a low-profile ferrite core. It replaces the round magnet wire of a conventional transformer with planar copper, which is why it suits switching frequencies from roughly 100 kHz to several MHz.

A Quick Definition (and Where the Idea Came From)

Where a wire-wound transformer winds round magnet wire onto a bobbin, a planar transformer uses a different conductor geometry: the turns are wide, thin copper layers stacked inside a PCB, around a flat ferrite set split into two clamping halves.

The format went mainstream as switch-mode supplies climbed into the hundreds of kilohertz in the 1980s and 1990s; wide-band-gap devices (SiC and GaN) pushed converters higher still in the 2010s, renewing interest in it (KSG PCB). The bargain: tooling investment and a limited turns count, for higher usable frequency, lower profile, and more repeatable parasitics.

PCB Winding Construction

The winding is the PCB. A typical design uses a 4-, 6-, or 8-layer board with copper weights of 2 oz to 6 oz (roughly 70 µm to 210 µm); each layer carries one or a few turns, connected through vias to form the full primary and secondary. Because the board fabrication fixes the geometry, every unit comes out nearly identical.

Insulation between the layers is FR-4 prepreg, polyimide, or a higher-CTI laminate. That dielectric sets isolation, creepage, and clearance, so planar designs answer to the same safety framework as any switch-mode transformer: IEC 61558-2-16 for SMPS transformer safety and IEC 60664-1 for insulation coordination, which ties the laminate's comparative tracking index (CTI) to the required PCB spacing.

A second style replaces the PCB with stamped copper lead frames, favored where higher current per turn is needed. febetek manufactures planar transformers under a UL-recognized insulation system (UL E533808); the quality and certifications page describes that recognition at the company level.

Core Geometry (E, ER, ELP, EQ, RM)

Planar cores are short in the vertical axis and split so the PCB slides over the center post, the two halves closing the magnetic path above and below the board. Ferroxcube documents the common half-set combinations — E + E, E + PLT (a flat plate), and E/R + PLT/S — in its Planar E Cores application note.

Nomenclature varies by vendor: TDK/EPCOS uses the ELP family, spanning ELP 14 to ELP 102 across the full power range, alongside ER and EQ shapes (TDK Planar Cores). E and ELP shapes carry a rectangular center post that is simple to lay out on a PCB; ER and EQ shapes use a round center post that lets the copper follow a smoother path and lowers winding loss.

The core material is almost always a MnZn ferrite matched to the operating frequency. Ferroxcube rates 3C85 to about 200 kHz, 3F3 to about 500 kHz, and 3F4 to about 3 MHz; TDK's comparable grades include N87, N97, and PC95.

Skin Effect and Proximity Loss — the Real Reason Planar Exists

This is the physics that justifies the format. At high frequency, current crowds into a surface layer of depth δ (skin depth) rather than flowing uniformly. In copper, δ ≈ 66 / √f mm with f in hertz — about 0.21 mm at 100 kHz and 0.066 mm at 1 MHz. Copper thicker than a few δ carries little current internally, so a fat round wire wastes most of its cross-section above 100 kHz. Planar copper is naturally thin (70–210 µm), keeping it near the optimum of the ξ = h/δ ratio used to size such conductors (MDPI Electronics review).

The larger loss mechanism above 100 kHz is the proximity effect: the field from one layer drives eddy currents in its neighbors, and the losses compound through a stacked winding. Planar wins because the layers can be interleaved in a precise, repeatable order — primary, secondary, primary, secondary rather than all-primary then all-secondary — cancelling much of the field between them. This interleaving order, not merely thinner copper, is what cuts the loss (MDPI Energies review).

Thermal Management — Why Flat Beats Round

A planar transformer rejects heat better mainly because of its shape. The flat ferrite set has a higher surface-area-to-volume ratio than a stick-and-bobbin design, so the same dissipated power produces a smaller temperature rise — manufacturer data commonly cites about 50% lower thermal resistance than a wire-wound part. The PCB copper also acts as a heat spreader to nearby copper pours or a baseplate.

Design targets follow from the ferrite. Magnetics Inc. recommends keeping the ferrite hot spot near 100 °C for low-loss operation, with insulation systems typically limiting the part to about 130 °C (Magnetics Inc. FC-S8). The same bulletin flags a handling limit teams often miss: ferrite tolerates only about 5–10 °C per minute of temperature change before cracking risk rises, which matters during reflow and thermal-shock testing.

Why Choose Planar — When the Trade Pays Off

Choose planar when the operating point and the production volume both line up — the decision comes down to four factors:

• Switching frequency at or above 100 kHz — the inflection point. Below it, the skin- and proximity-effect savings shrink and wire-wound's flexibility wins.
• Annual volume in the tens of thousands — enough to amortize the fixed PCB and core tooling cost. Below that, wire-wound is cheaper per unit (passive-components.eu worked comparison).
• Repeatability — etched windings give a tight unit-to-unit leakage-inductance spread, where hand-wound parasitics scatter.
• EMI margin — predictable parasitics let you target CISPR 32 / FCC Part 15 limits deterministically instead of tuning each build.

Within their target band, planar designs are typically 1 to 3 percentage points more efficient than a wire-wound equivalent. To match a power tier to your design, browse our planar transformer series.

Where Planar Fails (Limitations)

Planar is not a universal upgrade. It loses in several common situations:

• High turns ratios. PCB window area limits how many turns you can stack; ratios of about 1:1 to 1:5 are routine, and pushing well beyond that forces impractical board area.
• High interwinding capacitance. Large overlapping layers store more energy between primary and secondary. A 2024 measurement showed a parasitic-capacitance resonance shifting from 1.27 MHz to 1.63 MHz with just 0.4 mm of added PCB thickness (Springer Journal of Power Electronics, 2024) — a real concern for LLC and CLLC topologies.
• Prototyping friction. Changing one turn means a new board revision — often a two-week loop, against hours to rewind a wire-wound prototype.
• Low-frequency designs. Below roughly 50 kHz there is little skin-effect benefit, so the tooling cost buys nothing.

Weighing a planar transformer against your current wire-wound part? Request a quote for a custom planar design and tell us your frequency, power, isolation, and profile targets.