As an embedded planar device, the transformer's ferrite core halves are physically attached to a system board that also carries the circuitry for, say, a dc/dc brick converter. The transformer is fashioned using some portion of the board's copper traces designed specifically for it. The standalone type, on the other hand, is basically a component in the same context as a conventional wirewound transformer. It uses the traces from a small printed-circuit board or other foil, and suitable layers of insulation, that are all sandwiched between the ferrite-core halves.

Low-profile benefits
In general, planar transformers, with their ideally flat windings stacked thinly, are touted for their low profile for today's telecom requirements. While low-profile structures are not necessarily confined to planar construction, the planar transformer offers other significant advantages over traditional types in many applications. "Skin effect and proximity effects of the windings at higher frequencies is what's driving planar development," said Geoff Wildman, senior product manager at Pulse. "It's one reason planars have been so widely received for both embedded/integrated core-on-board transformers and for component-type planars." The latter, he noted, enjoy a sweet spot at switching frequencies of about 250 kHz for 200 watts and up to 750 kHz at lower powers.

The nature of the transformer's printed-circuit layers is such that primary and secondary windings can more ideally be interleaved to reduce leakage inductance and high-frequency losses. Further, the leakage inductance can generally be restricted to more controlled, or known, values from lot to lot. "Planars not only minimize parasitic inductances, they allow the designer to use those parasitics to his advantage in resonant-type situations," said Terry Booker, application and design engineer for Transpower Technologies. "He'll be able to do that because he knows those values aren't going to vary much from lot to lot."

Another advantage is the planar design's high ratio of surface area to its volume, which helps out a lot in handling heat dissipation. Planars are not without their trade-offs, though. While the transformer's windings can be closely wound, thus reducing leakage inductance, such a situation tends to increase the parasitic capacitance. In addition, the planar transformer's inherent construction, especially when printed-circuit boards are used, and the nature of its cores are such that they restrict the number of turns.

Many designers say that restriction tends to limit the flexibility to optimize the primary-to-secondary turns ratio for a given design. As for the core designs, "frequency is up, current is up, and most of new applications are for low voltage," said Booker. "Core saturation becomes a problem with higher currents. In these applications, the cross-sectional area has to be high, but with smaller core sizes it can't be and that works against you. That's one of the trade-offs."