Throughout my 25 years in the computer-aided engineering industry, some of the smartest people I know have told me that you can’t use simulation to design planar magnetic transformers. Is it true? No! What they’re really saying is that there isn’t an effective way to simulate the devices to predict the full behavior — which includes electromagnetic losses, harmonic content, EM-thermal coupling and ultimately how the electromagnetic fields and temperature affect the circuit — *in a reasonable amount of time* for simulation to be an effective design tool.

But now we have changed all that! With recent breakthroughs in simulation technology and high-performance computing, you can now design, simulate and optimize planar magnetic components in dramatically shorter time frames without compromising model fidelity.

But let’s take a step back and provide some context. Planar magnetic components consisting of a ferrite magnetic core and numerous conductor/insulation layers have been used for many years in switched mode power supplies (SMPS). These devices power PCs, TVs, charge mobile devices and are critical components of automotive electronics and telecommunications systems. The main benefits of planar components over conventional components with round conductors are that they have low leakage inductance and have a lower profile that requires less vertical space for the printed circuit board. Additionally, but of no less importance, is that planar components offer reduced weight, superior thermal management, higher power density and repeatability in manufacturing. Thus, planar magnetic designs have been the choice of SMPS designers in their drive to miniaturize power supplies while achieving a higher power density and improved overall circuit behavior.

However, the design of these components is not a trivial task. As switching frequencies rise, the increasing skin and proximity effects and the corresponding self and mutual impedances make modeling a planar magnetic challenging. Modeling and simulation efforts have focused on estimating AC resistance and validating physical measurements for electrical parasitic elements using 2-D frequency-domain Finite Element Analysis (FEA). These approaches have helped compress the design cycle, but they rely on assumptions and simplifications that require additional build and test iterations. This adds to the cost and limits the optimization of the design.

Ideally, you should apply a 3-D transient electromagnetic analysis to the planar design to provide the full harmonic content, and couple this with a thermal analysis to properly predict the losses without tedious build and test iterations. This simulation method allows you to analyze the dynamic system, including frequency dependent and thermally dependent materials, the ferrite core, and induced eddy currents under a variety of conditions and employing various excitations, including a pulsed waveform.

The transient EM process involves solving many time steps in a sequential fashion to calculate eddy currents in time and space. The transient electromagnetic analysis requires that many time steps be computed, where each time step employs large number of unknowns (degrees of freedom) due to very complex topologies and eddy effects on various conductive parts. Therefore the process is slow, which limits the size of problem that can be solved — even with today’s fastest computers!

Further complicating the design process is the application of integrated magnetics, where the planar transformers and planar inductors are integrated into a single structure to reduce the component footprint. This has a major impact on the overall component design since the magnetic design must be adjusted to optimally fit thermal constraints whose location is typically enforced by the location of the planar component on the PCB. Engineers must perform a thermal simulation to accurately predict heat distribution due to loss dissipation.

For complete thermal cooling analysis, thermal conduction (due to physical contact among bodies), thermal convection (due to fluid, such as air or a liquid, flowing away from the source) and thermal radiation (due to the emission of electromagnetic waves) must be simulated to predict the rate of heat transfer. These are dependent on the temperatures of the entire system and the properties of the intervening medium through which the heat is transferred.

Therefore, it is a huge computational undertaking to characterize an electronic transformer or a planar magnetic component. It can take hours or even days to complete a 3-D transient EM field simulation. Historically, engineers have not considered this methodology as an effective design method although its use could generate detailed and accurate design data, including power loss as a source of thermal management.

Now, the game has changed! Recent breakthrough developments in simulation technology and high-performance computing from ANSYS now make it possible for you to design, simulate and optimize planar magnetic components without compromising simulation accuracy or building physical models. The most notable breakthrough technology is the introduction of the patent pending Time Decomposition Method (TDM) to 3-D transient EM analysis. This high-performance computing technique delivers breakthrough computational speed by solving the time-steps simultaneously, instead of sequentially. TDM is used along with specialized solver technology with excellent parallel scalability. The result is an order of magnitude increase in computational speed that leads to a significant increase in the size of simulation that can be solved.

This leapfrog technology delivers a phenomenal increase in computational speed, allowing the 3-D transient EM methods to become a viable design tool. Furthermore, the technology deploys our full multiphysics engineering portfolio to provide a comprehensive end-to-end seamless design flow. Front-end design synthesis software streamlines the model creation. Two-way coupled thermal analysis tools predict joule heating from the losses in the windings and allow cooling strategies to be evaluated. Finally, reduced order modeling techniques pioneered by ANSYS allow the finite element simulation to be used within the circuit simulation to predict and validate final operation.