This article gives an example of how profound proximity effects can be.

Introduction

In this article, Dr. Ridley discusses one of the most difficult aspects of magnetics design and analysis: winding proximity loss. Without understanding proximity loss, there is a roadblock to reducing the size of the magnetics components.

The Size of Modern Magnectics

In the last ten years, tremendous strides have been made in semiconductors to increase the performance of high-frequency power supplies. New wide-bandgap technologies, and new packaging techniques have led to unprecedented levels of power density and converter efficiency. The future for power conversion has never been so exciting.

However, one component has stubbornly continued to defy proportional miniaturization—the magnetics. Invented in 1831, the power transformer remains a larger component than anticipated, and it does not seem to follow any version of Moore’s Law.

Figure 1: Modern Power Converter Size is Dominated by the Magnetic Components

Figure 1 shows modern power converters for different applications. A PFC circuit, LLC, and a point-of-load converter all feature prominent magnetics. In some cases, we find that the semiconductor devices no longer even need a heatsink and have disappeared from view underneath the circuit board. This emphasizes even more the need to work harder on the magnetics if we are to continue making progress with size reduction.

When starting a series of articles on magnetics design, it is customary to start at the beginning with design basics. Considerations of turns counts, saturation, core materials, and design optimization are usually at the forefront. You can learn more about some of these issues from ourmagnetics videos[1].

However, in this article, we are going to start with an advanced topic which causes severe problems for many designers. We will see in the example used in this article that conventional analysis of winding loss indicates that there will be just over 1 W of dissipation. However, detailed proximity loss analysis shows that, in reality, there isalmost 15 Wof power loss.

Forward Converter Design Example

Figure 2 shows an example forward converter running at 360 W. The converter runs at a switching frequency of 100 kHz. We will look at the detailed design of the transformer and show that it demonstrates excessive winding loss, despite initially good performance calculations.

Figure 2: Two-Switch Forward Converter for 360 W Output