Power architectures with nonisolated voltage regulation continue to evolve, and power inductor designs are fundamental to the success of the new product designs. The trend toward two-stage conversion with nonisolated point-of-load (POL) modules is fueling the demand for low-profile, high-power surface-mount inductors with current ratings up to 25 A. Power system design engineers have a wide variety of power inductors at 8 A and higher to choose from when working on a new design. This can be rather overwhelming for a new designer because datasheets can be confusing, and an experienced designer may not be aware of some of the newer core materials available such as powdered iron and ferrite materials. These materials can help reduce size, cost and power losses.
The Right Material for the Right Job
As power densities and current levels continue to increase and more competitors enter the market, product performance becomes critical for successful product sales. For the magnetics design engineer seeking to minimize power losses, the core material selection and physical package size are critical. The job becomes even more difficult as the size of the device decreases and the current increases.
The Right Material for the Right Job
The most common core materials are listed in Table 1. These materials have been around for a long time, but manufacturers are continuously developing new ones. The basic core material ingredient is iron, but it is the alloy blends or oxide formulations and the process controls that give materials their unique performance characteristics. Even though there are many categories, the majority of the materials are considered powdered irons. Categories 1 through 5 in Table 1 outline powdered irons.
Each material category occupies a niche where it is the best material for a given design application. If cost is the critical factor in the design application, then powdered iron is the traditional choice. Table 2 depicts a core material cost comparison using standard powdered iron as the benchmark.
The cost multiplier has been declining in recent years for the ferrite, powdered alloy and high flux materials, because more vendors are developing new materials and selling this type of product. The molypermalloy (MPP) material is cost-restrictive for most applications; however, it is often used for low-volume, height-restrictive applications where a toroidal shape is required. MPP manufacturers are starting to expand their MPP offerings into EIR core shapes, but this product will probably remain too expensive for most high-volume, high-current inductor designs.
For power module designs, the inductor's saturation level is critical. Saturation level is typically defined as the dc current level at which the device's inductance declines to 75% to 80% of its nominal inductance. The dc current can saturate the inductor quickly if an air gap is not introduced into the core's magnetic path. The powdered iron materials have an inherent air gap that is distributed throughout the core, which gives them a soft saturation curve. Ferrite material must have an air gap physically inserted or ground between the mating surfaces of the core halves. The saturation curve is steeper and saturation is abrupt. Typical material saturation levels, or maximum flux density levels, are shown in Table 3.
If the most critical design parameter is current saturation level and the customer's pc board size has either height- or board-area constraints, a high-flux material or a powdered-alloy is a good choice. Ferrite-core materials have typically struggled in this area because their maximum flux density levels are about one-half to one-third the level of powdered-iron materials. In addition, the ferrite material's saturation level will decrease at higher operating temperatures, while the powdered-iron-alloy saturation level does not decrease at higher operating temperatures. This consideration is often overlooked in the customer's product specification. If ambient temperatures are 70°C maximum and the inductor has a maximum temperature rise of 40°C, then the inductor must be able to operate above 100°C. Many inductor datasheets list saturation current ratings at room temperature conditions only; often the customer must specifically request current saturation level rating information.
A New Ferrite Material
Within the last five years, manufacturers have introduced a new type of ferrite material that has increased the maximum flux density level from a previous high of 5000 G to 5800 G at 25°C, and from 3800 G to 4800 G at 100°C. The material is often referred to as a high-BSAT material. An example of how this material can help is shown in Fig. 1.
Two devices were tested for inductance versus current at 130°C. Both inductors were constructed with an EIR12 core and a four-turn coil with a nominal inductance of 1.2 µH. The only difference between the parts tested was the core material. One used a high BSAT material and the other used a standard ferrite-powder material. The difference in saturation levels is approximately 4 A, which represents a significant improvement in performance.
Comparing the high-BSAT ferrite material to the high-flux material, the latter material's saturation performance is still much better. Given 1.2-µH inductor designs with an EIR9.5 core size, the high BSAT ferrite-core material will saturate at 18 A versus the high-flux material saturating at 22 A. Fig. 2 shows inductance versus current at 25°C.
The equivalent saturation performance in a ferrite part would require the core package to be 50% larger than the high-flux design. At 100°C, the difference in saturation levels is even more pronounced.
High-performance voltage regulators require a fast transient response, which has increased switching frequencies to 350 kHz and higher. Following this trend, the inductor's inductance value has decreased to 0.50 µH to 1.5 µH, with peak currents pushing to 30 A and ripple currents rising to 30% and higher. Under these conditions, core losses are becoming a significant factor in the selection of the inductor. Most magnetics design guides recommend that the core-to-copper loss distribution be at least 50-50, but 20-80 is actually preferred. It is easier to dissipate the heat from the winding versus the core material, as copper has a higher thermal conductivity than either ferrite or powdered iron. Core losses are a function of frequency and flux swing, and are defined in the following formula:
PV = k × (ƒx) × (By),
where PV is the core loss density (kW/m3), f is the frequency (Hz), B is the change in flux density in Gauss (G), and k, x and y are constants derived from actual test data. Ferrite materials have a higher resistivity than powdered-iron materials, which permits them to operate more efficiently at higher frequencies. Core loss will vary with the swing in flux density, frequency and temperature. No two ferrite materials or powdered iron materials are alike when it comes to core-loss performance. In general, the less expensive the material, the higher the core losses. Most empirical core-loss data is derived using a sinusoidal waveform, which makes the testing setup straightforward. Fig. 3 shows a comparison of core losses per volume for various materials.
The tolerance of this data is typically ±15%. This data is based on the core manufacturer's standard material characteristics and is taken at a flux density of 100 G at 25°C.
At switching frequencies of 300 kHz and higher, many power modules use ferrite-core inductor designs, requiring significant pc board area to accommodate the inductor. However, there are new powdered-iron materials that perform well with relatively low core loss levels up to 500 kHz and higher. The 60 perm-powdered alloy shown in Fig. 3 is close in performance to a MPP material. It is difficult to compare core losses in various types of materials because the flux-density level is inversely related to the turns and core area. These parameters can vary because the material permeability levels are so different, especially between a ferrite design and a powdered-iron design. With that said, a benchmark study was conducted to determine the difference in ac power losses (coil plus core) for several powdered-iron core inductor designs versus a high BSAT ferrite-core design. Included in this comparison was a molded powdered-alloy inductor. The core in this inductor is pressed around the coil instead of assembling two core halves. The design parameters for the inductor were as follows: 1 µH nominal at 0 Adc, peak current equals 16 A, ripple current equals 6 APK-PK. The various cores were assembled with flat ribbon wire coils and the finished package sizes are listed in Table 4.
Fig. 4 shows the power losses that were measured on a Clarke-Hesse V-A-W meter. The ferrite design has the lowest power losses. However, the package size is much larger than the powdered-iron options.
If board area is critical, a powdered-iron alloy may be a good choice. Conversely, if height is critical, a molded powdered alloy may be a good choice. However, either will not significantly increase the ac power losses at switching frequencies up to 500 kHz. Finally, pricing is comparable between all of these options.
Given that most engineers designing voltage regulators are looking for smaller size in both footprint and height, as well as higher current-handling capability, the new ferrite and powdered-iron materials will provide new solutions to meet this challenge.