Technical Article

Memory Inductors for Energy-Efficient Applications

June 25, 2023 by Alexander Gerfer

The advanced series of WE-MAPI inductors combine the best possible utilization of inductance and current carrying capacity with low intrinsic losses thanks to clever material selection and manufacturing technology.

This article is published by EE Power as part of an exclusive digital content partnership with Bodo’s Power Systems.

Energy-efficient devices are important building blocks for saving resources and protecting the environment. The more efficient the electronics, the longer the battery life in mobile devices, and in large industrial and server facilities with thousands of consumers, the energy requirements are noticeably reduced.

The basis for energy-efficient devices is significantly influenced by the power supply. Whereas linear regulators were once the most commonly used voltage regulators, switching power supplies are now predominantly found in modern power electronics circuits. The continuous reduction in processor voltages has played its part. A few years ago, switching frequencies of up to 300 kHz were still very common, but today modern switching regulators are mostly clocked at frequencies of 800 kHz and more. The switching losses, on the one hand, but also the losses of the storage choke, on the other hand, are important aspects in the design of switching power supplies.

New material mixtures from the group of iron alloys have further reduced the core material losses for high-current storage chokes. Based on this, the WE-MAPI series combines the best possible utilization of inductance and current carrying capacity with low intrinsic losses through clever material selection and manufacturing technology.

The developer of power supplies is supported by the REDEXPERT tool, which allows the DC and AC losses of storage chokes to be determined with an accuracy not previously available. This is done by a measurement-based method that allows the calculation of core losses considerably more accurately than would be possible with the Steinmetz formulas.

 

WE-MAPI Storage Choke Characteristics

In conventional coils, the enameled copper wire is usually wound around the core and soldered or welded to the terminal with a clip. Subsequently, the outer shield ring is mounted and bonded to the inner core and winding.

 

Figure 1. The WE-MAPI storage choke is characterized by its compact design. The coil wire is directly contacted at the connection pads. Image used courtesy of Bodo’s Power Systems [PDF]

 

With the WE-MAPI, the winding is contacted directly to the terminal pads of the component using a direct contact process without soldering or welding (Figure 1). By eliminating the need for a clip, the effective diameter could be increased, which means that fewer windings are required for the same inductance values. This is expressed in a significantly reduced direct current resistance (RDC) of the winding.

In the application, the start of the winding of the coil is usually connected to the switching node of the switching regulator–the component carries a marking for this purpose. This minimizes the spatial expansion of the “hot” switching node, and coupling effects from the outer part of the winding, which is at “quiescent” potential, are shielded (Figure 2).

 

Figure 2. Self-shielding winding and core design for improved EMC performance. Image used courtesy of Bodo’s Power Systems [PDF]

 

The core is made of an innovative iron alloy pressed around the winding. This gives the WE-MAPI high inductance values in a small package. The special design of the core also has a self-shielding effect. The core material itself is temperature stable with only low drift and soft saturation behavior. In addition, a protective layer is applied around the core to make the surfaces resistant to environmental influences.

The WE-MAPI is available in many different sizes, from 1.6² x 1.0 mm, up to 5.4² x 3.1 mm (Figure 3) and has been recently extended at the top end.

 

Figure 3. Overview of the available designs of the WE-MAPI storage choke. Latest versions are indicated by the red box. Image used courtesy of Bodo’s Power Systems [PDF]

 

Losses in Storage Reactors

The losses of a storage choke consist of core material losses and winding losses. The winding losses themselves can be divided into DC losses, significantly influenced by the DC resistance of the winding (equation 1) and the AC losses (RAC) of the winding, which result from the skin and proximity effect.

\[P=I^{2}\cdot R_{DC}\,\,\,(1)\]

There are several methods for determining the AC losses of the winding–e.g., the Dowell, Ferreira, or Nan/Sullivan method.

The importance of AC losses in modern switching regulators can be determined with a simple setup and measurement of the corresponding losses. As an example, a buck converter with an input voltage of 24 V is used. A voltage of 12 V at a current of 2 A is available at the output. The clock frequency is 500 kHz. In the comparison in Figure 4, a memory inductance with 2.2 µH of the WE-MAPI 4020 series was measured and compared with memory inductances of the same design. It can be clearly seen that for all coils measured, the AC losses exceed the DC losses.

 

Figure 4. Proportions of the AC and DC losses of a 2.2 µH coil in a buck converter with 24 V input voltage and 12 V output voltage and 2 A output current at a clock frequency of 500 kHz. Image used courtesy of Bodo’s Power Systems [PDF]

 

In switching regulators, the coil is one of the most important components. Therefore, accurate determination of losses and heating is a critical step in selecting the right component. To predict the heating, the AC losses must first be accurately determined.

One approach would be the Steinmetz models, whereby these offer an acceptable approximation, especially for sinusoidal excitations and a duty cycle of 50%. The Würth Elektronik model, on the other hand, provides more accurate results.

Würth Elektronik eiSos has developed its own model to precisely determine the complete AC losses in inductors. This model is based on empirical data obtained with a real-time application setup. Here, the first step is to divide the total inductor losses into AC and DC losses.

The empirical data is acquired using a DC/DC converter. A pulsating voltage is applied to the inductor, and the input power Pin and the output power Pout are measured. On this basis, Ploss = Pin - Pout is obtained, and the AC losses of the inductor PAC are separated. This process is measured for a wide variety of parameter settings - for example, variations in magnetic excursion, switching frequency, ripple current, etc. - and this empirical data is recorded. Using this empirical data, the model for calculating the AC losses is created (equation 2).

\[P_{AC}=f(\Delta I,freq,DC,kl,k2)\,\,\,(2)\]

Advantages of the Würth Elektronik AC loss model:

  • Empirical data is based on a DC/DC converter.
  • Accurate determination of losses for any given duty cycle
  • Accurate over a wide frequency range (10 kHz to 10 MHz)
  • Takes into account even the smallest changes in core material and winding structure
  • Valid for components with more than one material used
  • Accurate determination of losses of components with iron powder and metal alloys
  • Valid for any core design and winding structure
  • Also includes AC winding losses

 

Image used courtesy of Bodo’s Power Systems [PDF]

Figure 5. AC losses for the core materials MnZn and iron powder and a duty cycle of 33 % according to different Steinmetz models by simulation with REDEXPERT and real measured values. Image used courtesy of Bodo’s Power Systems [PDF]

 

Figure 6. Simulation for a buck converter in REDEXPERT with WE-MAPI components. Image used courtesy of Bodo’s Power Systems [PDF]

 

Würth Elektronik’s model has been extensively validated and compared with existing models and measured data. AC losses for various materials, such as WE superflux, iron powder, NiZn, MnZn, etc., were measured over wide duty cycle and frequency ranges and compared with theoretical models (Figure 5). The plots show the core losses determined using the Steinmetz Power Equation (Pst), Modified Steinmetz Equation (Pmse), and Generalized Steinmetz Equation (Pgse). “REDEXPERT” indicates the AC loss as calculated using the AC loss model from Würth Elektronik. “Real” is the measured AC loss.

 

Selecting WE-MAPI with REDEXPERT

WE-MAPI storage inductors, with their innovative core material and thoughtful design, are engineered for maximum performance in minimum space. For energy-efficient switching regulators, the best way to select the right WE-MAPI storage inductor is with REDEXPERT (Figure 6), Würth Elektronik’s online design tool. The world’s most accurate AC loss model is integrated here, which achieves very high accuracy over wide value ranges for the parameters frequency, ripple current and duty cycle.

 

This article originally appeared in Bodo’s Power Systems [PDF] magazine and is co-authored by Alexander Gerfer, Ranjith Bramanpalli, and Jochen Baier.