Showing papers by "Khai D. T. Ngo published in 2021"

A Double-Side Cooled SiC MOSFET Power Module With Sintered-Silver Interposers: I-Design, Simulation, Fabrication, and Performance Characterization

Abstract: Planar, double-side cooled power modules are emerging in electric-drive inverters because of their low profile, better heat extraction, and lower package parasitic inductances However, there is still a concern about their reliability due to the rigid interconnection between the device chips and two substrates of the power module In this article, a porous interposer made of low-temperature sintered silver is introduced to reduce the thermomechanical stresses in the module A double-side cooled half-bridge module consisting of two 1200 V, 149 A SiC MOSFETs was designed, fabricated, and characterized By using the sintered-Ag instead of solid copper interposers, our simulation results showed that, at a total power loss of 200 W, the thermomechanical stress at the most vulnerable interfaces (interposer-attach layer) was reduced by 42% and in the SiC MOSFET by 50% with a tradeoff of only 36% increase in junction temperature The sintered-Ag interposers were readily fabricated into the desired dimensions without postmachining and did not require any surface finishing for die bonding and substrate interconnection by silver sintering The porous interposers were also deformable under a low force or pressure, which helped to accommodate chip thickness and/or substrate-to-substrate gap variations in the planar module structure, thus simplifying module fabrication The experimental results on the electrical performance of the fabricated SiC modules validated the success of using the porous silver interposers for fabricating planar, double-side cooled power modules

Series-Resonator Buck Converter—Viability Demonstration

Abstract: The series-capacitor buck converter doubles the duty ratio, and equalizes the current between two phases A series-resonator buck converter is realized by adding a resonant tank in series with the series capacitor Cs All switches turn on at zero voltage, and the low-side switches turn off at zero current The resonant tank generates additional loss and increases the voltage stress of the low-side switches A 2-MHz prototype with a peak efficiency of 985%, 48 V at the input and 7 V, 20 A at the output was built to demonstrate the viability of the topology

Degradation of SiC MOSFETs under High-Bias Switching Events

Abstract: Evaluating the robustness of power semiconductor devices is key for their adoption into power electronics applications. Recent static acceleration tests have revealed that SiC MOSFETs can safely operate for thousands of hours at a blocking voltage higher than the rated voltage and near the avalanche boundary. This work evaluates the high-bias robustness of SiC MOSFETs under continuous, hard-switched, turn-off stresses with a dc-bias higher than the device rated voltage. Under this high-bias switching condition, SiC MOSFETs show degradation in merely tens of hours at 25°C and tens of minutes at 100°C. Two independent degradation and failure mechanisms are unveiled, one present in the gate-oxide and the other in the bulk-semiconductor regions, featured by the increase in gate leakage current and drain leakage current, respectively. The second degradation mechanism has not been previously reported in the literature; it is found to be related to the electron hopping along the defects in semiconductors generated in the switching tests. The comparison with the static acceleration tests reveals that both degradation mechanisms correlate to the high-bias switching transients rather than the high-bias blocking states. These results suggest the insufficient robustness of SiC MOSFETs under high-bias, hard switching conditions and the significance of using switching-based tests to evaluate the device robustness.

Synthesis of DC–DC Converters From Voltage Conversion Ratio and Prescribed Requirements

Abstract: This article proposes a systematic synthesis method to identify the optimum converter topology for a specified voltage conversion ratio. This gain-oriented converter synthesis is formulated as an inverse problem around the inductor flux balance equations. The proposed synthesis method consists of two processes. A set of distinct converter topologies is first synthesized from the specified voltage conversion ratio by obtaining the solutions for a set of six inverse equations in twelve unknowns. The optimal topology for any intended application can be identified based on the desired performance metrics. The complete synthesis method is illustrated with quadratic buck–boost gain as an example. In total, 25 distinct topologies are synthesized using the first process, and their feasibility is validated by PLECS simulation. In total, three case studies are reported to explain the second process. Ground-referenced output and the minimum number of switches are used as the first two selection criteria in all the case studies. Minimum peak inductor current, minimum component stress factor, and minimum voltage and current stress are the third selection criterion in Case Studies-1, 2, and 3, respectively. The steady-state operations of two optimal converters selected in Case Study-1 were verified experimentally.

High-Power Magnetoelectric Voltage Tunable Inductors

Abstract: Magnetoelectric voltage tunable inductors (VTIs) offer a new paradigm for power electronics circuit design. Here, air-gapped VTIs are demonstrated exhibiting a large inductance tunability of 220% with applied fields under 20 kVcm−1 and operational stability up to 5 MHz that covers the full frequency range of the state-of-the-art power electronics. The design of air-gapped VTI comprises of two C-shaped ferrite cores, one permeability-variable magnetic flux valve (MFV) in between C-shaped cores, and slotted air gap between cores and MFV. The tunability of VTIs is achieved through the electric field modulation of permeability in the MFV based on the magnetoelectric effect. There is a tradeoff between the tunability and saturation current. The introduction of air gap significantly reduces the tunability but provides methodology toward increasing the saturation current and power handling capability due to the increased reluctance. The inductance and tunability reduces by 50% with increase in the air-gap width from 0.07 mm to infinity (condition with no C-shaped cores). Phase field simulations demonstrate that the air gap affects the magnetization, permeability, and tunability of VTIs by tailoring the demagnetization field that further influences the magnetic domain rotation process. VTIs with increased frequency range and saturation current will strengthen the continued development of tunable power electronics.

Two-Dimensional Mapping of Interface Thermal Resistance by Transient Thermal Measurement

Abstract: Bonded interfaces in power converters add thermal resistances to heat dissipation. Under cyclic power, temperature, or chemical loading, these interfaces degrade, raising the thermal resistances. Reliability of the thermal interfaces is especially problematic when the bonded area is large because the larger the area the more likely it is to have preexisting defects from processing. To help qualifying the development of a bonding process and quantifying the interface reliability, it would be desirable to have a simple, reliable, and nondestructive measurement technique to obtain a 2-D map of the interface thermal resistance across a large bonded area. Based on the transient thermal method of JEDEC standard 51-14, in this article, we develop a measurement technique that involves moving a thermal probe discretely across a large-area bonded substrate and acquiring the thermal interface resistance under the probe at each location. The probe is made by custom-packaging an insulated-gate bipolar transistor (IGBT) power device. An analytical thermal model is developed to gain insights into the effects of probe materials and structural parameters on the sensitivity of the measurement technique. To obtain a 2-D thermal resistance map of a bonded substrate, the probe is thermally coupled to the substrate at one location through a thermal pad or grease; the device is powered up to a steady-state junction temperature; the power is cutoff; and then the junction temperature during cool-down is recorded. The recorded temperature data are analyzed to derive a thermal structure function of the multilayer material stack. The process is repeated at other locations until a 2-D map of the interface thermal resistance across the entire substrate is completed. This technique is demonstrated on copper–copper bonded samples using either a thermal grease or sintered silver. The resolution of the 2-D mapping technique is evaluated by a copper–grease–copper stack with defects implanted at the bond line.

Packaging of an 8-kV Silicon Carbide Diode Module with Double-Side Cooling and Sintered-Silver Joints

Abstract: Packaging innovations are needed for medium-voltage wide bandgap power semiconductor modules to enable their adaptation in grid applications. A unique challenge for packaging medium-voltage power modules is managing the trade-off between insulation demand and heat dissipation. The focus of this work was on developing a packaging innovation that improves the module heat dissipation and offers more flexibility to its insulation design. Two strategies were explored for the packaging of an 8-kV SiC diode rectifier module:(1) double-side cooling and (2) sintered-silver bonding. Double-side cooling was realized by using short metal posts rather than long and thin wire bonds for device interconnection, forming a low-profile package with devices sandwiched between two insulated metal substrates. Sintered-silver bonding enabled the devices to function reliably at over 250 °C. Simulations of the packaged module showed a low interconnect inductance of 2.67 nH and a 50% less heat transfer coefficient required to cool the chips. Prototypes of the module were fabricated, and preliminary electrical testing results validated the package design.

Desired Properties of a Nonlinear Resistive Coating for Shielding Triple Point in a Medium-Voltage Power Module

Abstract: This work explores a nonlinear resistive field-grading material to shield the triple points prior to silicone encapsulation, thus, increasing the power module's partial discharge inception voltage. The material is aimed at coating the metal edges of a trench on a direct-bond-copper substrate for over 10 kV applications. To find the desired coating properties that would significantly reduce the field intensity at the triple point, an application-specific guideline is first prescribed for determining the baseline properties of the coating. Then, electric field simulations show that with only 10% coverage of coating at each triple-point edge, the field stress across the trench could be reduced by as much as 40%, i.e., over 67% improvement in the partial discharge inception voltage. A parametric study is done to evaluate the effect of each of the material properties, as well as the coating width, and shows that the improvement in field-stress reduction could be further increased. The material properties prescribed in this study can serve as targets for materials designers and engineers to develop insulation materials for packaging medium-voltage power modules.

A Constant-Current ZVS Class-E Inverter With Finite Input Inductance

Abstract: A Class-E inverter with finite input inductance for constant output current and zero-voltage switching (ZVS) over a load range is presented, by combining a Class-E inverter (designed to generate a constant voltage) with a trans-susceptance network (transfers voltage to current and keeps ZVS). Both switching frequency and duty cycle are fixed under load variation. The tradeoffs between the output current value, the effective load range, and the input current ripple are discussed. Same output power capability as the optimum condition in the traditional Class-E inverter is achieved. The expectations were validated by a design switched at 6.78 MHz with 10-V input, 1.4-A output, and 12.6-W maximum output power. ZVS is maintained with 16% output current varying over a 10:1 output power range.

Circuit Modeling of the Mechanical-Motion Rectifier for Electrical Simulation of Ocean Wave Power Takeoff

Abstract: As is the case with several other mechanical power takeoffs (PTOs), the mechanical-motion-rectifier-based PTO consists of components, such as one-way clutches, gears, a ball screw, mechanical couplings, and a generator. Equivalent circuit models have been created in this article to describe the dry frictions, viscous damping, and mechanical compliances in these components, so the nonideal efficiency and nonlinear force of the PTO can be predicted in electrical simulations by integrating these subcircuit models. The circuit model is simplified, and its parameters are categorized as dc and ac unknowns. The dc and ac tests on the PTO are performed sequentially to extract two sets of parameters through linear regression or nonlinear curve fitting. Then, the model is validated through its prediction capability over 25 test conditions on input forces, output voltages, and efficiencies, with correlation coefficients of 0.9, 0.98, and 0.981, respectively.

Effects of SiO2 inclusions on sintering and permeability of NiCuZn ferrite for additive manufacturing of power magnets

Abstract: 3D-printing has the potential to ease fabrication of novel magnetic components that lead to miniaturization of power electronic devices. The main challenge lies in the lack of functional magnetic feedstock that has many choices of different magnetic permeabilities. NiCuZn ferrite feedstock was developed of which the permeability was tailored by adding different fractions of silica particles. This paper developed a formulation guideline of the feedstock, which prescribes the fraction of silica to add from the target permeability of the feedstock. To study the mechanisms of the effects of silica on permeability of the ferrite feedstock, the feedstock was 3D printed into toroid cores and the permeability, density, and microstructure were characterized. Formulas from existing models describing the different aspects of the silica’s effects on the ferrite were evaluated and combined to build the guideline. The guideline accurately predicted permeability of the composite ferrite when silica was

Additive Manufacturing of Hetero-Magnetic Coupled Inductors

Abstract: Coupled inductors are used in power electronics to improve transient response and reduce component size. To further increase inductance density and reduce fringing fields, a combination of different magnetic materials was introduced to create a hetero-magnetic coupled inductor (HMCI) core. To simplify and speedup the prototyping of HMCI designs, in this study a paste-extrusion additive manufacturing platform was used to co-process a custom-developed magnetic and conductive feedstock. First, ferrite plates of each having a winding were co-printed and co-sintered. Then, the HMCI was assembled by bonding together the ferrite + winding plates using a low-temperature curable magnetic paste. The finished HMCI was characterized by measuring its dc winding resistances, leakage inductance, and coupling coefficient. The coupling coefficient of the HMCI was easily adjustable to meet various requirements for high-frequency interleaved multiphase buck converters. This work demonstrates the feasibility of applying an additive manufacturing process for rapid prototyping of an unconventional power magnetic component having an inhomogeneous magnetic core.

Omnicoupled Inductors (OCI) Applied in a Resonant Cross-Commutated Buck Converter

Abstract: All input and output inductors of a resonant cross-commutated buck (rccBuck) converter are coupled to shrink the core size while maintaining efficiency. The steady-state current ripples of the omnicoupled Inductors (OCI) are modeled by separating common-mode and differential-mode voltages. The coupling coefficients are derived to minimize inductance and energy. In this article, the twisted E-E cores with one-turn printed circuit board (PCB) windings are proposed to realize the recommended coupling coefficients. The magnetic flux paths and reluctances are analyzed. A prototype was fabricated to demonstrate the OCI in a 2-MHz rccBuck converter with 12-V input and 3.3-V output. Magnetic volume was 57% smaller and efficiency was 0.5% higher (over the entire load range) than those of the reference rccBuck converter with discrete inductors.

New Way of Generating Electromagnetic Waves

Abstract: This article presents a new method for generating low-frequency electromagnetic waves for navigation and communication in challenging environments, such as underwater and underground. The key concept is to disturb the magnetic energy stored around a permanent magnet in a time-variant fashion. The magnetic reluctance of the medium around the permanent magnet is modulated to alter the magnetic flux intensity and direction (disturb the stored energy) in order to achieve this goal. The nonlinear properties of the surrounding magnetic material are a critical phenomenon for efficient and effective modulation. Since the proposed way of generating the electromagnetic field is not based on a second-order system (resonant structure), the bandwidth of any modulation schema is not limited to the overall system quality factor. A transmitter is prototyped as a proof of concept, and the generated field is measured. Compared with the rotating magnet, the prototyped transmitter can modulate up to 50% of the permanent magnet’s stored energy with much lower power consumption.

A Soft Magnetic Moldable Composite With Tri-Modal Size Distribution for Power Electronics Applications

Abstract: Soft magnetic moldable composites (SM2Cs) would be ideal for magnetic integration in power electronics converters because they can be readily processed at low temperature without pressure into the core shapes prescribed for the circuits. However, most of the existing SM2Cs have low relative magnetic permeability (<30) and low saturation flux density (<1 T). To improve the magnetic properties, we combined three different magnetic powders giving rise to a tri-modal size distribution and an acrylic polymer to produce a paste of tri-modal SM2C. The paste was molded and then cured below 200 °C without pressure to form cores offering a relative permeability over 30, a saturation flux density over 1.0 T, and a low core-loss density. The ease of its processing and high-performance properties make this SM2C a good candidate material for power electronics applications.

Thermal and Mechanical Design of a High-Voltage Power Electronics Package

Abstract: This paper presents the thermal and mechanical design aspects of a power electronics package with 5-kV GaN devices. Through finite element analyses, we investigated the impact of different cooling configurations and device locations on the thermal performance and reliability of the package. We found that placing the devices closer to the direct bond copper substrate as opposed to a centered approach in the proposed double-sided cooling configuration resulted in improved heat dissipation. This approach also reduced the total number of attachment layers, thereby likely improving the reliability of the package. Furthermore, simulations revealed that the device location had a negligible impact on the thermomechanical behavior of the attachment layers, as they are more prone to the local coefficient of thermal expansion mismatch.