Condition monitoring techniques for power electronics components are important for reducing maintenance costs and increasing reliability in systems such as aircraft. This article presents a noninvasive condition monitoring system that utilizes time-frequency analysis of conducted electromagnetic interference (EMI) to classify the health of the dc-link capacitor within a three-phase inverter. The approach proposes a combined EMI filter and measurement board which is placed on the dc bus of the inverter. This board filters conducted EMI effectively and enables the inverter to comply with MIL-STD-461 G. It also enables EMI measurements to be collected for condition monitoring applications. The EMI content obtained from this board is analyzed from 15–43 MHz during switching events using a continuous wavelet transform. These characteristic switching images are used to train support vector machine models that are able to classify dc-link health into one of five health stages with accuracy up to 100%.

Aerospace applications demand power electronics with high power density and high reliability. Many strategies exist to improve power density and researchers are examining new power devices as a tool to achieve this goal. Commercially available silicon carbide (SiC) power devices have Figures of Merit significantly higher than silicon (Si) components. Their smaller on-resistances facilitate lower conduction losses, and their small parasitic capacitances result in reduced switching loss and increased transient switching speed. These advantages, when applied strategically, can increase power density. The reliability of SiC power devices has also improved during the last decade, but questions remain about conducted and radiated emissions that result from SiC-based hardware due to their high switching speed. This research models the conducted electromagnetic interference (EMI) of a three-phase inverter and electric machine. Experimental results confirm the validity of the model under DO-160.

Carbon nanotubes (CNTs) have unique thermal/electrical/mechanical properties and high aspect ratios. Growth of CNTs directly onto reactive material substrates (such as metals and carbon based foam structures, etc.) to create a micro-carbon composite layer on the surface has many advantages: possible elimination of processing steps and resistive junctions, provision of a thermally conductive transition layer between materials of varying thermal expansion coefficients, etc. Compared to growing CNTs on conventional inert substrates such as SiO2, direct growth of CNTs onto reactive substrates is significantly more challenging. Namely, control of CNT growth, structure, and morphology has proven difficult due to the diffusion of metallic catalysts into the substrate during CNT synthesis conditions. In this study, using a chemical vapor deposition method, uniform CNT layers were successfully grown on copper foil and carbon foam substrates that were pre-coated with an appropriate buffer layer such as Al2O3 or Al. SEM images indicated that growth conditions and, most notably, substrate surface pre-treatment all influence CNT growth and layer structure/morphology. The SEM images and pull-off testing results revealed that relatively strong bonding existed between the CNT layer and substrate material, and that normal interfacial adhesion (0.2‒0.5 MPa) was affected by the buffer layer thickness. Additionally, the thermal properties of the CNT/substrate structure were evaluated using a laser flash technique, which showed that the CNT layer can reduce thermal resistance when used as a thermal interface material between bonded layers.

Carbon nanotubes (CNTs) are known for high thermal conductivity and have potential use as nano-radiators or heat exchangers. This paper focuses on the thermal performance of carpet-like arrays of vertically aligned CNTs on solid graphite substrates with the idea of investigating their behavior as a function of carpet dimensions and predicting their performance as thermal interface material (TIM) for electronic device cooling. Vertically aligned CNTs were grown on highly oriented pyrolytic graphite (HOPG) substrate, which creates a robust and durable all-carbon hierarchical structure. The multi-layer thermal analysis approach using Netzsch laser flash analysis system was used to evaluate their performance as a function of carpet height, from which their thermal properties can be determined. It was seen that the thermal resistance of the CNT array varies linearly with CNT carpet height, providing a unique way of decoupling the properties of the CNT carpet from its interface. This data was used to estimate the thermal conductivity of individual multi-walled nanotube strands in this carpet, which was about 35 W/m-K. The influence of CNT carpet parameters (aerial density, diameter, and length) on thermal resistance of the CNT carpet and its potential advantages and limitations as an integrated TIM are discussed.

Absolute number densities of atomic species produced by nanosecond (ns)-duration, repetitively pulsed electric discharges are measured by two-photon-absorption laser-induced fluorescence (TALIF). Unique to this work is the development of femtosecond-laser-based TALIF (fs-TALIF) that offers a number of advantages over more conventional nanosecond (ns)-pulse-duration laser techniques, such as higher-fidelity quenching rate measurements over a wide pressure range, significantly reduced photolytic interference (including photo-dissociation and photo-ionization), ability to collect two-dimensional images of atomic-species number densities with high spatial resolution aided by higher signal level, and efficient and accurate measurements of atomic-species number densities due to the higher repetition rates of the laser. For full quantification of these advantages, atomic-oxygen TALIF signals are collected from an atmospheric-pressure plasma jet. employing both ns-and fs-duration laser-excitation pulses and the results are compared and contrasted.

The axial and radial variations of the relative electron density distributions have been measured in an obstructed helium dc discharge by a laser collision induced fluorescence (LCIF) from 4 3D to 2 3P state at 447 nm. For this 1.75 Torr, 6.5 mm gap, 1.6 mA helium obstructed gas discharge, the on-axis LCIF intensity is mostly determined by the neutral atom collision. On the basis of this information, we have plotted the LCIF data to obtain a normalized electron density variation in the inter-electrode volume by plotting the ratio, I−Ion−axisIon−axis, where I is the radial variation of the LCIF intensity and Ion-axis is the on-axis LCIF intensity. The normalized LCIF data show an annular current density distribution in this obstructed discharge with the current peaking near the radial boundary of the electrodes.

Carbon nanotubes (CNTs), with exceptional thermal and mechanical properties as well as inherently high surface area, are an attractive candidate for integrating into thermal structures of advanced power electronics. Growth of vertically aligned carbon nanotubes (VACNTs) directly onto copper (Cu) substrates is a promising approach to apply CNTs as novel thermal interface materials (TIMs) in electronics packaging. However, compared to growing CNTs on conventional inert substrates such as SiO2, direct growth of controllable CNT arrays onto Cu substrates is significantly more challenging due to the diffusion of metallic catalyst into the substrate during growth. By depositing an appropriate buffer layer on the Cu substrate surface, VACNTs of good alignment and high quality were reproducibly synthesized on the Cu substrate via the chemical vapor deposition (CVD) method in this study. The effect of different buffer layers on the CNT growth, nanotube structure and quality was investigated (SEM, Raman), particularly in terms of the interfacial properties between the CNT array and Cu substrate (Tensile compression force tester, Laser Flash Analysis). Our experimental results indicated that the buffer layer material, deposition method, and thickness play a key role in regulating the CNT layer growth/structure, leading to variable mechanical and thermal properties. The fundamental understanding thus obtained allows the successful synthesis of VACNT on copper substrates with desired structure and properties.

In a previous paper, we introduced a modified Steinmetz equation to account for dc-bias field effects, which requires only a simple dc permeability measurement to predict total power loss. In this paper, we expanded our investigation to include Finemet nanocrystalline material and found that our modified Steinmetz formalism was effective in predicting dc-bias-related losses for this system as well. In this paper, it was observed that Finemet cores exhibit lower losses than Metglas cores under identical test frequencies and bias fields. In addition, we show that a full characterization of the dc loss component necessitates the consideration of higher order (n > 1) harmonic components. In order to quantify these higher frequency loss components, a dc-dc converter-based test system was built to intentionally introduce inductor current harmonics by varying the filter capacitance and parasitic inductance of the test system. Both core types were evaluated under fundamental frequencies of 20 to 150 kHz and dc-bias fields of up to 1.3 kA/m, with the inclusion of distorted waveforms obtained by varying filter capacitance. At higher frequencies, the Metglas cores were found to exhibit greater loss fractions associated with the higher order harmonic components. A detailed summary of the measured core loss characteristics for both core types is included and discussed. This paper includes the details of the measurements, the modified Steinmetz relation, and the loss extraction algorithm used for analysis.

Femtosecond, two-photon-absorption laser-induced-fluorescence (fs-TALIF) spectroscopy is employed to measure space-and time-resolved atomic-oxygen distributions in a nanosecond, repetitively pulsed, externally grounded, atmospheric-pressure plasma jet flowing helium with a variable oxygen admixture. The high-peak-intensity, low-average-energy femtosecond pulses result in increased TALIF signal with reduced photolytic inferences. This allows 2D imaging of absolute atomic-oxygen number densities ranging from 5.8 x 10(15) to 2.0 x 10(12)cm(-3) using a cooled CCD with an external intensifier. Xenon is used for signal and imaging-system calibrations to quantify the atomic-oxygen fluorescence signal. Initial results highlight a transition in discharge morphology from annular to filamentary, corresponding with a change in plasma chemistry from ozone to atomic oxygen production, as the concentration of oxygen in the feed gas is changed at a fixed voltage-pulse-repetition rate. In this configuration, significant concentrations of reactive oxygen species may be remotely generated by sustaining an active discharge beyond the confines of the dielectric capillary, which may benefit applications that require large concentrations of reactive oxygen species such as material processing or biomedical devices.

Controllable growth of carbon nanotube (CNT) arrays on carbon substrates requires an oxide buffer layer material such as Al2O3 and SiO2. This research analyzes the influence of these buffer layers on a model carbon substrate composed of pyrolyzed carbon film on flat silicon. Electron microscopy was used to analyze nanocatalyst morphology, particle density, CNT nucleation rate, and CNT morphology for selected growth conditions. It was observed that nanotubes grown on Al2O3 buffer layers had the fastest CNT nucleation rates and more uniform tube diameters, with residual nanocatalyst particles having elongated shapes. In contrast, CNTs formed on SiO2 buffer layers nucleated slower and had more variation in tube diameters, while the nanocatalyst particles remained spherical. Comparisons among SiO2 buffer layers deposited using three different techniques—atomic layer deposition, microwave plasma enhanced chemical vapor deposition and thermal oxidation—indicated that surface roughness plays an important role. Silica layers with the highest roughness had the highest catalytic particle density and tallest carpet height for identical growth conditions. These results provide new insight into the importance of buffer layer chemistry and morphology on growth and controllability of CNT arrays.

Direct current bias related hysteresis loss characteristics of three commercially available magnetic materials: (1) an iron based Metglas tape core, (2) a Sendust powder core, and (3) a Mn-Zn based ferrite in both un-gapped and gapped configurations were studied. The measurements are conducted for a fixed external field Hext, a fixed flux swing (ΔB), and a fixed maximum forward magnetization (Bmax) as a function of the external bias field. In all the measurements, a direct correlation is found between permeability and measured loss values as a function of dc bias field. Increased hysteresis losses are measured in the magnetization rotation region in which classical domain theory predicts minimal losses. The observed trends are discussed within the frame work of classical domain theory.

: Positive polarity applied voltage and gas pressure dependent scaling of cathode directed streamer propagation properties in helium gas flow guided capillary dielectric barrier discharge have been quantified from streamer velocity, streamer current, and streamer optical diameter measurements. All measurements of the non-stochastic streamer properties have been performed in a variable gas pressure glass cell with N2 co-flow and under self-consistent Poisson electric field dominated conditions to permit data comparison with 2-D streamer dynamics models in air/nitrogen. The streamer optical diameter was found to be nearly independent of both gas pressures, from 170 Torr up to 760 Torr, and also for applied voltages from 6 to 11 kV at 520 Torr. The streamer velocity was found to increase quadratically with increased applied voltage. These observed differences in the 2-D scaling properties of ionization wave sustained cathode directed streamer propagation in helium flow channel with N2 annular co-flow compared to the streamer propagation in air or nitrogen have been shown to be caused by the remnant ionization distribution due to large differences in the dissociative recombination rates of He2 + versus N4 + ions, for this 5 kHz repetition rate applied voltage pulse generated streamers.

Recent advances in silicon carbide (SiC) power semiconductor technology and resulting availability of SiC Schottky rectifiers and controlled devices (bipolar junction transistors, JFETs, and MOSFETs) make it possible to design and implement power converters capable of operating at 200 °C. The design, prototype development, operation, and testing of a 74 kHz, 2 kW, 100 V/270 V inversely coupled, in...

Soft magnetic components in electronic systems are often subjected to dc bias-flux conditions. These dc bias conditions result in distorted hysteresis loops, increased core losses, and have been shown to be independent of core material. The physical origin of these increased losses is not well understood and there is no simple model that can predict these losses without extensive measurements. Absence of a widely accepted model coupled with the complete lack of dc loss attributes on core manufacturers' data sheets result in a requirement to empirically determine loss values for specific design applications. These deficiencies have motivated our efforts to investigate dc bias dependent loss phenomenon in a Fe-based Metglas core inductor operating in a dc-dc boost converter. Since dc flux levels in the core are proportional to the controllable converter load currents, this topology is ideal to study dc-related losses. Inductor core B - H hysteresis loop characterization was accomplished as a function of switching frequency, input voltage, and load current operating conditions and parameters. In this paper, the core loss results were presented as a function of the dc bias conditions, and the results showed that the core losses increased with the pre-magnetized (B de ) fields. As a result of our observations, we have proposed a modification to the conventional Steinmetz loss equation to include the effects of dc pre-magnetization flux in the core.

A hermetic multichip power package for silicon carbide devices that will operate at 200 degrees C ambient and switch 50-100 A has been developed. The Al2O3/MoCu structure, in which the SiC junction field-effect transistors and diodes are attached, was designed to hermetically seal the device areas. Details of the materials and processes used to fabricate the package are discussed. Die attach, ribbon bonding, and lid attach, as well as thermal modeling, electrical testing, and thermal cycling results are also described.

An all SiC 600V / 6 m hermetic half-bridge power module has been developed to operate at ambient temperatures of 200oC and with junction temperatures near 250oC. The modules use SiC trench JFET technology and can output over 100A at Tj=250oC. Double pulsed switching was performed up to temperatures of 150oC with a measured total switching energy of 0.73mJ