Thanks to the increasing availability of SiC MOSFETs with outstanding static and dynamic performances, the number of applications in which these devices are used is rapidly growing. Despite that, the maximum current rating of such devices is usually limited at few hundred Amps, which sets an upper bound for the power level at which these transistors can be adopted. A viable solution to this problem consists in paralleling several SiC MOSFETs. However, the design of parallel configurations needs optimization since mismatched performances can cause the enhanced stress of a subset of devices, which can ultimately lead to premature failure of the whole module. In this contribution, several sets of Monte Carlo electrothermal simulations of parallel SiC MOSFETs are used to systematically relate the deviations of devices and circuit parameters to the resulting uneven power dissipation. To this purpose, a set of relevant parameters is identified and statistically described. Thereafter, the simulation of a 200 kHz synchronous buck converter relying on mismatched parallel MOSFETs is performed as case study. Eventually, a methodology to derive a guideline for the design of reliable multichip configurations is developed. The methodology is based on the iteration of Monte Carlo simulations for different tolerances of the parameters.

Myocardial Infarction (MI), commonly known as heart attack, is a cardiac condition characterized by damage to a portion of the heart, specifically the myocardium, due to the disruption of blood flow. Given its recurring and often asymptomatic nature, there is the need for continuous monitoring using wearable devices. This paper proposes a single-microcontroller-based system designed for the automatic detection of MI based on the Edge Computing paradigm. Two solutions for MI detection are evaluated, based on Machine Learning (ML) and Deep Learning (DL) techniques. The developed algorithms are based on two different approaches currently available in the literature, and they are optimized for deployment on low-resource hardware. A feasibility assessment of their implementation on a single 32-bit microcontroller with an ARM Cortex-M4 core was examined, and a comparison in terms of accuracy, inference time, and memory usage was detailed. For ML techniques, significant data processing for feature extraction, coupled with a simpler Neural Network (NN) is involved. On the other hand, the second method, based on DL, employs a Spectrogram Analysis for feature extraction and a Convolutional Neural Network (CNN) with a longer inference time and higher memory utilization. Both methods employ the same low power hardware reaching an accuracy of 89.40% and 94.76%, respectively. The final prototype is an energy-efficient system capable of real-time detection of MI without the need to connect to remote servers or the cloud. All processing is performed at the edge, enabling NN inference on the same microcontroller.

In this paper, an advanced electrothermal simulation strategy is applied to a 3.3 kV silicon carbide MOSFET power module. The approach is based on a full circuital representation of the module, where use is made of the thermal equivalent of the Ohm’s law. The individual transistors are described with subcircuits, while the dynamic power-temperature feedback is accounted for through an equivalent thermal network enriched with controlled sources enabling nonlinear thermal effects. A synchronous step-up DC-DC converter and a single-phase inverter, both incorporating the aforementioned power module, are simulated. Good accuracy was ensured by considering electromagnetic effects due to parasitics, which were experimentally extracted in a preliminary stage. Low CPU times are needed, and no convergence issues are encountered in spite of the high switching frequencies. The impact of some key parameters is effortlessly quantified. The analysis witnesses the efficiency and versatility of the approach, and suggests its adoption for design, analysis, and synthesis of high-frequency power converters in wide-band-gap semiconductor technology.

Diabetes is a chronic condition, and traditional monitoring methods are invasive, significantly reducing the quality of life of the patients. This study proposes the design of an innovative system based on a microcontroller that performs real-time ECG acquisition and evaluates the presence of diabetes using an Edge-AI solution. A spectrogram-based preprocessing method is combined with a 1-Dimensional Convolutional Neural Network (1D-CNN) to analyze the ECG signals directly on the device. By applying quantization as an optimization technique, the model effectively balances memory usage and accuracy, achieving an accuracy of 89.52% with an average precision and recall of 0.91 and 0.90, respectively. These results were obtained with a minimal memory footprint of 347 kB flash and 23 kB RAM, showcasing the system’s suitability for wearable embedded devices. Furthermore, a custom PCB was developed to validate the system in a real-world scenario. The hardware integrates high-performance electronics with low power consumption, demonstrating the feasibility of deploying Edge-AI for non-invasive, real-time diabetes detection in resource-constrained environments. This design represents a significant step forward in improving the accessibility and practicality of diabetes monitoring.

In this work, a comparison between the gate-driving requirements of p-GaN HEMTs with gate contact of Schottky and Ohmic type is presented. Furthermore, the presence of a gate current of different magnitude is experimentally verified for both types of devices. Successively, the possibility of using the gate current as a temperature-sensitive parameter and its monitoring during real circuit operation is proposed. The viability of monitoring the gate current without introducing additional complexity in the gate driver is examined through experimental measurements on commercially available p-GaN HEMTs.

In this document is described the design, the characterization, the development, and pre-industrialization of a novel interrogation system for Fiber Bragg Grating (FBG) sensors capable to take advantage of the Internet of Things (IoT) technology. Due to its complexity and cost, classical interrogation systems limit the design and use of a sensor network based on fiber optic technology that are often replaced with less performant electronic sensor systems. The proposed system has the potential to overcome the limits of today’s commercial interrogators, converting the classic FBG interrogator into an IoT node, where each circuit is represented by a sensory node that continuously exchanges and publishes data on the cloud. Moreover, the proposed concept may be applied in harsh environments, still taking advantage of the fiber optic technology. The interrogator is composed of a passive optical device section; 4-layer electronic boards section (one board for each connected FBG in the external array), which conditions the optical signal to a compliant voltage at the downstream analog-to-digital converter (ADC). In the subsequent digital processing section, created inside an ARM based board, the conversions are then carried out through dedicated algorithms to match the voltage variation to the Bragg wavelength shift and the subsequent conversion into the relative physical quantity (temperature, deformation, or humidity) via a subsequent conversion algorithm. Finally, physical values, in real time, are published on a remotely accessible server. The proposed system was tested and validated comparing data produced by commercial interrogations. The tests were performed using an array composed of 6 (six) independent FBG sensors. It was tested through stability and on-field measurements, detecting one temperature compensated strain point, and four temperature monitoring points.