Massimiliano de Magistris

Dynamic Electrothermal Macromodeling: an Application to Signal Integrity Analysis in Highly Integrated Electronic Systems

The problem of dynamic electrothermal equivalent extraction for system level and signal integrity analysis is considered in the framework of standard macromodeling techniques. The self-heating and mutual thermal impedances of the embedded elements of the system-preliminarily evaluated via 3-D thermal simulations-are reduced with passive identification methods, and equivalent electrical models are realized with a Foster multiport synthesis scheme. The identification is pursued with a convex optimization approach, which is found to be well suited for the thermal problem. As a case study, the coupled electrothermal analysis of an ultra-thin chip-stacking module including two thinned silicon chips interconnected by a copper line is fully carried out in SPICE.

On the concretely passive realization of reduced circuit models based on convex constrained positive real fractions identification

The paper deals with the problem of concretely passive realizations of reduced models directly associated to Foster (matrix) expansions. After considering the procedure named Positive Fraction Vector Fitting (PFVF), based on convex passive constrained optimization applied to a pole-residue scheme identification, we will show how such formulation allows unique and straightforward possibility of concretely passive realizations, by means of generalized Foster matrices. This remarkable property, not common to other passive macromodeling techniques, complements the already known favorable aspects of PFVF with a straightforward possibility of a purely passive elements circuit realization.

Electromagnetic Models of Plasma Breakdown in the JET Tokamak

This paper presents the electromagnetic modeling of the plasma current breakdown phase of the JET tokamak. The first part of this paper models the presence of the JET iron core up-down asymmetry and the effects of the eddy currents in the reconstruction of the magnetic topology needed for the plasma start. The second part describes the approach used to evaluate the ionized particle connection length inside the vacuum chamber at breakdown. The results obtained are validated using JET experimental measurements.

Synchronization of networks of non-identical Chua circuits: Analysis and experiments

This paper is concerned with the theoretical and experimental analysis of synchronization and pinning control of networks of non-identical Chuas circuits. The design and implementation is presented of an appropriate setup to carry out experiments on networks of these nonlinear circuits with easily reconfigurable parameter values. Then, the theoretical expectations are validated of the so-called Extended Master Stability Function approach to study the onset of synchronization in the presence of real parameter mismatches between the circuits at the network nodes. The validation is carried out both numerically and experimentally. For the first time, the EMSF is also used to investigate pinning synchronization in a network of non-identical circuits where a master node (or pinner) is used to drive the network dynamics towards some desired synchronous evolution. The numerical and experimental results confirm the viability of the EMSF as an effective analysis and design tool.

Modular experimental setup for real-time analysis of emergent behavior in networks of Chua's circuits

Summary This paper describes the design, realization and use of an analogical, fully reconfigurable experimental setup to analyze the complex dynamics of networks of chaotic circuits. It reports details of the implementation and characterization of the setup, together with representative results, showing its flexibility and potential. The setup allows to choose arbitrarily the coupling strength and interconnection structure among the circuits, the type of link and to select the parameters of the node dynamics. It has a modular structure, and it can accommodate up to 32 nodes interconnected by at most 32 links. The collective dynamics of a relatively large set of different network structures and configurations has been investigated using the setup. Synchronization, pattern formation and other interesting collective phenomena were observed experimentally, their evidence being reported here as an illustration of the potential of the proposed setup. Copyright

Electrothermal Analysis of Carbon Nanotubes Power Delivery Networks for Nanoscale Integrated Circuits

This paper presents a circuit-based self-consistent electrothermal model of Power Delivery Networks (PDNs) fabricated with carbon nanotubes, for nanoscale integrated circuits. It couples the electrical submodel with an equivalent circuit submodel of the thermal problem, and allows characterizing the PDNs in terms of both voltage drop and temperature rise. The temperature dependence of the PDNs electrical parameters is taken into account through physically meaningful models of the mean free path and the number of conducting channels in carbon nanotubes. The model is used to compare the performance of conventional copper PDNs with those of PDNs realized with bundles of carbon nanotubes in different configurations (single-wall, multiwall, and mixed). The adopted values of thermal and electrical parameters are those typically obtainable with the current fabrication processes, for technology nodes of 65 and 22 nm and different core power consumption.

Numerical validation of a procedure for direct identification of passive linear multiport with convex programming

The paper deals with the numerical validation, performance evaluation and robustness assessment of a procedure for the direct identification of passive transfer matrices based on convex programming. Validation is pursued by producing data sets from lumped multiport systems with random parameters (passive, non passive, and possibly affected by random noise), then evaluating the identification ability of the considered method. Results demonstrate how the considered approach satisfactorily covers the passive identification of a large class of data sets, even in presence of significant passivity violations or noise flawed data, at average accuracies comparable with non passive identifications obtained with standard Vector Fitting.

Effective Electrothermal Analysis of Electronic Devices and Systems with Parameterized Macromodeling

We propose a parameterized macromodeling methodology to effectively and accurately carry out dynamic electrothermal (ET) simulations of electronic components and systems, while taking into account the influence of key design parameters on the system behavior. In order to improve the accuracy and to reduce the number of computationally expensive thermal simulations needed for the macromodel generation, a decomposition of the frequency-domain data samples of the thermal impedance matrix is proposed. The approach is applied to study the impact of layout variations on the dynamic ET behavior of a state-of-the-art 8-finger AlGaN/GaN high-electron mobility transistor grown on a SiC substrate. The simulation results confirm the high accuracy and computational gain obtained using parameterized macromodels instead of a standard method based on iterative complete numerical analysis.

Analogic Realization of a Non-linear Network with Re-configurable Structure as Paradigm for Real Time Analysis of Complex Dynamics

A novel experimental set-up realized for the real time analysis of reconfigurable complex networks, with chaotic Chua’s circuits as nodes, is considered. It has been designed to easily perform large scale experiments on networks of chaotic oscillators, possibly exploring in real time the parameters space in terms of topologies, coupling strengths, nodes’ dynamics, with potential application in the area of neuro-computing and associative dynamic memories. A sample of the capabilities is given by considering diffusive coupling with a large range of coupling strengths in a set of topologies, and first experiments on a ring of 32 chaotic nodes are reported. Synchronization, chaotic waves and patterns are experimentally observed, and the potential of the realized set-up in terms of accuracy, flexibility and analysis time is fully revealed.

Experimental Dynamics Observed in a Configurable Complex Network of Chaotic Oscillators

New experimental results are presented, obtained with a recently developed set-up, designed and tailored to easily perform large scale experiments on networks of chaotic oscillators, and using Chua’s circuits as nodes. A ring of 16 chaotic nodes, diffusively coupled with a large range of coupling strength values is considered, and accurately characterized. Complete synchronization, patterns and chaotic waves have been observed and described, so revealing the potential of the realized set-up for realizing accurate experiments in configurable complex networks.