Daniele Romano

Efficient Frequency-Domain Analysis of PEEC Circuits Through Multiscale Compressed Decomposition

The solution of mixed electromagnetic/circuit problems is important for the electromagnetic compatibility/signal integrity/power integrity system designs. The ever-increasing frequency content of signals and decrease of geometrical features requires the 3-D electromagnetic methods, such as the partial element equivalent circuit (PEEC) method, to be used for the analysis and design of high-speed circuits. Very large systems of equations are often produced and their efficient solution can be extremely challenging. In this paper, we propose a new frequency-domain PEEC solver which is based on the adaptive cross approximation and singular value decomposition. Taking advantage of the rank deficiency of the dense partial inductance and coefficient of potential matrices, a multiscale block decomposition is adopted to explicitly compute the inverse of the admittance matrix of the PEEC circuit. The proposed approach provides both speedup and memory storage saving, while preserving the accuracy. The efficiency of the proposed method is demonstrated through its application to the PEEC modeling of typical interconnect problems.

Acceleration of Time-Domain Nodal-Analysis of Partial Element Equivalent Circuits Through Multiscale Compressed Decomposition

In this paper, we show the efficiency and accuracy of an implicit time-domain solver for the nodal analysis of circuits generated by the partial element equivalent circuit method. The multiscale block decomposition, based on the recursive partitioned matrix inverse formula and compression techniques, is applied to the implicit time-domain solver in order to further speeding up the computations. The effectiveness of the proposed approach is confirmed by the numerical results.

A comparison between backprojection and sensitivity methods in EIT reconstruction problems

Electrical impedance tomography is an imaging technique used to estimate the electrical conductivity distribution of human body sections. To solve the inverse problem of EIT many numerical methods have been developed. Two of these methods, namely the backprojection algorithm and the sensitivity method, are compared in this paper on some canonical models representing a section of the human body.

A surface PEEC formulation for the analysis of electrical networks in airplanes

This paper presents a systematic methodology based on the Partial Element Equivalent Circuit method to model aircrafts in order to characterize their current return network. The state-of-the-art of numerical modeling techniques for electrically large structures is used to allow accurate and fast simulations. The comparison of the numerical results with experimental data validates the procedure.

Full-wave modeling of inductive coupling links for low-power 3D system integration

In recent years, 3D system integration has emerged as a new paradigm to reduce the overall size of multichip systems (processor and memory stacks), improve data throughput, and lower assembly costs. Among the various techniques developed to connect multiple die in a 3D integrated-circuit (IC) package, inductive coupling links are very cost-effective solutions that require no specialized processing steps. While it is easy to use integrated inductors to implement inductive coupling links in standard CMOS processes, evaluating their electrical characteristics requires using an electromagnetic field solver software. And, integrating these links into a standard SPICE-like circuit design environment is not straightforward. In this paper, we describe a technique, based on the partial element equivalent circuit (PEEC) method, to model an integrated inductive coupling link as a simple lumped parameter circuit. Starting from layout and technology data, the lumped parameter circuit model can be used in a SPICE-like simulator for system design purposes.

Multiscale decomposition based analysis of PEEC models

The high level of integration has made the analysis and design of integrated circuits and packages increasingly challenging. Hence, there exists an urgent need to reduce the computational complexity of existing numerical methods. The integral equation based method known as the Partial Element Equivalent Circuit (PEEC) method naturally generates an equivalent circuit which can be analyzed in both the time and frequency domains. The enforcement of Kirchoff laws to the equivalent circuit can easily result into a very large set of equations whose solution can be extremely time consuming. In this paper, a new frequency-domain nodal analysis PEEC solver is proposed which is based on the adaptive cross approximation and recursive partitioned matrix inverse formula. The proposed approach provides a significant computational speedup, while preserving the accuracy. The efficiency of the proposed method is demonstrated through its application to a relevant interconnect problem.