Coupled thermal-fluid-electrical simulation for printed circuit board design

Abstract

In this work we describe a coupled thermal-fluidelectrical simulation for printed circuit boards and their surrounding enclosures. A finite-element based electrical solver is used to compute the distribution of electrical currents and ohmic power losses in the board's power and ground distribution network. The power loss information is then used as an input to a computational fluid dynamics (CFD) solver. The CFD solver computes the resulting air flow velocities and temperature distribution in the board and its surroundings. The temperature map is then returned to the electrical solver, which adjusts the electrical conductivity as a function of position and computes an updated set of current and power losses. This procedure is iterated until it obtains a converged, self-consistent solution for the steady-state temperature and ohmic power loss distributions in the board-enclosure system.

Impact of Ports Reference Choice on S-Parameter Modeling of BGA Package Interconnections

“…In this framework, the extracted S-parameters are the starting point to generate macro-models for circuit simulation, in addition to represent reflection, transmission and crosstalk losses of a passive network. The S-parameters extraction techniques used for complex BGA (Ball Grid Array) packages mostly derive from PCB modeling framework and use the same commercial tools, based on hybrid solvers and 2.5D approximation; in particular, the tool used for the extractions is Ansys SIwave [5], [6], [7]. Package interconnections are typically much shorter than PCB interconnections.…”

“…A typical S-parameter characterization is carried out by extraction tools by setting up a port between the terminals of the interconnection (i.e., the signal forward path) and a reference point, which constitutes the terminal of the current return path [6]. Ports are usually fictitious circuit elements used to apply an excitation (in the cases analyzed in this work, current sources are exploited) to the structure in a specific position.…”

Thermal-aware DC IR-drop co-analysis using non-conformal domain decomposition methods

“…Moreover, owing to the linear temperature dependency of metal resistivity, the non-uniform temperature distribution affects the electrical performance of the power delivery network (PDN) and substantially increases the IR drops in the power/ground planes (Bracken et al 2011;Farina et al 2011). In particular, modules that draw more current from the PDN, owing to higher power demands, will suffer worse IR-drop effects.…”