Created: 10/24/95 Updated: 8/18/98

Modelling of Natural Convection From Horizontal Printed Circuit Board

Project Number : C95-17

Point of Contact:

Dr. Yogi Joshi
CALCE EPRC
email: yogi@calce.umd.edu
Phone: (301) 405-5428
Fax: (301) 314-9269
Objectives Background Work Accomplised

Objective

Expand the capability of CALCE Thermal Analysis Tool to handle passive cooling of a horizontally oriented printed wiring board in a box.


Background

Temperature variations within electronic equipment are a primary source of undesirable stresses. These stresses are cyclic in nature based on power cycling of equipment and/or variations in the external environment. As part of the CALCE EPRC's effort to characterize failure mechanisms of components and printed wiring board, an effort has focused on evaluating temperature distributions within electronic equipment. Currently, the thermal analysis capability includes conduction, forced convection and passive cooling of vertically oriented printed wiring boards. Traditionally, horizontally oriented heated boards have received little attention. However, with increase in commercial applications and desire to simplify cooling system, passive cooling is extremely attractive.


Work Accomplished

For the present study, three configurations of a horizontal PCB in increasing order of complexity were selected. These include a uniformly heated PCB, a five by five array, and a three by three array of equally sized heated components on a horizontal substrate. A design of experiments study was carried out to identify the computational runs needed to examine the thermal behavior of these horizontal boards in an enclosure. First, a simulation model for a heated horizontal surface in an enclosure was developed using a computational fluid dynamics code. Thirty five detailed computational fluid dynamics (CFD) based runs to charac terize thermal behavior of horizontally placed PCBs were carried out. Based on the results of these runs, a simpler empirical convection correlation was derived and implemented within the existing CALCE Thermal Analysis Tool. The empirical correlation was calibrated based on results from the CFD simulation.