A series of experiments was performed to study forced convection from rectangular arrays of electronic components. Effects of channel height, planar spacing, component row number, and approach velocity are assessed. Correlations for the temperature rise of a component due to its own power and due to heating by upstream modules are presented. The components used were aluminum-capped ceramic pin grid array modules, of the same shape as “flatpacks”, measuring 37 mm square and 5.8 mm tall. The space between the rows and columns of modules was varied from near zero to about one module length. Channel height ranged from about two to five times module height. The air approach velocity ranged from 0.5 to 5.5 meters per second, corresponding to module length Reynolds numbers from 1000 to 11000. The heat transfer coefficient varied from 25 to 75 W/m2K. The temperature rise of a component due to its own power was found to be a strong function of velocity and less dependent on channel height and row number. The effect of velocity was weaker on the densest configuration than on the two sparser spacings. The two densest spacings had little dependence on channel height; the sparsest configuration had only a weak dependence on row number. The temperature rise of components downstream of a heated component exhibited similar but stronger dependence on velocity and channel height and the same dependence on row number.

This content is only available via PDF.
You do not currently have access to this content.