A study has been conducted to determine the tehrmal characteristics of three commercial-off-the-shelf electronic (COTS) circuit boards; and temperatures ahve been predicted for the boards whent heya re operated in vacuum conditions. The COTS boards used for this study are those in the WinSystems PC104 computer/controller. The primary thermal characteristic if interest is the chip power dissipation. The condidence level/accuracy associated with the chip power dissipation determination ahs been evaluated by comparing calculated temperature results for the operating boards against measured temperatures and measure power dissipation. The measured data came from tests in which the boards were operated in a room ambient environment while running the Norton Utilities software. Durin the tests, the boards rejected heat to the room via natural convection or forced convection hear transfer (in addition to the always present radiation heat transfer). Three air speeds were used for the forced convection tests. Chip heat dissipations and temperatures were calculated with SINDA85 models representing each board as it operated in each of the convective heat rejection situations.

A high level of confidence can be associated with the circuit board thermal characteristics derived from this study. The high confidence level derives from the less than 4 degrees centigrade of correlation (difference) between measured and calculated temperatures for almost all of the chip/board areas consisdered. The typical correlation was less than 2 degrees centigrade. Calculated total power dissipation was approximately 3.2 watts versus the 3.76 watts measured. The error in the calculated power might lead to a manimum error in temperature prediction of about 4 degrees centigrade.

The temperature predictions indicate that the boards will be able to operate in a vacuum with a maximum sink temperature of 21 degrees centigrade, assuming that the chips can tolerate a reasonable (although unconfirmed) maximum case temperature of 70 degrees centigrade.