The analysis and modeling of fine pitch laminate interconnect in response to large energy fault transients

摘要

In embedded applications, the miniaturization of circuitry and functionality provides many benefits to both the producer and consumer. However, the benefits gained from miniaturization is not without penalty, as the environmental influences may be great enough to introduce system failures in new or different modes and effects;Of particular interest within this research is the effect of fault transients in reduced geometries of printed circuit card interconnect, commonly referred to as fine pitch laminate interconnect. Whereas larger geometries of conductor trace width and spacing may have been immune to circuit failure at a given fault input, the reduction of the trace geometry may introduce failures as the insulating effect of the dielectric is compromised to the point where arcing occurs;To address this concern, a circuit card was designed with fine pitch laminate features in microstrip, embedded microstrip, and stripline constructions. Various trace widths and separations were tested for structural integrity (presence of arcing or fusing) at voltage extremes defined in avionics standard. The specific trace widths in the test were 4 mils, 6 mils, 8 mils, and 12 mils, with the trace separation in each case equal to the trace widths. The results of the tests and methods to artificially improve the integrity of the interconnect are documented, providing a clear region of reliable operation to the designers and the engineering community;Finally, the construction of the interconnect and the results from the test were combined to create an empirical model for circuit analysis. Created for the Saber simulator, but readily adaptable to Spice, this model will describe high-speed operation of a propagating signal before breakdown, and uses data from the experiment to calculate threshold values for the arcing breakdown. The values for the breakdown voltages are correlated to the experimental data using statistical methods of weighted linear regression and hypothesis testing.