Accelerated Testing for the Selection of Most Reliable MicroVia Design

Charles Recchia - MACOM and Milena Krasich - Raytheon

In modern circuit card designs, micro vias are the most common connection for the electronic components. The intricate details in the design, the small sizes along with the type of materials used, make the micro vias susceptible to cracks due to stresses encountered in product use. This paper describes in detail the case of development of micro cracks in micro vias of a newly designed system during the extended operational tests and the accelerated test which was designed to select the best new micro via design solution to prevent their appearance for the duration of the predetermined useful life of the product. The paper also outlines the test data analysis with the original unique approach for selection of the best design solution. The micro crack propagation through micro vias compromised connections of high power amplifier components and the operability of the assembly. Given the high number of the specific connections in the system (approximately 40,000) the probability of some or many of them failing during the 30 year life of the system was unacceptably high. The re-design effort considered several different design configurations and materials to correct the problem, but for the selection of the viable solution was needed for a sufficient confidence that the connections would not fail at some time during the system life. To select the best design solution and to gain this assurance, the accelerated testing of the coupons containing the selected proposed design solutions was prepared and carried out. The accelerated testing comprised of the environmental and use stresses which were identified as the most likely cause of development and propagation of micro cracks in vias or any connections, the thermal cycling and the thermal exposure (dwell). The thermal cycles in use from diurnal/nocturnal temperature changes and from the ON/OFF operation in various seasons and climates were recalculated to a number of thermal cycles with a single thermal variation which was then accelerated for the tests. Similar recalculations were done for the life duration of thermal exposures in ON and OFF conditions. To simulate life and stress synergism, thermal exposures were distributed over the number of thermal cycles. To determine the values for the constants in thermal cycling and dwell acceleration factors, the coupons were first tested with two different temperature ranges and the dwell temperatures. Test data consisted of cycles to failure or survival of the test cycles without a failure. The failure criterion was the resistance value of the vias connected in series on each coupon. The original and new approach in data analysis was that the strength/stress criteria (fatigue strength as number of survived test cycles recalculated to the use conditions, vs. required cumulative fatigue stress - the number of required cycles in use with added margin for reliability demonstration) were used to determine failure rates of vias in each assembly in each of the coupons (proposed design solution). Those failure rates were then compared to their failure rate allocation and the design solutions with the acceptable failure rates were considered candidates for the solution