Computational simulations have attracted a great deal of attention in a number of manufacturing industries. This technology is widely used in the design aspects and has been introduced to the study of reliability issues. In electronic packaging industries, many standard reliability tests have been established over the past 15 years. These tests, however, takes a long time to conduct and it is not economically viable to carry out these time consuming tests before selling the products to the market. Therefore, design for reliability (DfR) through the use of computational analyses is adopted. The basic advantage of computational analyses is it shortens the time for the investigation of the product performances at the very beginning stage by predetermining all the possible failure mechanism...[ Read more ]

Computational simulations have attracted a great deal of attention in a number of manufacturing industries. This technology is widely used in the design aspects and has been introduced to the study of reliability issues. In electronic packaging industries, many standard reliability tests have been established over the past 15 years. These tests, however, takes a long time to conduct and it is not economically viable to carry out these time consuming tests before selling the products to the market. Therefore, design for reliability (DfR) through the use of computational analyses is adopted. The basic advantage of computational analyses is it shortens the time for the investigation of the product performances at the very beginning stage by predetermining all the possible failure mechanisms in the prolong operations.

Accelerated temperature cycling test is one of the reliability testing methods for qualifying the electronic products. However, performing such tests for verifying different product designs usually time consuming. In this study, based on the data from previous research works, finite element analyses are carried out to simulate the accelerated thermal cycling tests. Comparison between computational and experimental approaches is made and it is suggested that 2-D finite element analysis is sufficient for estimating the thermal fatigue life of PBGA assemblies with the advantages of reduced time and effort in model building.

In addition, thermal cycling test involves many uncertainties. For example, test results obtained fiom different ATC machines may not be consistent since the conditions provided may vary. Therefore, parametric studies on the effects of various conditions during ATC tests are investigated by the computational method. During actual thermal cycling tests, the ideal temperature profile with four linear segments as specified in the industrial standard is usually difficult to achieve. Non-linear ramping rate in the profile, however, is the common observation when the temperature approaches extreme values. In this study, finite element analyses have been applied to find out the effects of such profile. Also, there is always a concern about the sudden break down of the thermal cycling machine during ATC tests. FE simulations have been carried out by introducing a period of service interruption in the middle of the thermal cycling test. Suggestions have been provided for the possibility of the resumption of the ATC tests when encountering this kind of machine breaks down. Stress free temperature of ATC test is another controversial topic in simulation which may also be a factor influences the test results. So, different starting temperatures have been investigated.

Finite element analysis of the failure prediction of passive components under depaneling condition has also been proposed in this research work. Correlation between the bending strain on the PCB (which is an index of the local curvature of the bent PCB) and the bending stress in the passive components (which is the reason to crack capacitors/resistors) is established. Discussions are given on the modeling and the methodology of obtaining component stress distribution plots on PCBs. The outputs of this part of the thesis can provide a guiding rule to avoid mounting passive components in the "high risk" areas on the PCB at the design stage.