| Field Report : Top : Evaluation validity in systems with continuous, automatic measurement |
| |
 |
| Evaluation validity in systems with continuous, automatic measurement |
|
|
|
| |
 |
 |
|
| |
 |
Evaluation validity in systems with continuous, automatic measurement
|
 |
 |
|
| |
|
Kunikazu Ishii *
Rapid strides are being made in miniaturizing electronic equipment while simultaneously offering improved functions and performance. Companies are under pressure to reduce product development time, while equal pressures demand that product safety and reliability be maintained. To handle these opposing demands, both the effectiveness and the accuracy of environmental testing require further improvement. To meet such requirements, we at ESPEC CORP are incorporating measurement of electrical characteristics in our environmental test equipment. This arrangement enables us to continually measure specimen characteristics under environmental conditions, and thus grasp real time characteristics when determining malfunction modes. We are currently developing a full line of automatic measuring systems. This report will introduce some aspects of Evaluation Systems and discuss the effectiveness of automatic measurement systems.
|
 |
 |
|
|
| |
|
|
| |
To produce ever-more-compact electronic equipment, companies are striving to miniaturize chip sizes, especially for chips used in semiconductor IC packages and electronic parts. Complicating the problem has been rapid pace of the development of build-up circuit boards *1 , which make high-density mounting possible. Studies of new methods and materials for connecting IC packages are also proceeding at a fever pitch. To further complicate matters, developers are not only working under the constraints of market competitiveness with regard to cost and functions, but they are also working under the constraints of global environmental protection and societal regulations, such as using Pb-free solder *2 , meeting VOCs regulations *3 for flux, and developing environmentally friendly printed circuit boards. While handling such a wide variety of requirements, developers must also shorten the development time while maintaining product reliability. To meet these conditions, developers must have a means of accurately and more effectively assessing safety and reliability.
At ESPEC CORP, in addition to checking functions and evaluating reliability, we are combining our widely used environmental equipment, and while running environmental tests we are continuously measuring electrical characteristics of specimens. By extracting real time data, we are developing automatic measuring systems that are able to detect malfunctions and failures. In this report, I would like to present some real world examples of evaluations involving continuous measurement of electrical characteristics of specimens in the environments used for these tests. I would also like to discuss the effectiveness of these measurement systems.
|
|
| |
 |
| 2. How evaluation needs to be done |
|
|
|
|
|
| |
To guarantee reliability of products with complicated advanced functions while covering a wide variety of development requirements in less time than ever before, developers must be able to improve the effectiveness and accuracy of product evaluation. Fig. 1 presents some suggestions for methods to improve the effectiveness and the accuracy of evaluation.
2-1 Evaluating more effectively
To evaluate more effectively, first of all, the time required to make evaluations must be shortened. Next, test times must be reduced by cutting down the time taken for the test operation itself as well as by revising the determination methods for evaluation. For example, when running a product life test, in the most general evaluation method, failure and relative advantages of different specimens is extrapolated from data measured at specific intervals. In this case, if evaluation could be made with data from shorter measurement intervals, it would be possible to assess product life and malfunctions more quickly and shorten test times. Also, applying accelerated life testing would be very effective in reducing test times. At this point, I would like to discuss reducing test times, and making more efficient evaluations based on data from continuous measurements and more efficient test operations.
2-2 Evaluating more accurately
Current products boasting advanced functions with high-density mounting are experiencing failure modes that did not occur in previous products. These advances mean that even slight changes in characteristics can lead to failure, and so current evaluation practices must be much more strict. To improve evaluation accuracy, there is greater need to be more precise in capturing failure phenomena, and a concurrent need to increase the amount of data used for analysis.
In other words, it is difficult to determine whether failures only occur in specific environments and whether failure will occur unless the product is in such a specific environment. Reproducibility (the ability to predict test results) is also difficult to obtain. For example, continuously measuring specimen characteristics as well as selecting measurement methods that are appropriate to the characteristics of failure modes occurring in the test environment used both are indispensable for detecting insulation deterioration due to ion migration, defective contact due to differences in coefficients of thermal expansion, and fluctuations in resistance values due to cracking.
To clarify failure modes and to more accurately predict product life, we must be able to accurately detect not only the fact of failure occurrence but also the time of occurrence. We must also be able to make our analyses based on data that can contrast fluctuation in specimen characteristics with such environmental conditions as temperature and humidity. We must further maintain a large enough quantity of data to enable us to make statistical determinations that consider dispersion. Since the time of running the test does not change, if we can even slightly increase the amount of data used for analysis, we will be able to obtain more accurate evaluation results.
|
|
 |
| |
|
Fig. 1 Requirements for evaluation
|
|
|
| |
|
|
|
|
|
|
