Will ASIC Technology Demand a New Interconnection Technology instead of Soldering in Automotive Electronics? Available to Purchase

Automotive electronics, after 1995, will be similar to aerospace electronics because there is a demand for low weight and volume together with very high speed and ultra‐high reliability. What is different is that automotive electronics must achieve all those properties at very low cost. It will be shown that when using ASIC chips the chip area is determined by the number of pins instead of the number of components of the active circuit. As ASIC technology proceeds towards line widths in the submicrometre range, the ratio of active Si area to total chip area is becoming much less than 1. This means that on the ASIC chip there is Si area which is ‘empty’. This ‘empty’ Si area can be used for designing self‐test circuits and redundant functions on the ASIC chip at a cost penalty slightly higher than the design cost. It will also be shown that these ASICs can work in the order of 100 MHz at the chip level. Such ASIC chips will therefore have a very high reliability at the chip level due to the inherent properties of the Si and the built‐in redundancy. At the same time they can work at very high speed. From a performance point of view the best solution should be a highly miniaturised packaging technology. With self‐test circuits on the chip, there is a good correlation between wafer test and final test. Therefore, from an economical point of view, working with chips will then have an economical advantage compared with working with packaged circuits. From a reliability point of view it will be shown that the solder joints are the limiting factor. A critical review is presented of the reliability problems plaguing the SMD and soldering technology of today. It will be shown that, if SMD technology is to meet the reliability demands in a future automotive environment, it will have to have solder joint failure rates better than 30 ppm over the life, 17 years, in automotive applications. The conclusion is that a multi‐ASIC chip approach has the best potential as the solution for the future, post 1995, automotive electronics, provided there is a highly reliable chip interconnection technology available at that time.