E.A. Bondareyko

High level of electronic components reliability maintaining allows to reduce the cost of electronic equipment. Due to the rapid development of radio electronics, appearance of new electronic components generation and permanent tightening of requirements for quality and reliability of electronic components the urgency of the reliability forecasting problem for electronic components increases at the early stages of design. At the first it is modern very large scale integrated circuit (VLSI). It is very important to predict reliability for this type of product using available information that is prediction average lifetime and failure estimate of the product. The literature analyses showed that the modern methods of the electronic components reliability prediction (including modern VLSI) are based on statistical and physical methods. The statistical method considers that the elements of the semiconductor integral circuit are formed with the help of forward and reverse bias p-n-transitions whose failure rate is equal to the failure rate of a diode. Transistor and diode p-n-transitions, in-circuit connections and pads of the package are considered to be the main unreliable components of the integral circuit. The failure rate of a chip is calculated using the failure rate of nominal transistor and diode p-n-transitions and their number. This number is equal to the total number of diodes, resistors and capacitors. But the active and passive elements of semiconductor chips are formed on the die with the help of the definite number of manufacturing operations and for this reason they cannot be considered as the discreet elements. The physical method (method of failure physics) takes into account not only the number of reliability components but also the quality of the designed topology, the number of the manufacturing operations, operation mode and maintenance impacts. The unreliability components of the integral circuit are structure elements and integral circuit structure: transistors, diodes, diffused resistors and bridges, metallization and the die, diffuse operations, area of the elements, metallization and the die. The elements of unreliability are also the package of integral circuit and connections. The analyses of the existing electronic components reliability predicting methods is performed in this research, the advantages and disadvantages of the methods are detected. Mathematical calculating models of the VLSI reliability based on statistic and physical methods are viewed. At present the technologies are rapidly evolving. The result of elaborations is a significant reduction in electronic component failures. Statistic models become more and more imprecise and for this reason these methods are used for approximate calculation of the integral circuit reliability at the stage of conceptual design. Being more complex and expensive physical methods commonly used in the process of reliability calculation at the stage of the documentation design.