I.L. Akhramovich - Dr.Sc. (Eng.), Professor, General Director of Interregional Association "Morinformsystem", Quality Deputy Technical Director of JSC "Concern "Morinformsystem-Agat?, Member of International Information Academy and Russian Academy of Reliability
S.A. Kogut - Head of Reliability Department, JSC "Concern "Morinformsystem-Agat -
F.V. Tereshchenko - Leading Engineer of Reliability Department, JSC "Concern "Morinformsystem-Agat?. E-mail: luxma@rambler.ru

When a technical all requirement document on estimation of any device reliability is issued, it should necessarily contain the value of ambient temperature around the component. This is the so called total (common) index, set as a rule according to the regulatory documentation. However it is not entirely reasonable to consider total device temperature or temperature of the whole system. It is obvious that the temperature of different kinds of electronic components may vary, and quite often the temperature values in a device may differ by several times or even orders. In the course of the test it was found out how in particular the specified electric component failure density depends on the increase of the component ambient temperature. At the initial stage the exploitation failure density values of various kinds of electronic components under different ambient temperature conditions were calculated. During the research, the experiments (temperature tests) were conducted, which allowed to determine the practical pattern for temperature levels of the components included in the articles under development. The statistics derived from the experimental data shows that total device temperature or temperature of the whole system does not reflect the real picture of component heating. The calculated reliability values turn to be either understated, which is undesirable (in case high temperature values for calculation were specified), or overstated, which can be quite dangerous (in case low temperature values for calculation were specified). Considering the given calculated and experimental data it is concluded that it is reasonable enough to conduct temperature tests with the biggest possible number of electronic components of different kinds. The estimation of temperature values of electronic components in operational modes will help OEMs to increase the calculating accuracy of article reliability, to optimize the configuration of SPTA sets, to receive initial data for the development of methods for rapid reliability tests.