DonNTU   Masters' portal

Abstract

Content

Introduction

An important component of the operations with computer technology is the maintenance, including diagnostics and troubleshooting. Now in today's digital devices are increasingly used testability design, which allows you to automate testing and troubleshooting. However, Ukraine has a large number of obsolete computer equipment, which upgraded or replaced by a new is very expensive. There is a need to diagnose and repair it. Objectives for the diagnosis of this technique solved the "classical" methods, thus requiring specialized diagnostic systems. For such a large-scale application of diagnostic needs an optimal algorithm for finding fault, but an analysis of domestic and foreign literature shows a practical lack thereof. This is due to complexity of developing such algorithms.

Formulation of the problem

For an extensive and widespread use of diagnostics requires an optimal algorithm for troubleshooting , but the analysis of domestic and foreign literature shows a practical lack thereof. In the area of ​​probe diagnostics systems remains unresolved a few problems, among which should first identify the problem of minimizing the permutations of the probe (also duration of the process of diagnosis), fault isolation, up to a removable components and the problem of minimizing external memory of these systems. Is urgent to find solutions to these problems in order to obtain sufficiently generic algorithm for troubleshooting. Experts in the field of diagnostics for a long time doing research in this direction. Particular attention is paid to probe the troubleshooting process and how the algorithm itself, "the slave probe". It also focuses on the optimization problems solved in the probe diagnostics on various criteria, mainly to reduce the number of permutations of the probe. To solve this problem there are different approaches. Purpose of the development of an algorithm for the probe troubleshooting digital circuits, taking into account the technological features of the real circuits, and optimization of the average number of permutations of the probe.

Expected scientific novelty and practical value

The problem of finding a universal algorithm for troubleshooting and minimize the number of permutations of the probe is not new, but obviously the lack of optimal solutions in this area. In this paper author proposed to develop an improved methodology for the algorithm, "slave probe" in particular, with the possibility of discharge from the facility diagnostics units with local feedback contours. This allow to take a group of elements with local feedback as an element that can be called a pseudoelement.

If the developed algorithm became quite successful, its use in the probe diagnostics bring significant benefits to the time of testing real computing devices, integrated circuits and other equipment.

Review of research and development

The development of this direction is engaged in the Donetsk National Technical University by Associate Professor of Computer Engineering, Ph.D. Yuri Zinchenko. Results of his development and research activities can be found here. Under the leadership of Yuri Zinchenko was also written several works of masters, each of which contains an extensive overview of the topic and offers a solution to the problem of the lift. Abstracts in this work can be found in Master's work Links.

Troubleshooting and diagnostic systems

Technical diagnostics of computing devices (CD) solves the following main tasks: 

Depending on the degree of automation of solving the problem of the second type of diagnosing CD divided into two main classes:

Automated systems perform troubleshooting by identifying the faulty state of the object of diagnosis (OD), through its external (boundary) and comparing the findings of this state with pre-built so-called diagnostic fault dictionary. Diagnostic each fault dictionary of OD associates with the test reaction, which is obtained by modeling this state on a given OD test [8].

With eliminating human intervention, automated system designed to completely automate the process of troubleshooting. However, this ideal goal was basically impossible. The impossibility to automatically search fault in CD because, as studies have shown that the definition of the source fault of digital devices up to a separate removable component is a task of insurmountable difficulties [2].

Proved to be impractical as the idea is to use a dictionary shooting, because firstly, the model is stuck-at faults, which is determined with respect to vocabulary, have served to gate circuits, is inadequate for today's complex clause. Second, the length of the test of modern CD reaches a million or more vectors, which casts doubt on the possibility of modeling faults at this test. Therefore, automated systems are used for the diagnosis of combinational and sequential simple CD [7–10].

The failure of automated systems has led to the necessity of building automation, ie allowing intervention in the diagnostic process rights, systems troubleshooting. Human intervention, referred to in such systems is reduced to an active or passive exposure (sensing) the internal control points OD, so called probe system, and the process of finding fault with the system - the search probe or probe diagnostics (PD)[3–5].

Troubleshooting by the method of "slave probe" based on "tracking" path fault manifestations of the topological scheme of OD. Tracking is carried out probing the internal control points the way to contact the output jack, which revealed a fault to the site where her source, and comparison with the real test reactions with reference values. The reference test reaction is evaluated by simulation facility in good condition or removed in advance with a known good (reference) sample OD [6].

Expected results

In this master's work author proposed to develop an improved methodology for the algorithm, "slave probe" in particular, with the possibility of discharge from the object of diagnostics units of local feedback contours. This allow to take a group of elements with local feedback as pseudoelement. To find a fault in the combinational circuit can be obtained by the algorithm GALOP, which requires a less number of permutations of the probe, rather than SCAN, used in circuits with feedback. If the fault has been detected in the pseudoelement, it is obvious that she was in a group of elements with feedback. For this group will apply the algorithm SCAN.

For clarity, consider the example in Figure 1. A trigger is a group of elements that have local feedback. The scheme is weakly consistent. To scan and diagnose the scheme by the method of "slave probe" will need to use the algorithm SCAN.

Insert of pseudoelement

Figure 1 – Insert of pseudoelement
(animation: 4 frames, 5 cycles of repeating, 105 Kb)

To prevent these complications make the change of the local group of elements with feedback on the pseudoelement having similar inputs and outputs. We obtain a diagram of the facility of diagnosis, which has no feedback, is a combination and can be processed faster algorithm GALOP.

Conclusion

Work in a designated area is associated with the development and testing algorithm with an emphasis on finding and replacing elements of the groups with local feedback wint the equivalent pseudoelemen. If the developed algorithm became quite successful, its use in the probe diagnostics bring significant benefits to the time of testing real computing devices, integrated circuits and other equipment.

This master's work is not completed yet. Final completion: December 2012. The full text of the work and materials on the topic can be obtained from the author or his scientific adviser after this date.

References

  1. Малышко, Ю. В., Чипулис, В. П., Шаршунов, С. Г. Автоматизация диагностирования электронных устройств [Текст] / Ю. В. Малышко, В. П. Чипулис, С. Г. Шаршунов. // Москва : Энергоатомиздат, 1986. – 216 с.
  2. Горяшко, А. П. Синтез диагностируемых схем вычислительных устройств [Текст] / А. П. Горяшко // Москва : Наука, 1987. – 288 с.
  3. Зинченко, Ю. Е., Козинец, А. М., Жилин, К. Н. Проблемы зондового поиска неисправностей и пути их разрешения [Текст] / Ю. Е. Зинченко, А. М. Козинец, К. Н. Жилин // Сборник трудов Донецкого государственного технического университета. Серия: Информатика, кибернетика и вычислительная техника, выпуск 6. – Донецк : ДонГТУ, 1999. – С. 212-217.
  4. Воротынцев, Н.В., Зинченко, Ю.Е. Поиск неисправностей цифровых ТЭЗ по алгоритму ведомого зонда со «статической оптимизацией» [Текст] / Воротынцев Н.В., Зинченко Ю.Е. // Сборник трудов первой международной студенческой научно-технической конференции – 15 декабря 2005 г. – Донецк : ДонНТУ, 2005. – С. 321-322.
  5. Деменко, А.Г., Ханаев, В.В., Зинченко, Ю. Е., Зинченко, Т. А. Проблемы минимизации количества перестановок зонда в процессе поиска неисправностей по методологии «ведомого зонда» [Текст] / Деменко А.Г., Ханаев В.В., Зинченко Ю. Е., Зинченко Т. А. // Сборник трудов III научно-технической конференции студентов, аспирантов и молодых ученых – 16-18 апреля 2012 г. – Донецк : ДонНТУ, 2012.
  6. Зинченко Ю. Е. – Научные интересы и разработки [Электронный ресурс]. – Режим доступа: http://hardclub.donntu.ru/zinchenko/science.htm
  7. Молов В.К., Тарасенко В.П. Структурно-аналитический метод поиска неисправностей [Текст] / Молов В.К., Тарасенко В.П. // Автоматика и вычислительная техника. – 1984., с. 72-78.
  8. Зинченко Ю.Е., Хатейт Ю. Стратегия поиска неисправностей системы зондовой диагностики [Текст] / Зинченко Ю.Е., Хатейт Ю. // Теоретическая и прикладная информатика. Науч.-техн. конф.. Тезисы докладов. – Донецк, 1993. – С. 59.
  9. Тестово-Диагностический Комплекс – Автоматизированная система диагностики ТЕСТ-Д для ремонта электронных промышленных устройств [Электронный ресурс]. – Режим доступа: http://mages.chat.ru/tdk3e.htm
  10. Бухтеев А. – Методы и средства проектирования систем на кристалле [Электронный ресурс]. – Режим доступа: http://www.chipinfo.ru/literature/chipnews/200304/1.html