- Email: [email protected]
A reliability growth model
World Abstracts on Microelectronics and Reliability
Thermal stress in bonded joints. W. T. CHEN and C. W. NELSON. I B M d. Res. Dev. 23, (2) 179 (March 1979). This paper considers the stress distributions in bonded materials induced by differential expansion or contraction of these materials. The analytical approach is similar to the lap joint theories attributed to Volkersen and expanded by Goland and Reissner. Several simple and typical analytical models are presented to bring out the relative importance of different geometrical and material parameters and to give some insight into different modes in which the bonds might fail.
Causes of wavesoldering defects. CASIMIR D. BERNARD. Electron. Prod. p. 35 (June 1979). This article is concerned with what meets the eye as the wavesoldered printed circoit
411
board emerges from the wavesoldering machine. Successful wavesoldering of a printed circuit board depends upon the pcb being : • Well designed for wavesoldering. • Well prepared for wavesoldering. • Processed under optimum conditions. • Wavesoldered on the most appropriate wavesolder machine. Common defects of wavesoldered pcb's are: • No solder joint or only a partial one. • Solder bridging two or more conductors or pins. • Solder may protrude from the board as a sharp iciclelike spike. • The board may be covered with a webbing of solder.
3. CIRCUIT AND SYSTEMS RELIABILITY, MAINTENANCE AND REDUNDANCY A design review approach toward dynamic RAM reliability. J. A. ROBERTS. Microelectron. Reliab. 19, 97 (1979). For many equipment manufacturers the selection, test and use of Dynamic Random Access Memories (RAM) has become the prime component problem. Many semiconductor companies have been attracted by the vast market size and this has resulted in a large number of design attempts. Cost, market and performance pressures have forced semiconductor manufacturers into producing designs which have not been fully developed. This paper describes design review techniques that have been used to identify design weaknesses.
Beurteilung der Funktionssicherheit von Nachrichtennetzen mit mehreren Hierarchieebenen (Estimation of the reliability of hierarchical communication networks). GERD MROZYNSKI. Frequenz 33,(3) 62 (1979). (In German.) A comparison of communication networks with regard to their vulnerability through disturbances caused by failure of system components and cable flaws requires a criterion independent of the special structure of the network. This criterion should take into account the traffic load and the traffic distribution of connected sources and sinks. A measure of system failure is introduced and the results are presented by evaluating special examples. As a final conclusion it can be said that under given conditions a bidirectional double ring network is least vulnerable to cable flaws which are responsible for the majority of network disturbances.
An improved algorithm for reliability optimization. KRISHNA GOPAL, K. K. AGGARWAL and J. S. GUPTA. IEEE Trans. Reliab. R-27, (5) 325 (December 1978). This paper proposes a modification of the Aggarwal et al. algorithm for reliability optimization by introducing a new heuristic criterion for selecting the subsystem where redundancy is to be added. This criterion accounts for the relative decrement in unreliability versus the largest of the relative increments in resources. The method applies to multiple separable constraint problems (which need not be linear) and to systems which may be complex or series. The method is simple, fast, and easily programmed. The results are compared with those of the Aggarwal et al. algorithm and are better in many problems.
Die Erkennung latenter Defekte in Sehaltungen durch eine Phasenrausehmessung tion of latent defects in electronic circuits phase noise). HORST RODER. Frequenz 33, (4)
elektronischen (The recogniby measuring
101 (1979). (In German.) In most cases the test methods, which are predominantly used in quality control of electronic circuits at present, do not allow to recognize those hidden defects which lead to failure only during operation. In this paper, both a method for the measurement of phase noise and
the corresponding equipment are described which enables us to detect circuits containing such defects. Experimental results and error analyses on thick-film electronic circuits as well as high-voltage stress investigations on CMOS circuits show the power of this new experimental technique.
The logistics of life cycle cost. JOHN R. PERONNET. Microelectron. Reliab. 19, 23 (1979). Program Support has always been a multifaceted discipline and therefore has exhibited a sensitivity to both the range and depth of available program planning data upon which to project downstream logistic consequences. This paper addresses the key aspects of Life Cycle Cost (LCC) methodology synthesis and highlights the ease of program application and integration on such programs as the Canadian New Fighter Aircraft (NFA) concept. Also discussed in the paper is the interrelationship between hardware design, ILS resource requirements and program cost impacts of prime support postures as well as alternate support plan concepts. The illustrations used are drawn from actual Aircraft Program and Ground Based Electronic System applications and, in summary, the paper highlights the effectiveness of this new technology and points out its front end applicability to new programs.
lnteraetiom of industrial integration and standardization. CHARLES A. MILLS. Qual. Prog. p. 22 (February 1979). True product integrity results from an integrated program, involving each segment of an organization.
Developing the sampling plans. ALBERT H. BOWKER. Qual. Prog. p. 20 (February 1979). The 1977 Shewhart medalist talks about pioneering efforts in applying mathematical statistics to solve sampling inspection problems. Interactive graphics in new product quality assurance programs. JOHN A. CLEMENTS. Qual. Prog. p. 14 (February 1979). A special department at Gillette uses an effective method to manage and report vast quantities of quality control information. A reliability growth model. G. L. SRIWASTAV.IEEE Trans. Reliab. R-27, (5) 306 (December 1978). A system can be in one of two states: perfect (so) or imperfect (sl). At any trial, the system can succeed or fail if it is in sl, but will never fail if in So. After every failure, some corrective action is taken which, with a chance ct, will lead to a transition to so. Even if the corrective action does not succeed, i.e. the system is still in sl, the probability of failure reduces by a constant multiple at the next trial. Expressions for the probability that the system will be in sl at trial n and the unconditional probability that the system fails in trial n, are
412
World Abstracts on Microelectronics and Reliability
derived. The latter can be expressed in terms of the former. These probabilities depend not only on the number of successes in n trials but on their sequence.
Component replacement and the use of Relcode. N. A. J. HASTINGS and A. K. S. JARDINE. Microelectron. Reliab. 19, 49 (1979). The paper reviews briefly the standard age and block replacement models for setting component replacement rules, highlighting the importance of correct identification of the failure distribution of the component. The main focus of the paper is on procedures for identification of the failure distribution. Three procedures are covered : (i) classical statistical distribution fitting (ii) use of probability paper (particularly Weibull) to estimate the parameters of the failure distribution (iii) use of a computer package Relcode. Of the above procedures emphasis is on use of the Relcode computer software package which is used for the analysis of data relating to the failures and successful performance of machines or components. The analyses range from simple calculations such as mean life, to the fitting of the Weibull and other distribution models, and renewal and planned replacement analyses. Relcode may use data arising from reliability trials or from other day-to-day records of failure and successful performance of the items concerned. The package allows for the possibility of multiple censored data, that is items (called suspended items) which perform successfully without failure--as well as for items which fail.
Selling reliability engineering to the company. M. B. DARCH. Microelectron. Reliab. 19, l (1979). Reliability engineering is a relatively new field of engineering that blossommed into prominence with the demands of the space program and military procurement. Current consumer awareness is leading to an increasing demand that reliability play a much larger role in consumer goods. For this to happen, the reliability engineer must understand his role in the design process and convince management within the company of the benefits to be derived from better reliability and maintainability. On fault trees and other reliability evaluation methods. BALBIR S. DHILLON and C. SINGH. Microelectron. Reliab. 19, 57 (1979). This paper provides a brief description and comparison of fault trees and several other reliability evaluation methods. It is expected that this material will be useful to reliability engineers in providing a perspective on various methods of reliability analysis. Reliability learning model: Application to color TV. J. IMAI, H. KARASAWAand H. MACHIDA. Microelectron. Reliab. 19, 73 (1979). This paper discusses reliability learning for consumer color TV receivers, categorizing learning by product model. Each product model has a hierarchical construction involving various levels, and learning occurs at each level. The characteristics of the convergence point for a chassis category a r e studied. Application of this learning model to practical reliability activity is explained, methods related to this application are indicated, and the process of total reliability learning is explained. Particularly, the B.M.B method, which indicates inherent reliability capability, and the idea of design review to implement this method at the present level of theoretical understanding is indicated. Stochastic behaviour of some 2-unit redundant systems. P. K. KAPUR and K. R. KAPOOR. IEEE Trans. Reliab. R-27, (5) 382 (December 1978). This paper discusses the stochastic behaviour of some 2-unit cold-standby redundant system under most generalised assumptions of arbitrary failure and repair time distributions and derives various reliability characteristics e.g.,
(i) Distribution of First-passage time to system failure, (ii) Point availability and unavailability, (iii) s-Expected number of system failures, disappointments, interferences, etc. Models discussed are : (1) (II) (III) (IV)
Intermittently available redundant system, Intermittently used system, Redundant system with repair and post-repair, Repair limit suspension model.
We employ Markov Renewal Processes technique to study the above systems.
Eigenvalue approach for computing the reliability of Markov systems. LEIF KANDERHAG. IEEE Trans. Reliab. R-27, (5) 337 (December 1978). This paper solves the computation of the reliability of Markov systems. The solution is especially useful for systems where the number of states is large. Not only is the mean time-to-failure computed but an analytic expression is given for the probability of the system's being in a certain state as a function of time. The solution is derived by solving the eigenvalues and eigenvectors of the transition-rate matrix of the Markov system. In the solution, numerical algorithms have been chosen to keep the computer cost down. The method uses just a fraction of the time that, for example, the Runge-Kutta method uses. Empirical Bayes estimators of reliability for Iognormal failure model. W. J. PADGETT and J. A. ROBINSON. IEEE 7tans. Reliab. R-27, (5) 332 (December 1978). Empirical Bayes (EB) procedures are considered for estimating the reliability R(t ; y, tr) = gaufc[(ln t - y)/~] for the lognormal failure model. EB estimators are obtained for the 2 cases: (i) y is unknown and a is known, and both 7 and a are unknown. The empirical Cdf of the maximum likelihood estimators of the parameters is used to obtain the EB estimators. (ii) A smooth EB estimator of R(t; 7, o-) is developed when 7 is unknown and cr is known. A modification of this estimator is proposed for both y and 6 unknown. In both cases, EB estimators are obtained for complete samples at each testing stage. Monte Carlo simulations are presented to compare the EB estimators and the maximum likelihood (ML) estimators of R(t ; 7, 6). The simulations indicate that the smooth EB estimators have smaller mean squared errors than the other EB estimators or the ML estimators. Computation-based reliability analysis in real time applications. A. PEDAR and V. V. S. SARMA. IEEE Trans. Reliab. R-27, (5) 345 (December 1978). Reliability analysis for computing systems in aerospace applications must account for actual computations the system performs in the use environment. This paper introduces a theoretical nonhomogeneous Markov model for such applications. Determinations of MTTF from the transition probability matrix and diagram. J. M. KONTOLEON. IEEE Trans. Reliab, R-27, (5) 347 (December 1978). In designing complex repairable systems it is often important to assess system MTTF for a variety of component reliability characteristics and repair capabilities. Two procedures are given for systems for which a state transition diagram is possible. The first allows such calculation to be performed iteratively without solving each time the system of state equations from the beginning. The approach involves matrix manipulations which split the matrix formulation into two parts; one contains the solution from the previous iteration (at the beginning, the MTTF is calculated in the usual manner) and the other incorporates the changes in the transition probabilities. The second procedure allows the determination of MTTF by decomposing the transition probability
Thermal stress in bonded joints. W. T. CHEN and C. W. NELSON. I B M d. Res. Dev. 23, (2) 179 (March 1979). This paper considers the stress distributions in bonded materials induced by differential expansion or contraction of these materials. The analytical approach is similar to the lap joint theories attributed to Volkersen and expanded by Goland and Reissner. Several simple and typical analytical models are presented to bring out the relative importance of different geometrical and material parameters and to give some insight into different modes in which the bonds might fail.
Causes of wavesoldering defects. CASIMIR D. BERNARD. Electron. Prod. p. 35 (June 1979). This article is concerned with what meets the eye as the wavesoldered printed circoit
411
board emerges from the wavesoldering machine. Successful wavesoldering of a printed circuit board depends upon the pcb being : • Well designed for wavesoldering. • Well prepared for wavesoldering. • Processed under optimum conditions. • Wavesoldered on the most appropriate wavesolder machine. Common defects of wavesoldered pcb's are: • No solder joint or only a partial one. • Solder bridging two or more conductors or pins. • Solder may protrude from the board as a sharp iciclelike spike. • The board may be covered with a webbing of solder.
3. CIRCUIT AND SYSTEMS RELIABILITY, MAINTENANCE AND REDUNDANCY A design review approach toward dynamic RAM reliability. J. A. ROBERTS. Microelectron. Reliab. 19, 97 (1979). For many equipment manufacturers the selection, test and use of Dynamic Random Access Memories (RAM) has become the prime component problem. Many semiconductor companies have been attracted by the vast market size and this has resulted in a large number of design attempts. Cost, market and performance pressures have forced semiconductor manufacturers into producing designs which have not been fully developed. This paper describes design review techniques that have been used to identify design weaknesses.
Beurteilung der Funktionssicherheit von Nachrichtennetzen mit mehreren Hierarchieebenen (Estimation of the reliability of hierarchical communication networks). GERD MROZYNSKI. Frequenz 33,(3) 62 (1979). (In German.) A comparison of communication networks with regard to their vulnerability through disturbances caused by failure of system components and cable flaws requires a criterion independent of the special structure of the network. This criterion should take into account the traffic load and the traffic distribution of connected sources and sinks. A measure of system failure is introduced and the results are presented by evaluating special examples. As a final conclusion it can be said that under given conditions a bidirectional double ring network is least vulnerable to cable flaws which are responsible for the majority of network disturbances.
An improved algorithm for reliability optimization. KRISHNA GOPAL, K. K. AGGARWAL and J. S. GUPTA. IEEE Trans. Reliab. R-27, (5) 325 (December 1978). This paper proposes a modification of the Aggarwal et al. algorithm for reliability optimization by introducing a new heuristic criterion for selecting the subsystem where redundancy is to be added. This criterion accounts for the relative decrement in unreliability versus the largest of the relative increments in resources. The method applies to multiple separable constraint problems (which need not be linear) and to systems which may be complex or series. The method is simple, fast, and easily programmed. The results are compared with those of the Aggarwal et al. algorithm and are better in many problems.
Die Erkennung latenter Defekte in Sehaltungen durch eine Phasenrausehmessung tion of latent defects in electronic circuits phase noise). HORST RODER. Frequenz 33, (4)
elektronischen (The recogniby measuring
101 (1979). (In German.) In most cases the test methods, which are predominantly used in quality control of electronic circuits at present, do not allow to recognize those hidden defects which lead to failure only during operation. In this paper, both a method for the measurement of phase noise and
the corresponding equipment are described which enables us to detect circuits containing such defects. Experimental results and error analyses on thick-film electronic circuits as well as high-voltage stress investigations on CMOS circuits show the power of this new experimental technique.
The logistics of life cycle cost. JOHN R. PERONNET. Microelectron. Reliab. 19, 23 (1979). Program Support has always been a multifaceted discipline and therefore has exhibited a sensitivity to both the range and depth of available program planning data upon which to project downstream logistic consequences. This paper addresses the key aspects of Life Cycle Cost (LCC) methodology synthesis and highlights the ease of program application and integration on such programs as the Canadian New Fighter Aircraft (NFA) concept. Also discussed in the paper is the interrelationship between hardware design, ILS resource requirements and program cost impacts of prime support postures as well as alternate support plan concepts. The illustrations used are drawn from actual Aircraft Program and Ground Based Electronic System applications and, in summary, the paper highlights the effectiveness of this new technology and points out its front end applicability to new programs.
lnteraetiom of industrial integration and standardization. CHARLES A. MILLS. Qual. Prog. p. 22 (February 1979). True product integrity results from an integrated program, involving each segment of an organization.
Developing the sampling plans. ALBERT H. BOWKER. Qual. Prog. p. 20 (February 1979). The 1977 Shewhart medalist talks about pioneering efforts in applying mathematical statistics to solve sampling inspection problems. Interactive graphics in new product quality assurance programs. JOHN A. CLEMENTS. Qual. Prog. p. 14 (February 1979). A special department at Gillette uses an effective method to manage and report vast quantities of quality control information. A reliability growth model. G. L. SRIWASTAV.IEEE Trans. Reliab. R-27, (5) 306 (December 1978). A system can be in one of two states: perfect (so) or imperfect (sl). At any trial, the system can succeed or fail if it is in sl, but will never fail if in So. After every failure, some corrective action is taken which, with a chance ct, will lead to a transition to so. Even if the corrective action does not succeed, i.e. the system is still in sl, the probability of failure reduces by a constant multiple at the next trial. Expressions for the probability that the system will be in sl at trial n and the unconditional probability that the system fails in trial n, are
412
World Abstracts on Microelectronics and Reliability
derived. The latter can be expressed in terms of the former. These probabilities depend not only on the number of successes in n trials but on their sequence.
Component replacement and the use of Relcode. N. A. J. HASTINGS and A. K. S. JARDINE. Microelectron. Reliab. 19, 49 (1979). The paper reviews briefly the standard age and block replacement models for setting component replacement rules, highlighting the importance of correct identification of the failure distribution of the component. The main focus of the paper is on procedures for identification of the failure distribution. Three procedures are covered : (i) classical statistical distribution fitting (ii) use of probability paper (particularly Weibull) to estimate the parameters of the failure distribution (iii) use of a computer package Relcode. Of the above procedures emphasis is on use of the Relcode computer software package which is used for the analysis of data relating to the failures and successful performance of machines or components. The analyses range from simple calculations such as mean life, to the fitting of the Weibull and other distribution models, and renewal and planned replacement analyses. Relcode may use data arising from reliability trials or from other day-to-day records of failure and successful performance of the items concerned. The package allows for the possibility of multiple censored data, that is items (called suspended items) which perform successfully without failure--as well as for items which fail.
Selling reliability engineering to the company. M. B. DARCH. Microelectron. Reliab. 19, l (1979). Reliability engineering is a relatively new field of engineering that blossommed into prominence with the demands of the space program and military procurement. Current consumer awareness is leading to an increasing demand that reliability play a much larger role in consumer goods. For this to happen, the reliability engineer must understand his role in the design process and convince management within the company of the benefits to be derived from better reliability and maintainability. On fault trees and other reliability evaluation methods. BALBIR S. DHILLON and C. SINGH. Microelectron. Reliab. 19, 57 (1979). This paper provides a brief description and comparison of fault trees and several other reliability evaluation methods. It is expected that this material will be useful to reliability engineers in providing a perspective on various methods of reliability analysis. Reliability learning model: Application to color TV. J. IMAI, H. KARASAWAand H. MACHIDA. Microelectron. Reliab. 19, 73 (1979). This paper discusses reliability learning for consumer color TV receivers, categorizing learning by product model. Each product model has a hierarchical construction involving various levels, and learning occurs at each level. The characteristics of the convergence point for a chassis category a r e studied. Application of this learning model to practical reliability activity is explained, methods related to this application are indicated, and the process of total reliability learning is explained. Particularly, the B.M.B method, which indicates inherent reliability capability, and the idea of design review to implement this method at the present level of theoretical understanding is indicated. Stochastic behaviour of some 2-unit redundant systems. P. K. KAPUR and K. R. KAPOOR. IEEE Trans. Reliab. R-27, (5) 382 (December 1978). This paper discusses the stochastic behaviour of some 2-unit cold-standby redundant system under most generalised assumptions of arbitrary failure and repair time distributions and derives various reliability characteristics e.g.,
(i) Distribution of First-passage time to system failure, (ii) Point availability and unavailability, (iii) s-Expected number of system failures, disappointments, interferences, etc. Models discussed are : (1) (II) (III) (IV)
Intermittently available redundant system, Intermittently used system, Redundant system with repair and post-repair, Repair limit suspension model.
We employ Markov Renewal Processes technique to study the above systems.
Eigenvalue approach for computing the reliability of Markov systems. LEIF KANDERHAG. IEEE Trans. Reliab. R-27, (5) 337 (December 1978). This paper solves the computation of the reliability of Markov systems. The solution is especially useful for systems where the number of states is large. Not only is the mean time-to-failure computed but an analytic expression is given for the probability of the system's being in a certain state as a function of time. The solution is derived by solving the eigenvalues and eigenvectors of the transition-rate matrix of the Markov system. In the solution, numerical algorithms have been chosen to keep the computer cost down. The method uses just a fraction of the time that, for example, the Runge-Kutta method uses. Empirical Bayes estimators of reliability for Iognormal failure model. W. J. PADGETT and J. A. ROBINSON. IEEE 7tans. Reliab. R-27, (5) 332 (December 1978). Empirical Bayes (EB) procedures are considered for estimating the reliability R(t ; y, tr) = gaufc[(ln t - y)/~] for the lognormal failure model. EB estimators are obtained for the 2 cases: (i) y is unknown and a is known, and both 7 and a are unknown. The empirical Cdf of the maximum likelihood estimators of the parameters is used to obtain the EB estimators. (ii) A smooth EB estimator of R(t; 7, o-) is developed when 7 is unknown and cr is known. A modification of this estimator is proposed for both y and 6 unknown. In both cases, EB estimators are obtained for complete samples at each testing stage. Monte Carlo simulations are presented to compare the EB estimators and the maximum likelihood (ML) estimators of R(t ; 7, 6). The simulations indicate that the smooth EB estimators have smaller mean squared errors than the other EB estimators or the ML estimators. Computation-based reliability analysis in real time applications. A. PEDAR and V. V. S. SARMA. IEEE Trans. Reliab. R-27, (5) 345 (December 1978). Reliability analysis for computing systems in aerospace applications must account for actual computations the system performs in the use environment. This paper introduces a theoretical nonhomogeneous Markov model for such applications. Determinations of MTTF from the transition probability matrix and diagram. J. M. KONTOLEON. IEEE Trans. Reliab, R-27, (5) 347 (December 1978). In designing complex repairable systems it is often important to assess system MTTF for a variety of component reliability characteristics and repair capabilities. Two procedures are given for systems for which a state transition diagram is possible. The first allows such calculation to be performed iteratively without solving each time the system of state equations from the beginning. The approach involves matrix manipulations which split the matrix formulation into two parts; one contains the solution from the previous iteration (at the beginning, the MTTF is calculated in the usual manner) and the other incorporates the changes in the transition probabilities. The second procedure allows the determination of MTTF by decomposing the transition probability