Laser diode burn-in and reliability testing

More than 99 percent of all lasers manufactured in the world today are semiconductor laser diodes. Reliability is a concern in every laser diode application, whether it is a simple $10 laser pointer or a space-qualified optical transmitter link. The commercial success of a laser supplier rests largely on his ability to develop a robust manufacturing process that consistently produces reliable devices combined with the quantitative assurances he can provide to his customers proving the reliability of his devices. Over the past two decades, laser diode reliability testing techniques and equipment have evolved to support the diverse development of laser diodes.     

A nonparametric approach to estimate system burn-in time

System burn-in can get rid of more residual defects than component and subsystem burn-ins because incompatibility exists not only among components, but also among different subsystems and at the system level. There are two major disadvantages for performing the system burn-in: the high burn-in cost and the complicated failure rate function. This paper proposes a nonparametric approach to estimate the optimal system burn-in time. The Anderson-Darling statistic is used to check the constant failure rate (CFR), and the pool-adjacent-violator (PAV) algorithm is applied to “unimodalize” the failure rate curve. Given experimental data, the system burn-in time can be determined easily without going through complex parameter estimation and curve fittings     

通用微处理器等效老化试验方法

针对不同工艺、不同设计的功能全兼容集成电路等效老化的需要，提取出了集成电路等效老化的特征参数“归-化老化电流”指标a，并讨论了等效老化信号的确定方法。结合集成电路等效老化信号确定方法，以CPU486为研究对象，给出通用CPU等效老化试验方案，为评估和比较不同CPU的质量和可靠性提供了统一的试验平台。     

Mixed-Weibull parameter estimation for burn-in data using the Bayesian approach

A Bayesian approach is presented for both grouped and ungrouped burn-in test data, to come up with the posterior parameters of the bimodal mixed-Weibull distribution. The concepts of belonging probabilities and fractional ranks are introduced for this approach. A numerical comparison is made by conducting the Kolmogorov-Smirnov goodness-of-fit test on the parameter estimates obtained by Jensen's graphical method, by the Bayesian approach using the conventional separation plotting method and by the Bayesian approach using the fractional rank plotting method. It turns out that the Bayesian approach with the fractional rank plotting method yields the best results.     

Real impact of dynamic operation stress during burn-in on DRAM retention time

The burn-in (BI) mechanism in connection with the dynamic operation stress (DOS) has been investigated to examine the real impact on dynamic random access memory (DRAM) reliablity. In this paper, the wafer burn-in (WBI) method with equivalent screening efficiency as the package burn-in (PBI) is implemented by employing DOS. It is found that retention time degradation by BI stress in DRAM with potentially lethal defects is mainly attributed to DOS-induced hot carrier (HC) degradation of DRAM cell. Hot electrons injection in Si-SiO/sub 2/ interface brings about lots of interfacial states as well as the electrical field modification at the gate-overlapped region, causing the degradation of retention time. This is clarified by an anomalous threshold voltage (V/sub T/) shift, and an increase of gate-induced drain leakage (GIDL) after dc HC stress having the identical stress voltage as DOS. Moreover, it is proved that a WBI procedure with the relevant DOS can screen out weak bits effectively, compared to that with only static stress.     

Complex system reliability with general repair time distributions under preemptive repeat repair discipline

The behaviour of a complex system having “N” components in series in class L₁ and “m” identical components in parallel redundancy in class L₂, has been investigated. It has been assumed that a failure in L₁ brings about the complete break-down of the system whereas the failure of any two components of class L₂ causes the system to work in a state of reduced efficiency. The repair of the failed components in the class L₁ and L₂ is carried out under Preemptive Repeat Repair Discipline⁽⁵⁾. Laplace Transforms of various state probabilities, viz. the system is in up state, reduced efficiency state and down state, have been obtained. In the end, asymptotic behaviour of the complex system, has also been discussed.     

The reliability functions of the general source with fixed-lengthcoding

The reliability function problems with fixed-length source coding for the general source are studied for all rates R. Our fundamental philosophy in doing so is to convert all of the reliability function problems to the pertinent computation problems in the large derivation-probability theory. It turns out that this kind of new methodology, which was previously developed by Han (see ibid., vol.43, p.1145-64, 1997), enables us to establish quite compact general formulas of the reliability function for general sources including all nonstationary and/or nonergodic sources with countably infinite alphabet. Such general formulas are presented from the information-spectrum point of view     

LSI reliability prediction based on time

The proposed reliability prediction model takes into account the development of integrated circuit technology as well as the general improvement of the reliability with time.The basic factor of the model is the circuit complexity. Besides the complexity the chip size and temperature factors are also included in the model. The model is applicable to both SSI-, MSI- and LSI-devices.     

A dual burn-in policy for defect-tolerant memory products using the number of repairs as a quality indicator

In most existing studies on the optimization of burn-in for semiconductor products, all chips are treated equally and subjected to burn-in of the same duration (i.e. a single burn-in (SBI) policy is employed). However, the quality levels of chips before burn-in are not the same in general, and therefore, it may be more advantageous to treat chips differently at the burn-in process based on appropriate quality indicators. This paper considers defect-tolerant memory products and develops a dual burn-in (DBI) policy in which the chips submitted to burn-in are classified into two groups according to the number of repairs, a quality indicator that can be obtained from the wafer probe test results, and different burn-in durations are applied to different groups of chips. Then, cost models are developed for the SBI and DBI policies, and their relative performances are compared in terms of the expected total cost per chip. The effectiveness of the proposed DBI policy is demonstrated using the actual data for a certain type of 256M DRAM products.     

A general formulation of Faraday's law of induction

An alternative method for deriving the electromotive force generated by the time-varying magnetic flux in a deformable and moving closed circuit is shown by introducing the modern vector analysis to the original Maxwell's treatment.     

VLSI老化筛选试验技术的挑战

通过对集成电路老化试验技术的研究,指出了VLSI老化筛选试验技术仍然是集成电路产品质量和可靠性保障的重要手段。但是随着集成电路技术的飞速发展,老化技术面临许多有待解决的技术问题,从VLSI产品质量和可靠性保障的角度出发,急需制定可操作的VLSI老化试验技术规范。     

A general software reliability model for performance prediction

In this paper we give a general Markov process formulation for a software reliability model and present expressions for software performance measures. We discuss a general model and derive the maximum likelihood estimates for the required parameters of this model. The generality of this model is demonstrated by showing that the Jelinski-Moranda model and the Non-Homogeneous Poisson Process (NHPP) model are both very special cases of our model. In this process we also correct some errors in a previous paper of the NHPP model.     

Estimation of the reliability function using the delay-time models

The reliability model developed by Christer is used in which the notion of delay-time is involved, which is the span of time from when a defect is first detected upon inspection to when it is considered to have failed. In this paper, a new simplified formula for the reliability function is obtained, in which the delay time and the time to failure densities are exponential distributions. The MLE of the reliability function are also obtained. A numerical procedure is used to solve the results for the new model.     

Distribution of time to non-availability of a reliability system

The author derives in this paper the distribution of the time to non-availability of a system which under-goes the following states, viz. (i) operative (ii) failed, and (iii) repair. Special cases are discussed and tables are provided at the end.     

Microcomputer real time software reliability and fault recovery

An analysis of microcomputer software reliability is formulated; it is particularly applicable to real time control systems. Program structure and error recovery techniques are presented. An example of an oven controlled by a microcomputer is given and the practical implications of the theory are outlined and discussed.     

Analysis of trap effect on reliability using the charge pumping technology in La-incorporated high-k dielectrics

In this paper, reliability characteristics of nMOSFETs with La-incorporated HfSiON and HfON and metal gate have been studied. HfLaSiON shows greater device degradation by hot carrier (HC) stress than by positive bias temperature (PBT) stress, while HfLaON exhibits similar degradation during HC stress and PBT stress. To evaluate the contribution of bulk trap during PBT stress, a novel charge pumping (CP) technique is applied to extract the distribution of bulk trap (N_bt) before and after PBT stress. To evaluate permanent damage during HC stress, an appropriate selection of frequency range in CP method is considered. The initial interface trap density of HfLaSiON and HfLaON is similar, while the near-interface trap (N_IT) density of HfLaSiON after HC stress is equal or greater than that of HfLaON.     

Nomination time: CPR's 2008 law firm award

The competition opens for CPR's second Law Firm Award for Excellence in ADR, with entries due this month. Also, details on next month's meeting in Houston; the CPR National Task Force on Diversity in ADR; the keynoter for January's 30th Anniversary CPR Annual Meeting, and much more.     

Using repair-depot system reliability to determine the distributionof supportability turn-around time

Assuming constant repair times, Linton, et al. (1995) used an `expression for the reliability of the system for repairing failed units (FU) at a repair-depot' to compute the longest repair time for a newly failed unit (NFU) which assures a given reliability level (also termed the NFU supportability turn-around time, STAT) in terms of the: (1) constant failure rate for all components, number of spares (s) on-hand; (2) number (n) of FU either `under repair' or `scheduled to begin repair in the future'; and (3) downstream repair completion times (DRCT) for FU. Since subtraction of the repair time for a NFU from its STAT-value yields the NFU's latest repair start-time (LRST) which assures a given repair-depot system reliability (RDSR), STAT-values are important for scheduling RST. This paper assumes that repair time is a random variable and, consequently, DRCT is a random variable. As shown in Linton, et al, (1995), STAT is the zero of a nonlinear, nonpolynomial function of DRCT; thus, STAT is also a random variable, and determining the distribution of STAT is a stochastic root-finding problem. For n=1 and s⩾0, numerical analysis and probability theory are used to find the Cdf and pdf of STAT in terms of any repair time pdf. Using the pdfs for STAT and repair time, expressions are derived for: (1) E{LRST} for a NFU; and (2) q=Pr{(repair time+c)相似文献   

Improvement of reliability of radiation parameter estimates in real time

The paper analyzes the operation conditions of radio-engineering systems from the viewpoint of their reliability of measurement of radiation parameters in the event of a priori unknown carrier frequency and signal power level in the reception point. A possible approach to treatment of this problem consists in adaptive selection of the tunned frequency channel. A block diagram of the relevant device, together with estimation of its noise stability, is included.     

Analysis of microstrip circuits using three-dimensional full-waveelectromagnetic field analysis in the time domain

In this method, the circuit is excited by a raised cosine pulse source that includes broadened frequency components. The frequency characteristics are then computed from the Fourier transform of the output transient responses. To evaluate the validity and capability of the method, a side-coupled microstrip filter is analyzed and the frequency characteristics are calculated. A quasistatic analysis of this filter is also presented and the results compared with measurements. The frequency characteristics calculated with the full-wave analysis in the time domain show excellent agreement with the measured values, thus demonstrating the validity and the power of the analytical method