MTBF-公式介绍.doc

MTBF-公式介绍精品文档MTBF Introduction 利用加速寿命试验的手法及增加取样数量的方式，来缩短测试产品出现失效的时间，以便在短期内达到符合产品的平均失效间隔时间MTBF (Mean Time Between Failure)，进而保证产品在可用期的可靠度。其实施步骤可参考平均失效间隔时间验证作业程序 (TG-SX00-Q066) It is using accelerated life tests and increasing sample sizes to shorten the failure time of products so that we can reach the MTBF(Mean Time Between Failure) for products in a short period. As to its procedures, we can refer to “MTBF prediction/Demonstration Standard”(TG-SX00-Q066). (1). 根据客户要求之测试条件及标准，计算出在正常环境下累积MTBF所需的时间(T)。 Calculating the required time (T) being accumulated as MTBF in normal environment according to the demanded test conditions and criteria of customers.(a). MTBF之累积一般均采用单阶时间检剔数据法(即预先设定试验截止的时间点)，在试验中若发现不良品，就必须参考固定时间试验GEM(Generalized Exponential Model)表，计算出累积MTBF所需的时间。由GEM表之试验比，可以看出不良数越多，则累积MTBF所需的时间(T)越长。 Time screening method is generally used for the accumulation of MTBF. If there is any defected sample found during testing, it is a must to calculate, referring to fixed time tests GEM (Generalized Exponential Model), the time needed toaccumulate MTBF. It can be seen that the more failure quantity is, the longer time (T) is needed to accumulate MTBF. 试验比(M) = T = M MTBF 累积MTBF所需的时间Time needed to accumulate MTBF (T) 平均失效间隔时间Mean time interval for failure (MTBF)(b). 使用GEM表须按照客户指定的信赖水平来查表，若客户无特别要求，则建议选择90%信赖水平为基础，此为目前业界常用的水平。 We must follow the confidence level assigned by customers to use GEM table. Confidence level of 90%, which is the most common used in the field, is suggested to be adopted. (c). 在试验中若发现不良品，则将不良品修复，并登记一次不良。修复之不良品，可继续加入试验累积MTBF，同一件产品可重复修复，直到无法修复为止。 If there is any NG sample found, it should be repaired and this log needs to be recorded. Repaired NG samples can continue being tested to accumulate MTBF, and one product can be repaired repeatedly until it can not be repaired anymore.(d). 假设试验时每天开机的时数为H，待测产品数量为N，则可算出累积MTBF所需的时间(T)需要几天(D)： Assume that the sample quantity under test is N and the daily power on time is H, and then we can calculate how many days (D) the time (T) needed to accumulate MTBF. H N D = T = D = T/( H N ) (2). 利用加速寿命试验的手法，缩短累积MTBF所需的时间。 Shorten the time needed to accumulate MTBF by using accelerated life test.(a). 若加速寿命试验所考虑的环境应力为温度，且失效模式符合指数分布，则可采用Arrhenius模式作为加速寿命试验之模式，其加速因子A可表示为： We can use Arrhenius mode as accelerated life test mode if the considered environment stress is temperature and the failure mode meets exponential distribution. Its accelerated factor can be expressed as:Tn1Ta1A = e(E/K)( ) 式中K为Boltzmann常数(= 8.62310-5 eV/K)，E为活化能，Tn 为正常环境之温度，Ta 为加速寿命试验之温度。上式系假设试件在两种不同温度下之活化能相同所获得的结果，故应用时两种试验温度差距不宜过大，否则所预估之加速因子可信度将降低。一般电子产品在正常有用期失效之活化能趋近于0.6 0.8 eV。 Where K is Boltzmann constant (=8.623x10-5 eV/K), E is activation energy, Tn is the normal temperature, and Ta is the temperature for accelerated life tests. The formula above is derived from assuming that DUTs have the same activation energy are under two different temperature, therefore the test temperature differences had better be limited otherwise the credibility of predicted accelerated factors will lower. The activation energy of common electrical products during normal useful period approaches 0.60.8 eV. (b). 在加速寿命试验中，一般皆采用提高环境温度来达到加速劣化的目的。假设在应力环境中老化的加速因子为A，则可算出在应力环境中累积MTBF所需的天数 (Da)为： Raising temperature is usually adopted to reach the purpose of accelerated aging. Assume that aging factor in stress environment is A, and then we can calculate the days (Da) needed to accumulate MTBF in stress environment:H N A (Da) = T = Da = T/(H N A)(c). 假设一般使用者每天开机时数为U小时，则由Da之天数可预估使用者在正常环境下使用P天之状况： Assume that the power-on-hour per day of common users is U, and then we can predict the usage status under normal environment of users after P days:H A Da = U P = P = (H A Da)/U上式中若P = 30，就是预估使用者开机30天(I+30)的状况；若P = 90，就是预估使用者开机90天(I+90)的状况。In the aboveif P = 30, it is to predict the status of users for 30 days (I+30):if P = 90, it is to predict the status of users for 30 days (I+90)(3). 在试验初期若发现重大之遐疵或大量之不良品出现，则可考虑停止试验，将对策导入所有的待测产品，并重新开始累积MTBF。如此可避免过多的不良品，造成试验时间之延长。 If there is any major defect or a great quantity of NG products turns up, stopping testing can be considered. After phasing in countermeasure to all DUTs, restart to accumulate MTBF. Thus it helps to avoid too many NG samples or the extension of the test time. (4). 执行MTBF之环境条件及累积时数，须符合客户之要求。 The environmental conditions and accumulated time for performing MTBF must meet customers requirements. (5). MTBF累积，可使用不同时间点所产出之成品分批累积，但必须具备相同设计及相同制程之产品才可累积。 It is allowed to use in batches the finished products at different time points to accumulate MTBF; however, the products must have the same design and process. (6). 待测物必须是经过早夭期筛选(例如：Run-in)之良品。 Devices under test must be good samples passing the screen of early life period (Ex: Run-in). (7). 加速寿命试验中所选用的温度条件(环境应力)，必须有适当的考虑，其可施加应力之大小应受所造成失效模式与实际作用时发生者相同的约束。加速因子与应力之大小不一定有成正比的关系，若无限制的提高应力大小，则可能会有异常的失效模式出现。 The temperature conditions (environment stress) selected in accelerated life test must be considered properly. The strength of applied stress must be limited by “the failure mode caused is the same as that happens in practical operation”. It is no necessary that the relationship of direct proportion exists between the accelerated factors and the strength of stresses. Abnormal failure modes may occur if there is no restriction on raising the strength of stresses. (8). 若考虑待测物作完试验后，仍然维持可以出货之状态，则必须设定待测物试验时间的上限值，以避免因过度劣化而影响出货的质量。 Maximum time limit for testing must be set to avoid over aging if we still hope that the DUTs can keep in good condition after testing and be sold. 收集于网络，如有侵权请联系管理员删除