最坏分析

yulzhu

最坏分析 [复制链接]

 

 

 

最坏分析其实就是极端值分析：让你知道怎么样才会出问题。

Assure acceptable operation throughout the entire product life cycleunder the most unfavorable combination of anticipated conditions
        Define Critical Components and Spec. Control Drawing (SCD) Limits
        Provide Acceptance Test Procedure (ATP) Limits
        Define Need for and Range of Select-At-Test (SAT) components
        Improve Reliability through Parts Stress and Derating analysis
        Identify design concerns which during test, alignment, and use could result in circuit damage or premature degradation.       

        WCCA helps ensure increased product reliability. We accomplish this through         rigorous mathematical and simulation-based models along with hardware correlation.         Correlated models are then used to determine part stress margins, and EOL/BOL product         operating specifications. A single over-stressed component can cost your company millions        of dollars. A thorough Worst Case Analysis can eliminate this from happening.

Background & Definition

    A Worst Case Circuit Analysis is a quantitative assessment of the equipment         performance, accounting for manufacturing, environmental and aging effects.          In addition to a circuit analysis, a WCCA often includes stress and derating         analysis, Failure Modes and Effects         Criticality (FMECA) and Reliability         Prediction (MTBF).

                The specific objective is to verify         that the design is robust enough to provide operation which meets the         system performance specification over design life under worst case conditions        and tolerances (initial, aging, radiation, temperature, etc.).

一般我们做的比较多的就是stress和derating，这个最直观也最明显了。
        The Stress and Derating Analysis is intended to increase reliability by providing         sufficient margin compared to the allowable stress limits. This reduces overstress         conditions that may induce failure, and reduces the rate of stress-induced parameter         change over life. It determines the maximum applied stress to each component in the         system.

Methodology

 

 

 

•     A WCCA follows this general form:
• Generate/Obtain circuit model
• Obtain Correlation to validate model
• Determine sensitivity to each component parameter
• Determine component tolerances
•         Calculate the variance of each component parameter as sensitivity times absolute tolerance
• Use at least two methods of analysis (eg. hand analysis and SPICE or Saber, SPICE and measured data) to assure the result
• Generate a formal report to convey the information produced       

自己列公式+Pspice+测试数据，选两个对照一下罗
            The design is broken down into the appropriate functional sections.     A mathematical model of the circuit is developed and the effects of various     part/system tolerances (see below) are applied. The circuit's EVA and RSS     results are determined for Beginning-of-Life and End-of-Life states.

        Two methods of analysis and/or hardware correlation are always used to    confirm results.

        These results are used to calculate part stresses and are applied to other     analyses. In order for the WCCA to be useful throughout the product’s life cycle,     it is extremely important that the analysis be documented in a clear and concise     format. This will allow for future updates and review by other than     the original designer. A compliance matrix is generated that clearly     identifies the results and all issues.


Factors Addressed During WCCA

Analog Circuit Analysis
Some of the factors to be considered during analysis include:
Maximum line voltage variations and line transients（电源电压和暂态脉冲）
Maximum input and output variation（输入输出变化）
Maximum part parameter variation元件参数变化
Maximum performance demands and variations性能需求
Maximum and minimum environmental conditions（环境条件）
Fail-safe provisions（失效还是安全的，高可靠性的必备条件）
Redundancy provisions
Radiation effects, as applicable（辐射效应-一般用于太空拉）
Parameter drift due to aging（参数老化）
Transients due to turn-on, turn-off, and state changes（开关和状态变化引起暂态）
Fatigue due to cyclical loading and temperature cycling（周期性负载和温度循环）
Interface conditions between modules and modules to test equipment（模块间接口）

Digital Circuit Analysis
Digital circuit worst case analyses involves one or more of the following:
Timing Margin Analysis
Transmission Line Effects传输线效应
Noise Due to Crosstalk and Grounding 串扰
Meta-Stability Analysis
Decoupling Analysis解耦分析
Fanout Analysis 扇出分析—驱动能力
Logic Compatibility/Interface Analysis逻辑功能
Supply Power Application and Sequencing Analysis.
State Machine Analysis状态机
Unused/Tri-stated Input Analysis
No Connection Analysis
Test Point Current Limiter Analysis
Physical Layout Analysis
One-Shot Margin Analysis

Component Tolerances

Initial 初始误差
Temperature and Environmental Factors 温度和环境效应
Radiation 辐射
Aging or End-of-Life Factors 老化

yulzhu

今天发生了这样一件事情，是有关于芯片的验证的问题。我们知道考虑模块的温度的时候一般根据模块的位置以及极限环境，然后根据模块本身的发热状况再考虑整个附加的模块散热引起的内部温升。如下图所示：



我们知道ISO16750中气候已经定义了不同的安装位置的模块的外部环境温度，一般而言是-40～85度这个范围，而由于模块内部有发热元件，会导致模块内部的温度比外部的温度高的情况，这是因为一般模块我们都采用空气散热的方式。如下图所示：



我们在考虑的时候给这个温度一定的数量，如果模块的发热较严重，这个数值则偏高，一般可选择加10摄氏度。
我们知道我们正常的电压范围一般为 9～16V，而又有18V（60分钟）和24V（1分钟）的要求，这里需要提一下的是，我们选择电压条件和环境组合的时候，判断元件的阈值是不同的
16V，85+10degC的时候，需要在150度（普遍半导体的损坏温度，元件说明书中都有要求），在这个基础上，我们要选择下降一定的幅度，如果汽车在夏日高温下暴晒，其内部温度可能非常高的时候，这种情况就可能发生。
18V和24V都属于汽车发生故障的时候的情况，因此只要在这个时候模块不损坏（150度为阈值），设计就可以接受。注意24V按照实验条件是从55度开始考虑的。
以电阻为例，贴片电阻是随着环境温度的增加，超过70的时候散发功率下降了。


因此需要注意以上几点问题。
贴片电阻一般采用比较其耗散功率与实际功率的大小来判断电阻的状态。
一般的元件都是按照热阻去计算芯片温度的，我们一般假定芯片的热阻不变。
一般的我们认为耗散功率=（150-环境温度）/芯片热阻【一般假设芯片热阻在全温度范围内不变，实际上这也是取最值的方法】。
这是因为电阻的热阻随着环境的温度变化而变化（小于70度的时候，热阻一直在变化）
而电阻的耗散功率:
环境温度<70,耗散功率维持不变，电阻热阻一直在变化。
环境温度>70，耗散功率=P_额定*(150-环境温度）/150-70,此时热阻是不变的为P_额定/80。

decho0821

不错啊

lin_john

lightvsdark

看看，谢谢楼主分享。

greyfeather77

这个 楼主的博客好久没更新了啊
加油啊

70120662

楼主文章不错！学习了！收藏起来，表示感谢！

东方遗

看看，谢谢楼主分享