PLANNING AND OPTIMIZING ENVIRONMENTAL STRESS SCREENING

Abstract

Environmental Stress Screening (ESS) is a process where environmental stresses (eg. temperature cycling, random vibration) are applied to an electronic product to remove latent defects (eg. cold solder joint and intermittent pin contact). The common practice in the industries is to use recommended profiles in military standards without a means to assess their effectiveness. This project aims to define an approach to optimize ESS. Literature Search A literature search revealed the following methods of stress screen planning: a. Weibull graph plot of percentage failures against number of temperature cycles. The optimum number of cycles is that which removes 95% of latent defects; b. Modelling equations to relate the quantity of defects after screening to the equipment characteristics and screen strength. The constant terms are derived from historical data of similar equipment; and c. Use of the Mil-Hdbk-344, ESS of Electronic Equipment. This handbook estimates the quantities of latent defects based on the quality of parts and usage environment. It also provides screen strength modelling equations. The low-volume production and lack of historical data do not allow us to use methods (a) and (b). Method (c) provides a means to quantify defects and screen strength, but the profiles obtained are not necessarily optimized with respect to cost. It appears to be the most suitable for further refinement. Relevance of Mil-Hdbk-344 Models To determine if the modelling equations in Mil-Hdbk-344 can be used in our local context, a statistical test on available data was recommended. Screening data for four defence electronic systems was obtained from Chartered Industries of Singapore for this purpose. A Chi-Square goodness of fit test showed that the observed fallout data did not deviate significantly from that predicted by screen strength models in Mil-Hdbk-344. The models given in Mil-Hdbk-344 may be used to plan an ESS program. Insufficient Parts List Information The parts list for ESS planning may be incomplete if a system contains some commercial boards. In addition, Mil-Hdbk-344 does not provide fraction defectives (defects/part) for parts such as batteries, oscillators and sensors. Any resultant inadequacies in the screening profiles should be corrected through feedback of the fallout data. ESS Optimization The method of stress screen planning based on Mil-Hdbk-344 was refined to include the following considerations: a. The sensitivity of the incoming latent defects to different screen types. Temperature cycling removes 80% and random vibration 20% of latent defects; b. More screening should be performed at lower assembly levels (parts or PCB) due to the lower facility cost and higher stress levels allowed. But the fault coverage is lower than the unit or system level; and c. The optimum point to perform ESS is where the expected life-cycle cost (sum of the screening cost and field repair cost) is at a minimum. A Mathematical Programming software tool, GAMS (General Algebraic Modelling System) was used to code the ESS optimization problem. A database program was written to calculate incoming defect densities. The GAMS program then uses the incoming defect density values to work out the optimal ESS profiles.