Design for Reliability Techniques - Identification and Management of Reliability Critical Components

Transcription

1 Design for Reliability Techniques - Identification and Management of Reliability Critical Components Created by Michael Shover, Ph.D., Advanced Energy Industries, Inc. Abstract Markets that require rapid design cycles and high field reliability must invest robust Design for Reliability techniques during the earliest development stages. In line with industry best practices, Advanced Energy (AE) undertakes a Reliability Critical Component Control (RCCC) program to minimize the risk of incorporating certain components into power and control products. The process progresses from requirements gathering to hypothesis development, followed by hypothesis evaluation and control plans. An example is used for demonstration. Introduction Table of Contents Elements of Reliability Critical Component Control 2 Summary 5 References 5 At the outset of a technology or product development program, AE technical teams distill customer and market needs into hardware, software, and firmware requirements. Iterations to the design can be made with relative ease at this stage, allowing a high return on investment for improvement in key reliability parameters. Cross functional groups collaborate expediently during early development to maximize product reliability from the outset Advanced Energy Industries, Inc.

2 A crucial element of early development is an RCCC program, which aims to identify, analyze, and mitigate the negative impact of components which present an elevated risk to product reliability [1-4]. An RCCC is comprised of multiple techniques, both qualitative and quantitative, and considers similar existing products as well as new information. Elements of Reliability Critical Component Control The implementation of an RCCC process is illustrated in Figure 1. At the inception, requirements are collected, from which working hypotheses regarding component criticality are developed. These hypotheses are formed using baseline information and handbook assumptions, providing initial theories on component failure modes as an exploratory tool. Specifications, schematics, design analyses Theoretical analysis and laboratory testing System and performance requirements Develop criticality hypotheses Evaluate criticality hypotheses Corrective action and/or control plans Reliability improvements to design Supply chain and manufacturing controls Process and application data Identification techniques Figure 1 General process of Reliability Critical Component Control. System and performance requirements Ensuring an accurate and useful RCCC requires an understanding of both the product or technology under consideration and how it interacts with the environment in which it operates. Inputs at this initial step typically include product and module specifications, circuit schematics, and analyses from various technical disciplines, such as electrical, mechanical, software, and thermal engineering. As AE power and control products are integrated into highly complex systems, process and application data are equally important and are incorporated as necessary. Beginning the example, the system under analysis could be a processing and communications module, a mixed-technology PCBA with a microprocessor as the core component from which functionality is derived. The design engineering team proposes to employ the newest variant of a known platform of microprocessors. The system requirements and process data show that the new microprocessor is theoretically well-suited to the task. 2 advanced-energy.com

3 Development of reliability risk hypotheses The conception of hypotheses regarding component reliability risk can begin once the supporting information is collected [5, 6]. In alignment with cross-industry best practices, certain techniques have been adopted at AE as primary methods for developing reliability risk hypotheses: Field data from similar assemblies; Schematic analysis; Worst Case Circuit Stress Analysis (WCCSA); Failure Mode Effect and Criticality Analysis (FMECA); Model-based failure rate estimation, such as RIAC 217Plus and FIDES; and Physics-of-failure. Much of the useful information is contained in the details, such as the design margin noted in a WCCSA or the modal frequencies calculated in a physics-of-failure study. These represent known or estimated quantities that engineers can improve upon. To augment the quantitative details, it is also beneficial to categorize the specific hypotheses that have been developed for both immediate and future RCCC programs. Table 1 shows the categories used at AE to describe the hypotheses. Table 1 Categorized hypotheses for reliability risks. Customer Usage Failure Rates and Modes Design and Architecture Supply Chain Physical environment Part failure causes catastrophic unity failure New technology Unsatisfactory part history Application and usage Part failure would prevent data gathering Obsolete technology Insufficient part history Preventive maintenance schedule Stringent performance requirements Custom component Unsatisfactory supplier history Special handling required Stringent tolerance requirements High quantity Insufficient supplier history Single point system failure Component contains intellectual property Total traceability required Single/sole source Manufacturing process/materials Continuing our example, field data shows that the previous variant of microprocessor has a favorable failure rate and is sourced from a known supplier. It is neither unproven nor obsolete technology and is readily available as an off-theshelf component. A WCCSA indicates that the microprocessor is expected to operate well within the manufacturer specifications and AE derating guidelines under all use cases. The cross-functional DFMECA highlights an elevated criticality failure mode in a switching regulator supplying the microprocessor, which is easily rectified. A physics-offailure analysis indicates that the solder joints on this component, in a ball grid array (BGA) package with 100s of pins, are a potential reliability risk due to PCBA flexure, which could occur with thermal excursions during normal operation. advanced-energy.com 3

4 Evaluation of hypotheses Following the identification of potential reliability risks, it is necessary to determine if those hypotheses are supported by evaluations. Typically, further analysis is considered first, such as more detailed circuit modelling and thermal or structural FEA techniques. If theoretical studies are impractical, laboratory testing is devised and conducted. AE has invested in world-class engineering and reliability infrastructure across multiple sites, enabling internal evaluations including Highly Accelerated Life Test (HALT), Reliability Growth and Accelerated Life Tests, humidity exposure, and high-speed infrared thermography. Leveraging expertise from the supply chain, specifically in the applications engineering groups, is also highly relevant. Suppliers have deep knowledge about the capabilities and limitations of their components and have the benefit of understanding many use cases which may be complementary to those in which AE products operate. In the ongoing example, a pragmatic approach would be designing a HALT. This standardized test sequence, combining reduced and elevated temperature environments, random vibration, and rapid thermal changes, would activate the failure mode of solder joint fatigue, if it is indeed present, to support the hypothesis. Corrective action and/or control plans Depending on the outcome of the evaluations, the technical team will consider the options available [7, 8]. Possibilities for further action include: Collaboration with supplier applications engineering; Further derating of stress parameters; Reduction in quantity for subsystems with high component counts; Design changes to increase reliability; Re-evaluation of requirements or specifications; Parametric part screening. These actions are loosely ordered in most- to least-desirable. A closer alliance with a supplier can clarify component capabilities and perhaps a design can be adjusted to accommodate. On the other end of the spectrum, non-standard part screening at a supplier or in the AE factory would be considered only after other options had been exhausted. 4 advanced-energy.com

5 If it is found that the microprocessor is susceptible to premature solder joint fatigue, potential mitigations could be: Altering the PCBA layout to minimize strain in a certain area; Adding an underfill material to constrain the microprocessor to the PCBA and therefore minimize solder joint strain; or Minimize the thermal excursions with optimized cooling. Summary In conjunction with other Design for Reliability methods, an RCCC program which is properly integrated into the early phases of design and development is key to minimizing component level reliability risks. AE follows best known methods for requirements compiling, risk hypothesis development and evaluation, and corrective action and control plans. The example of a BGA-packaged microprocessor provides a relevant illustration of the process in a typical application at AE. References [1] Reliability Assessment of Critical Electronic Components. Rome Laboratory, United States Air Force, Jul [2] Identification, Control, and Management of Critical Items Lists. National Aeronautics and Space Administration, Oct [3] Failure Modes and Effects Analysis (FMEA), Critical Items List (CIL), and Fault Tree Analysis (FTA). National Aeronautics and Space Administration, Dec [4] Blueprints for Product Reliability Part 3 - Designing for Reliability. Reliability Information Analysis Center, 01-Oct [5] J. Setreus, P. Hilber, S. Arnborg, and N. Taylor, Identifying Critical Components for Transmission System Reliability, IEEE Transactions on Power Systems, vol. 27, no. 4, pp , Nov [6] R. Ghorani, M. Fotuhi-Firuzabad, P. Dehghanian, and W. Li, Identifying Critical Components for Reliability Centred Maintenance Management of Deregulated Power Systems, Transmission Distribution IET Generation, vol. 9, no. 9, pp , [7] H. Wang, K. Ma, and F. Blaabjerg, Design for Reliability of Power Electronic Systems, in IECON th Annual Conference on IEEE Industrial Electronics Society, 2012, pp [8] N. Nakka, A. Choudhary, G. Grider, J. Bent, J. Nunez, and S. Khalsa, Achieving Target MTTF by Duplicating Reliability- Critical Components in High Performance Computing Systems, in 2011 IEEE International Symposium on Parallel and Distributed Processing Workshops and Phd Forum, 2011, pp advanced-energy.com 5

6 ABOUT ADVANCED ENERGY Advanced Energy (AE) has devoted more than three decades to perfecting power for its global customers. AE designs and manufactures highly engineered, precision power conversion, measurement and control solutions for mission-critical applications and processes. AE s power solutions enable customer innovation in complex semiconductor and industrial thin film plasma manufacturing processes, demanding high and low voltage applications, and temperature-critical thermal processes. With deep applications know-how and responsive service and support across the globe, AE builds collaborative partnerships to meet rapid technological developments, propel growth for its customers and power the future of technology. For international contact information, visit advanced-energy.com. sales.support@aei.com Specifications are subject to change without notice Advanced Energy Industries, Inc. All rights reserved. Advanced Energy and AE are U.S. trademarks of Advanced Energy Industries, Inc. ENG-RCC