With today’s rapid product development cycles and time-to-market pressures, there’s not always time to perform reliability testing. This situation leaves many manufacturers with the question of how to ensure their products will be dependable when reliability testing and the possible resulting re-engineering are too time consuming or expensive.
The key to ensuring reliability when testing is not possible is to employ rigorous Design for Reliability (DfR) methodologies. DfR is a process for ensuring a product or system will perform the specified function within the customer’s use environment and over its expected lifetime. DfR occurs at the design stage, beginning with concept feasibility and before physical prototype construction. It is often part of an overall Design for Excellence (DfX) strategy. In addition to the obvious mechanical considerations, DfR needs to include electrical and software reliability, supplier and part selection, design for manufacturing (DfM) and Physics of Failure (PoF) analysis/modeling.
Not every product needs the same level of reliability to meet customer expectations. Four broad product reliability categories can be identified, from low-end consumer products like toys, pedometers, etc., to highly regulated electronics such as avionics and telecommunications. To ensure product reliability will meet customer expectations, the desired product lifetime and performance must be established during the product concept phase and should be used to drive sourcing and design decisions throughout product development. Expectations can be quantified; for example 95% reliability with a 90% confidence level over 15 years.
An important part of setting specific reliability goals is to establish the field environment. This can be done based on actual measurements (data for similar products may be used if time is limited) or use of industry specifications. The latter approach may be best if time is a severe constraint, and it is low cost and comprehensive. Specifications, such as MIL-STD-810, MIL-HDBK-310, IPC-SM-785, Telcordia GR3108, IEC 60721-3, etc., represent agreement throughout the industry for certain classes of electronics, though results are likely to be less accurate than measurement for actual products.
A large number of hardware reliability problems are driven by arbitrary size constraints set too early in the design phase. Problems that may ensue include poor interconnect strategies and poor choices in component selection. Therefore, it is best to keep dimensions loose in the initial portions of the design stage.
Reliability problems can also stem from inadequate analysis of electrical impacts. Power stability concerns exist in many markets worldwide. For example, in China there are issues with grounding, in India brownouts are common, and in Mexico voltage surges often occur. Furthermore, reliable designs will take into account electrical emissions (EMI/EMC), not just from a regulatory standpoint but from a product performance perspective. Finally, design for electrostatic discharge (ESD) is ever more important because components are increasingly sensitive to ESD.
Supplier and Component Selection
Supplier selection plays a critical role in the success or failure of the final product. OEMs must build manufacturing supplier partnerships that can consistently identify and align with design and functionality requirements, delivering quality products and services without interruption. Unwittingly engaging with manufacturers that lack these fundamental capabilities can dramatically increase the potential for defects, resulting in expensive warranty claims, loss of market share and possible irreparable damage to your brand.
Component selection is equally important and has several aspects. First, keep it simple. New technology can be attractive but is not always appropriate for high reliability systems. Be conservative, but not overly so since old technology quickly becomes private labeled—that is, “lawful counterfeiting.”
Next, employ a strategy of component derating. Derating is the practice of limiting stress on electronic parts to levels below the manufacturer’s specified ratings. Derating guidelines vary based upon the environment (severe, protected, normal, space/aircraft/ground). Sources for derating guidelines include governmental organizations, third parties, OEMs and component manufacturers. Finally, derating must have a practical or scientific foundation. Thus, component derating should be assessed through component stress analyses.
Design for Manufacturing
Design for manufacturing is the process of ensuring a design can be consistently manufactured by the designated supply chain with a minimum number of defects. DfM works best when there is an understanding of best practices (what fails during manufacturing and how to avoid it) and an understanding of the limitations of the supply chain.
DfM includes a variety of tools and processes, some or all of which could apply to a given product. A DfM checklist is given in DfR Solutions’ Design for Reliability webinar, which includes items related to the bare board (trace width/spacing, laminate material, etc.), the system (e.g., blind connections), and the assembly process (eliminating hand soldering, proximity of components to points of board flexure, etc.).
Component fracture due to excessive board flex during manufacture or handling is an example of how DfM can help avoid failures in the field. Processes of concern can include depaneling, in-circuit testing (ICT), screw attachments, connector or daughter card insertions and the attachment of structures such as heat sinks, covers or plates. Each design should be evaluated through every post-assembly process for excessive flexure. Until recently, such analyses were problematic due to cost and time. Now, however, fast and cost-effective tools are available such as DfR Solutions’ Sherlock Automated Design Analysis™ software ICT module which can perform analyses in a few minutes. Solutions to problems identified by analysis may be to avoid case sizes greater than 1206, maintain 30-60 mil spacing from flex points or to measure board-level strain during manufacture and maintain it below 750 microstrain.
Physics of Failure
Increasingly, companies need powerful algorithms and tools to accurately predict the probability of failure over the lifetime of the product. Even cell phones, having power amplifiers with high power cycles and dissipation, can experience wear out in three years. Over the past two decades, sophisticated Physics of Failure models have been developed and have proven to be useful tools for design. Such models exist for solder fatigue (including Pb-free), thermo-mechanical models for power cycling, board-level strain calculations for mechanical shock using finite element analysis (FEA) and vibration analysis using analytical and FEA modeling. If such capabilities are not available in house, they can be outsourced to experts in modeling for reliability. See, for example, the modeling resources provided by DfR Solutions, as well as the Sherlock Automated Design Analysis™ software.
Best in Class DfR
The companies that exhibit best in class DfR to ensure product reliability that meets customer expectations employ the following steps.
Step 1: Don’t paint yourself into a corner too early in the design process
Step 2: Be aware of ALL requirements
Step 3: Perform concurrent engineering
Step 4: Use a design checklist (don’t rely on tests to develop a robust design)
- Part selection
- Power Stability
- EMI / EMC
- Design for Manufacturability / Testability / Environment
- Components that wear out
Including a thorough test plan as part of the design phase will help ensure expected product reliability and success. Learn more about failure analysis and how and why test plans work by accessing our free guide, “Test Plan Development: How To Do It.”