Some simple reminders for maximizing test implementation ROI.

Despite the prevalence of papers, case studies, consultants and so on, a great deal of confusion remains surrounding the selection of an optimum test strategy. Those investigating test strategies – be it solder paste inspection, AOI, automated x-ray inspection, or in-circuit – have much to consider when deciding what test strategies to implement. Inspection occurs for:

• Improvement of process defect coverage.
• Improvement in ability to meet customer specifications.
• Detection of process defects as quickly after they occur as possible.
• Ability to decrease process defect rates through SPC.

However, those four categories alone do not ensure test and inspection system implementation. Often, a manufacturer will be required to show ROI data to justify the purchase of equipment, and this is not always easy to come by. Countless companies have purchased equipment based on an ROI calculation; however, their studies were internal and most manufacturers are gun-shy about sharing their results publicly or even at all, for fear that others will view the data and become concerned that they have quality issues. (Ironically, if the before-and-after data did reside in the public domain, others would see how the equipment implementation has positively contributed to their company’s quality.) Figure 1 shows a general depiction of this concept and indicates that for most processes, inspection will cut defects by about half by identifying systematic errors, while measurement-based SPC will reduce errors in half every six to 12 months. To clarify, systematic defects are typically defects that occur on a larger number of boards and are due to a specific cause, like a bent nozzle on a pick-and-place machine, soldering issues with one type of component, or a DfM issue.


While Figure 1 is powerful evidence, this is a limited view of ROI, as other aspects of inspection cannot be fully represented by just counting defects. For example, consider that most companies view WIP (work-in-progress) as a capital cost. Let’s consider a production line where the labor cost is very low and an argument against implementing automated test and inspection is that the manufacturer can use human resources instead. The risk is, in the end, this will amount to a hidden cost rather than a savings! From a dollar-for-dollar perspective, the labor investment may be far less than the initial capital investment. However, that system will continue to generate a stable return over time, paying back its initial investment cost. By contrast, personnel will likely change frequently – as happens in low-cost regions – thus forcing ongoing training investment. Even if that still makes the labor significantly cheaper, consider that workers (who are prone to random and inconsistent mistakes) then manage the repair and test strategy as defects arise; more things get built up in the WIP, and WIP is a capital expense that keeps accruing!

Another issue that impacts WIP is defect clustering. Defects typically do not come in isolated cases; they come in groups. While a board with a single defect may be easy to troubleshoot and repair, a complex board with multiple defects may be nearly impossible to repair. These boards then become permanent occupants of the WIP pile. As products continue to evolve with technology, the complexity is also increasing and, therefore, the human error effect is greater.

Where does a company go from here? To maximize test and inspection implementation ROI, these tips and guidelines should be considered. For process control and faster process feedback, the inspection system should be positioned as early in the process as possible. The short loop allows the user to control and optimize processes. SPI and pre-reflow AOI would have an advantage here. For defect containment or complete process feedback, the inspection system should be positioned as late as possible in the process, if possible, after wave. Post-reflow techniques such as post-reflow AOI, AXI and ICT would have advantages here.

Further, SPI and AOI are best used in the SMT line with the main focus on process control and secondary focus on defect detection, especially for pre-reflow and paste inspection. AXI, post-reflow AOI, and ICT are best used after all soldering process steps, with the main focus on defect detection with process feedback and the secondary focus on process control. Therefore, in many cases, AOI and SPI could be used for process control in the SMT line, and AXI, post-reflow AOI and ICT can be used for end-of-line defect detection, especially for high-complexity assemblies.

These tips coupled with the knowledge of a product’s complexity, cost to manufacture, defect levels, etc. can be used to select an optimum test strategy.

Stacy Kalisz Johnson is product marketing engineer at Agilent (agilent.com); stacy_johnson@agilent.com.