From Percent Rejects to Parts Per Billion Moving Towards Zero Defects
As a strategy, product or process reuse can reduce launch costs
When Philip Crosby announced zero defects as a philosophy during the 1970s, it was met with incredulity. There were already many articles written on the fallacy of such a strategy and the enormous costs of moving toward zero defects. Fast forward 40+ years, and zero defects has become a reality.
The automotive and semiconductor/high-tech industries that we have worked with have achieved this strategy over a 30-year time period. The strategies to shift organizations that were at 20,000 to 50,000 defective parts per million (PPM) during the late 1980s to 1,000 PPM to 100 PPM and even parts per billion are now quite straightforward. Let us understand how this is done.
New-product development, defect prevention, and FMEA
Quality improvement and defect prevention start with a disciplined, new-product development process that includes phase gates. Most important, three tools are key to overall success:
1. Failure mode and effects analysis (FMEA) that is used with design and processes
2. Design for manufacturing (DFM) and design for assembly (DFA)
3. Upfront involvement of manufacturing in design
These are not the only tools, but they are the most important that contribute to overall success. This article will focus primarily on FMEA.
Our understanding of design failure mode and effect analysis (DFMEA) has grown over time. DFMEA looks at the functions of a product and its requirements and then works on the failure modes of that product. During the design phase, the causes of potential failure are focused on just design failures and not manufacturing failures. The fear is that instead of improving the design, the design team will assign all issues to manufacturing. DFMEA identifies both preventive and detective controls. Preventive controls can go into a preventive checklist, but the detective controls are linked directly to test plans.
DFMEA and test plans
At Omnex, DFMEA at the system level is called system FMEA. There is a link between system FMEA, subsystem FMEA, and component DFMEA. This link is defined by the functions and requirements of each of the design elements.
The link between DFMEA and test plans, as well as the links between DFMEAs are high-level topics only understood by best-in-class organizations. Furthermore, these links between DFMEAs and test plans are really only possible using software. See figure 1:
Figure 1: Requirements manager/flow down and risk analysis
Critical and significant characteristics
The definition of critical and significant characteristics is an important part of the journey to zero defects. Both design and manufacturing need to define safety and key functional characteristics. These characteristics get defined during design, and this definition needs to populate manufacturing tools such as process flow, process failure modes and effects analysis (PFMEA), and control plans.
Process flow, PFMEA, and control plans
Figure 2 shows the links between process flow, PFMEA, control plans, and work instructions. Similar to the links between DFMEA and test plans is the link between PFMEA and control plans. An important element when doing the DFMEA and PFMEA is to look at the scrap and rework data, and ensure that the FMEAs are populated by the “things actually going wrong.”
Omnex has made dramatic improvements in company quality simply by using a methodology that we call Process Review. The Process Review is conducted by cross-functional shop-floor teams created around families of products or processes. Process flow, PFMEA, and control plans are conducted from the viewpoint of overall improvement using historical data. Most important, the updated FMEAs need to become “lessons learned” for future new-product launches.
Figure 2: Links between process flow, PFMEA, control plans, and work instructions
Disciplined problem solving—corrective action
When a product is launched, no matter how good the team is, problems or issues are going to occur both internally and at the customer. These problems are solved using a disciplined problem solving process, typically the 8D process. While the 8D methodology deserves great credit, this article focuses on the preventive nature of taking the problem, its root causes, and the corrective action, and populating both design and process FMEAs. In this way, all known problems are taken into the FMEA and used in the next product design.
Design and process reuse
This then brings us to the topics of design and process reuse. At a high level, this topic is about reusing design and processes that are known to work. What if an organization has data that shows the PPM levels of each process or product—not just the product itself, but to its functions? When product or process reuse is taken as a strategy, organizations can launch products with little to no risk. Very easily, the areas of risk can be pinpointed. In many ways, this is the next generation of the zero defects philosophy, using design and process reuse to reduce new-product launch costs.