Establishment of a Quality Assurance System

Figure 1:
　　The architecture of the quality assurance system, as shown in Figure 1, is the framework we aim to establish. Those who have undergone training in the five TS tools might wonder: In APQP, the sequence for quality planning calls for Process FMEA (Failure Mode and Effects Analysis) first, followed by the Control Plan; yet we have reversed this order. Is this a mistake? Let me state clearly: it is not. First of all, the architecture we are developing does not simply replicate the TS systems established by Chrysler, Ford, and General Motors. Instead, we are creating a quality assurance system tailored specifically to the printed circuit board manufacturing industry, with its own distinctive characteristics. This system integrates a variety of quality management tools. Anyone familiar with quality management knows that the Cause-and-Effect Matrix is a key tool within Lean Six Sigma; “Control” and “Improve” are routine components of daily quality‑management activities—what we are now doing is standardizing these processes and formalizing them into a structured working model. OCAP, or “Out-of-Control Action Plan,” refers to procedures for addressing abnormalities when they occur. Meanwhile, improvement initiatives such as IE and TPM have been continuously implemented under the guidance of U‑WIN consultants. TPS, or Toyota Production System, encompasses JIT, Kanban management, corporate culture, and core values; among these, JIT—literally translated as “just-in-time production”—is known in quality management as “zero‑inventory” management.
　　The development of P‑FMEA (Process Failure Mode and Effects Analysis) and the CONTROL PLAN also differs from the TS system, because what we are doing is not quality planning in advance. In April 2006, when TPC Plant No. 2 went into production, if we had prepared the P‑FMEA and CONTROL PLAN at that time, following the APQP principle of “P‑FMEA first, then CONTROL PLAN,” it would have been correct. However, now we are retroactively addressing work that was either not done or inadequately executed in the past. Therefore, we must first create a CONTROL PLAN for a specific product model, from which we can logically derive the P‑FMEA for each process (or operation). Subsequently, based on the product’s manufacturing flow and the P‑FMEA, we will develop CONTROL PLANS for all products. It is important to note that the P‑FMEA is developed on a per‑process basis. A “process” refers to a specific operation that accomplishes a particular machining function; it constitutes a segment within the overall manufacturing sequence. For example, our outer-layer patterning process can be divided into three distinct operations: outer-layer pre‑treatment, outer-layer dry‑film application, and outer-layer etching. We are required to establish a P‑FMEA for each of these processes. Should there be any changes to the process, equipment, materials, product, or personnel, the P‑FMEA must be revised or revalidated. To date, we have completed the P‑FMEAs for the outer-layer pre‑treatment, outer-layer dry‑film application, and outer-layer etching processes (see Figure 2: P‑FMEA for the Outer-Layer Etching Process).

Figure 2:
　　The outer-layer etching engineering P‑FMEA and the CONTROL PLAN are developed based on the product; in principle, a separate CONTROL PLAN is prepared for each process step. Moreover, the CONTROL PLAN must be structured in accordance with the sequence of the product’s manufacturing process. The CONTROL PLAN shown in Figure 1 (Quality Assurance System Architecture) refers to the CONTROL PLAN created specifically for implementing the P‑FMEA, and it covers only one or two process steps. At present, I have completed the CONTROL PLANS for two process steps: the six‑layer immersion nickel–gold surface‑treated product T29983 and the double‑sided hot‑air solder‑leveling surface‑treated product T27082 (see Figure 3: Partial CONTROL PLAN for T29983).

Figure 3:
　　The CONTROL PLAN for T29983 (partial) is now as follows: We will refine the CONTROL PLANS for T29983 and T27082, then develop P‑FMEAs for all engineering projects. Finally, we will establish CONTROL PLANS for all products, with particular emphasis on new part numbers and repeat‑order items that present significant manufacturing challenges. In addition, we will formulate OCAPs for those items in the P‑FMEA that have high RPN values; I will prepare the OCAP sampling plan in the next phase of work.
As the ancients said, “To believe everything written in books is worse than having no books at all”—and there’s certainly wisdom in that. PPAP, SPC, MSA, APQP, and FMEA are known as the “Five Major Tools of TS”; they’re called “tools” precisely because we should draw inspiration from them rather than slavishly copy them. Zhao Xuetao, a consultant at Naiweige, argues that, apart from SPC and MSA, the other three aren’t truly tools—but I disagree with her view. PPAP, APQP, and FMEA are indeed tools, and we must apply them effectively to the PCB manufacturing process. As the saying goes, “A foreign monk may chant the sutras,” but we must never forget that “Chinese culture is profound and extensive.” Every Chinese dictionary includes the phrase “to infer other cases from one example.” Quality management is a task that clarifies the “direction” but not the specific “path,” making it quite challenging to implement. In recent years, Western countries led by the United States, along with Japan, South Korea, Taiwan, and other regions, have each produced successful quality‑management case studies and introduced various quality‑assurance systems, tools, and methodologies. At first glance, this might seem to have provided a clear roadmap for quality work—but that’s far from the truth. If we simply adopt their tools without critical thinking, we will not succeed. Instead, we need to “infer other cases from one example,” leveraging our own unique circumstances to make these quality tools truly effective.
As quality managers, we must never cling rigidly to a single set of tools or attempt to solve every issue with just one approach—no single quality tool is universally effective. That’s why the quality assurance system at TPC Plant No. 2 adopts a diversified framework, seamlessly integrating multiple management methodologies and tools. We leverage Lean Six Sigma principles, the five core tools of ISO/TS 16949, the Toyota Production System, and many others. Moving forward, we may also introduce additional quality tools and approaches, harnessing cutting-edge global expertise to elevate TPC’s quality standards and forge a remarkable legacy in the PCB industry.

Quality Engineer Jia Ru