Estimated reading time: 4 minutes

Problem Solving

Related to TQC and a very important role of an engineer is solving problems. Using a problem solving methodology is a job that all engineers will use sooner or later, but if you are in product or process engineering in manufacturing, it will be sooner!  This was the situation that introduced me to printed circuit manufacturing.

I started at Hewlett-Packard in the very new department of integrated circuit manufacturing. What a wild environment that was in 1970!  After six months of orientation and familiarization, the bosses came to me one day with an urgent offer: “You are our only chemical engineer and we are in a crisis in our new printed circuit manufacturing plant. Will you go down the hill and help solve the process problems that are plaguing them?”  

“Sure,” I said. “What are printed circuits?” 

He answered, “They’re just large integrated circuits!”  

I went down the hill (from Page Mill Road to Porter Drive in Palo Alto), solved their process problems in a few weeks…and never left! I had two tools the electrical and mechanical engineers in the PCB plant didn’t have: engineering statistics, including design of experiments (DOE—more in the next column) and experience with problem solving—TQC.

In electronics manufacturing today, problems in production will involve numerous customers. Customer communication is essential. The timing of this communication depends on how quickly the supplier is expected to correct the issue. Feedback should be specific and detailed, including part number, lot number, invoice, date received, and supporting evidence such as photographs and test results if available. The supplier will need this information to conduct an investigation of root cause and develop a corrective action plan.

For serious quality problems that generate scrap or rework, customers will insist that the supplier submit a written document that describes the investigation and corrective actions, or a corrective action report (CAR). The purpose of the document is to provide a record of the problem solving, and establish confidence that the supplier has successfully addressed the issue and that the issue will not recur.

When selecting a problem solving process, it is important to understand when you should—and should not—use structured problem solving. Therefore, an understanding of problem solving methodology is crucial. Once you have selected a process to use, be sure to document and communicate progress throughout the project. 

Already written about in this column is the TQC PDCA Process:  Plan—Do—Check—Act. Also, the six-sigma DMAIC process: Define—Measure—Analyze—Improve—Control. Familiar to many of you would be the general scientific method:

• Define the question/make observations
• Gather information and facts
• Form hypothesis
• Perform experiments and collect data
• Analyze data
• Interpret data and draw conclusions
• Summarize results

A common methodology used by many suppliers is ‘Eight Disciplines Problem Solving’ (8D), created by the U.S. Department of Defense and popularized by Ford (Table 1), and the CAR based on this process is sometimes referred to as an 8D report. Another popular methodology is the ‘7-Step Problem Resolution Process’, attributed to Toyota and described in Table 2.

My favorite is the problem solving methodology that is taught by Kepner-Tregoe[1]. This is a rigorous three-day course, usually referred to as KT, which has expanded the problem solving processes into four areas:  Situation Appraisal—Problem Analysis—Decision Analysis­—Potential Problem Analysis (Figure 1). Its three action sequences (Figure 2) summarize the important steps in the KT process.

Figure 1: Kepner-Tregoe Problem Solving Process©® (courtesy Kepner-Tragoe).

Figure 2: Kepner-Tregoe Problem Action Sequence©® (courtesy Kepner-Tragoe).

Table 1. 8-Step Problem Resolution Process

Table 2. 7-Step Problem Resolution Process

All of these processes emphasize the need to understand why the issue is occurring under one set of circumstances but not another, and to consider more than one possible root cause. Regardless of which process is used, the customer should require a systematic approach to problem solving from the supplier to avoid these pitfalls: 

• Poorly defined and characterized problem
• Rapid convergence on a single root cause without considering others
• Confusion about who is working on the resolution
• Failure to segregate suspect material
• Poor verification of root cause or solution
• No schedule to track deliverables for verification
• No leverage of knowledge or defects discovered on similar products

There is often some urgency to find and resolve the issue, and the sourcing team should be wary of quick solutions proposed by the supplier without thorough analysis. The supplier may propose additional testing and inspection on their side to check for defects and prevent future escapes, but the sourcing team should push for permanent process changes to prevent defects from occurring in the first place. Finally, the corrective action request should not be considered closed until improvement has actually been measured and noted by the customer. A corrective action plan is not the same thing as the successful implementation of that plan.

Failure to promptly or successfully resolve an issue is a serious matter that could jeopardize the relationship between a supplier and a customer, particularly for repeat occurrences of the same issue. The sourcing team should closely monitor supplier investigations and their implementation of corrective action plans. Although any failure to meet customer requirements is a concern, issues will happen from time to time, and a supplier who takes swift and effective action on behalf of the customer can still be considered a valued partner.

My next column is on DOE. This was my secret weapon for problem solving when I first encountered printed circuit process problems. Further reading and resources is the ASQ, American Society for Quality^[2] . Villanova University has a number of on-line six-sigma and Agile courses and certificates^[3]. 

References 

1. Kepner-Tregoe: www.kepner-tregoe.com
2. ASQ: www.asq.org
3. Villa Nova University www.villanovau.com

Happy Holden has worked in printed circuit technology since 1970, with Hewlett-Packard, NanYa/Westwood, Merix, Foxconn and Gentex. He is the co-editor, with Clyde Coombs, of the Printed Circuit Handbook, 7th Ed. To contact Holden, click here.