Total Quality Management 305

To effectively "slay the sacred cows" requires changing direction and learning to think in new ways. The following is a list of areas that will constitute such re-direction:

The process of total quality management begins with an involve­ment in the determination of customer requirements. This requires an understanding of how the product is intended to be used and how it will in fact be used. This requires excellent communications and active listening. Quality Function Deployment (QFD) offers a structured process for having customer needs drive the design of the product and the design of the process for making the product.

With full recognition of customer needs and expectations, there is a need to recognize how this is to be integrated into the design of both the product and the process. Engineers must be fully conversant with the capabilities of both suppliers and the com­pany. The new techniques to facilitate this is concurrent and participative design and engineering.

Variability here is from lot to lot, from piece to piece, and over time. This applies not only to dimensional and performance parameters but also to such areas as lead time and yield and shrinkage. Determining the root cause for variability is an excel­lent means of improving absolute performance.

The measure of variability is process capability, also known as Cp and Cpk. It is beyond the scope of this presentation to detail the methods of calculation, except to indicate that this is a measure comparing process capability with specification limits. A Cpk of 2.0 means that the variability of the process uses one-half the specification tolerance. Some companies are attaining a Cpk as high as 5.0 on critical metrics.

The best way to deal with defects is to prevent them. While it is certainly desirable to avoid defects from reaching the customer, external and internal, the ideal is to avoid producing defects. This means that there must be a ...

The accepted approach to monitoring process quality is Statistical Process Control (SPC). A simpler method is "Precontrol." Developed in the 1950s by Frank Satterthwaite of Rath and Strong, the method has several advantages. First, the mechanics of precontrol can be taught to anybody in industry, including line operators, in less than ten minutes'. There are no complicated calculations. The rules are simple and transparent. Charting precontrol greatly eases the operator's burden. A full description of the methodology is beyond this paper. The referenced text will give an overview of the method.

The use of either statistical process control or precontrol should

have as its objective the elimination for the need of SPC or precontrol. These techniques should be intended as the foundation for implementing "Quality Cycles." Quality cycles are the deter­mination of how many pieces a process can deliver before an adjustment is needed. For example, if a die can produce 250,000 pieces before it needs to be sharpened, or a grinding wheel can produce 500 pieces before the wheel needs to be dressed, then the die should be pulled at a quality cycle of about 240,000 pieces and the wheel should be dressed after producing about 480 pieces. This is a basic view of the use of quality cycles. The method offers considerable simplicity over any method that lets the product determine the need for correction and/or adjustment. The actual cycle is a function of variability targets established and the capability of the process to compensate for change in the process.