I wrote this and presented this at the 1990 ASQ Quality Congress but thought some may find it interesting and informative.

ABSTRACT

This paper describes a real life application of Quality Function Deployment (QFD) to a factory of the future in the Aerospace and Defense industry. The factory of the future is a proposed high rate low cost microwave hybrid manufacturing facility. It is felt that without the application of QFD the facility will never achieve its goal of high rate and low cost. More a diary than a historical account, this paper describes an application that is still in progress. The completion of the project is planned for 1992.

Besides the goals of high rates and low cost mentioned above, several other reasons for the choice of QFD as a tool for managing this project will be covered. Some examples are: customer requirements, better company communications, and the integration of customer requirements into production requirements.

This paper discusses the resources required by QFD on this project. These resources include personnel needed for cross departmental teams; personnel needed for formal and informal facilitation and training; time spent in training and team meetings; and other costs such as software, tools, and equipment required to implement process controls.

A description of the QFD tools (matrices) employed and how they were employed, as well as, why certain QFD tools were not used and why others were modified to better fit the situation is discussed. This project incorporated a mix of the ASI basic waterfall of four basic, but large, matrices and the GOAL/QPC approach of 32 smaller matrices.

An in-depth look into the obstacles encountered will be presented. Included will be the understanding and overcoming of any resistance to this new approach by management, as well as the removal of any barriers found between departments. The obstacle of ignorance toward QFD is also discussed, with the approaches used to educate all of its advantages explained.

Finally, this paper provides the complete description of the results achieved. This includes the accomplishments, failures,

and the current status of the project. Also it will look at the future plans and goals that the project has and will try to project a completion date when the development of new matrices will essentially stop and the maintenance of the old ones begin.

INTRODUCTION

In October of 1988, a team of engineers from Hughes Aircraft, Delco Electronics, General Electric, AT&T, and Pacific Monolithics, was formed to write a proposal for a factory of the future. The Air Force had requested bids from several companies to assemble a manufacturing process capable of building Microwave Modules (Hybrids) at a rate of at least 1000 Modules a day and at a cost of less than $400 each. The intent was to use this module in a low cost active array radar system that was as good as or better than the current radar systems found in today's aircraft.

Hughes calculated it could meet the Air Force challenge by using the right mix of technology and people. A proposal team of experts from all areas of engineering and manufacturing, including a group of quality engineers was pulled together. From the start, the program management understood the need to design and build quality into the product. In our case management decided to get involved and stay involved supplying momentum to this quality effort throughout the program. They felt that a new management approach was needed to obtain high rates, low cost, and maintain the product quality and reliability. Management believed that by insuring quality in the product, rework and scrap would be reduced or eliminated, which would cause yields to increase and reduce costs to our customers. This new approach required quality involvement at every stage of the program. Involvement at the proposal writing stage, to designing of the module, to high rate production of the final product. The quality team introduced the rest of the staff to Quality Function Deployment (QFD) .

The Hughes team was one of the teams awarded a contract to develop this module production line. The project is now in Phase 1 and 2 of 4, the Module Design for Affordability Manufacturing Phase. Currently there are ten QFD teams developing matrices, with several more teams anticipated before the completion of Phase 4.

WHY QUALITY FUNCTION DEPLOYMENT (QFD) ?

There were three reasons for using QFD on the project: customer requirements, improved communications, and better integration of customer requirements into production requirements. The customer did not specifically call out QFD, but his request for quotes placed a heavy interest in the implementation of a Total Quality Management (TQM) System.

The Department of Defense is restructuring its management system toward a TQM system. Project management envisioned that a team of quality experts could pull together the best tools for the TQM system. The TQM system would require process controls with little or no formal inspection of the product. Convincing the Air Force that the process controlled the product quality and that we had control of the process would be critical.

In addition, the product needed to be characterized by a high rate process that could be manufactured at a low cost. Both of these require designing quality into the product. Designers now realized they had another customer, process or manufacturing engineering. To address these diverse customer requirements and not lose sight of the external customer needs, QFD became a major element of the TQM system.

The second reason for QFD -communications -became apparent right at the start of writing the proposal. We, the internal customers, were transmitting complex information (Whats) and ideas (Hows) thousands of miles, from Delco in Indiana, to Hughes in California, to General Electric in New York, to AT&T in Pennsylvania, to Pacific Monolithics in California, to Hughes in Arizona, and then back again. A tool was needed to keep the team focused and on track with the customers needs. Team members familiar with QFD knew that QFD could fill this need. The matrix structure of QFD allowed each team member to bring his or her expertise to the table and have direct input into the design and manufacture of the module. Everyone's contributions were recorded on the matrix and reviewed by the team through this process, lines of communication were clearly established and exploited. Once completed the matrix represented the best of all areas that met both the internal and external customer needs.

The final reason for QFD is better integration of customer requirements into production requirements. QFD requires teamwork and simultaneous engineering since each company, group, and department involved in building the product is another smaller, but just as important; customer. As customers they have wants (or Whats) that need to be filled by their suppliers. Design engineering can always design a product that the external customer wants but if manufacturing cannot produce it, in volume and at a low cost, then the company will fail in its attempt to produce the product. In our case we already had a working product, but it was too costly and volumes were too low to meet our external customer requirements. Some of this excess cost and time was due to old technology, but most was due to a lack of understanding of the internal customers (other companies, groups, and departments) requirements. Through teamwork and simultaneous engineering each area was represented on the QFD team. Each area had an input into how we were to meet our external customers needs (Whats), as well as that external customer (see Fig. 1) . Included were the difficulties associated with each "How" so that the team could decide the optimal approach for solving and meeting our customers needs.

RESOURCES REQUIRED

The primary resource needed was our experts and their time to work on the team activities. The first matrix developed was the "Requirements Matrix" which laid the foundation for development of new teams and matrices. Several of the Requirement Matrix team members were on subsequent matrix teams. In fact, at least one member of the Requirements Matrix team is found on every team developed. This aided in keeping the voice of the customer moving into and through each subsequent matrix. Each team is made up of four to thirteen members. As each team became larger, it was harder for that team to reach a consensus. with the size of each team being restricted, it was extremely important that we selected the right team members. Each team member had to bring to the team an expertise that was unique and important in fulfilling the teams goals. On our Requirements Matrix we had thirteen team members from the areas listed in Table 1.

Each member received four to eight hours of training on QFD before any work started on the matrix development. In fact everyone working on the program was given this training. In some cases, a second day of training was included. On the second day, Taguchi methods of Design of Experiments was covered. A select few went through several weeks of Design of Experiments (DOE) training covering Taguchi, Box, and Shainin methods. This extensive training augmented the QFD process. Once trained, the teams met for two hours twice a week, which often requires the use of other Quality Tools such as DOE, until the initial matrix was completed in about two to three months. Thereafter, the frequency and duration were reduced to an as needed basis. Currently once every two weeks the House of Quality team formally meets to update the matrix. This matrix maintenance will continue for the life of the program. QFD is never completed, it is continuously refined and improved.

Methods Used

The program TQM (QFD) facilitators were trained in two basic approaches to QFD. The two were the American Standards Institute (ASI) approach, which uses a few large matrices, and the GOAL/QPC approach, which uses several smaller matrices (see Fig. 2) . Adhering to a principle stated by Bob King (President of GOAL/QPC) , we used "the pieces" of both methods that worked best for us. Our method employed a foundation of one small matrix (the Requirements Matrix) .The matrix was developed to gain the "Voice of the Customer" and determine critical design requirements (see Fig. 3) .Then a functional tree diagram was developed to aid in selecting design concepts and technologies to be used for each function. At this stage, about ten specialized materials matrices were developed. These matrices were used in trading off different design concepts or for developing the processes to be used with the above technology. In retrospect, we trained our teams basically in the ASI approach and used more of a GOAL/QPC approach in the- actual designs.

Obstacles

Initially, ignorance of QFD was the biggest obstacle. Very few of the proposal team members knew what QFD was. Our saving grace was the customer interest in three basic elements.

1.              TQM system (understand that TQM includes QFD)

2.              The requirement for a low cost module (less than $400.00)

3.              The requirement for a high rate manufacturing operation (over 1000 modules per day)

As mentioned before, these three elements of the request for bid sent a direct message to management: control of product quality is mandatory. Thus management supported the quality group efforts for training and implementation of the TQM system which included QFD. without managements support and active participation, QFD had no real chance of being effective.

A second obstacle was the resistance to change that was felt from many areas or departments involved. QFD was new to each area and that usually means resistance to change. Most have found ways of doing their job, and doing it fairly effectively. They understood that things could be improved but unless "others" followed suit, their efforts would be wasted. Management led the way, convincing others that all involved should listen and learn the tools needed to complete their jobs in the best possible way. Management also participated in both the training and design of the matrices. This role modeling had high payoffs. Engineers entered the training with open minds and exited with new tools to help them do their job better. The program, in turn, now had a new way of doing business.

Do we still find resistance? Yes, some. Just when one group thinks it is finished and does not need the QFD matrices, it discovers other ways to improve the process. This means reassessing the QFD matrices to see how the new approach will effect the overall system and the customer needs. There is also resistance by the teams to integrate their jobs with other TQM tools such as DOE and SPC. Although many are using QFD in their jobs, other elements remain to be integrated. There are elements in the lower level QFD's that only a DOE will show the true interaction between. It is a logical step that tends to be missed. Our planned solution is to train the staff sufficiently in DOE to make that logical step, and then do the same for SPC.

RESULTS

To date, the QFD effort has been very successful. All the disciplines involved in the program have been trained on QFD, and over 10 teams (matrices) have been developed or are being worked (see Fig. 4). The QFD process has also been very successful at helping us truly hear and understand the voice of the customer. By clearly and effectively communicating with the customer through the QFD process, we are now better equipped to satisfy his needs. QFD is definitely putting us and our customer in the "win-win" situation. QFD is also fostering teamwork and simultaneous engineering, the benefits of which are always desirable.

Even though we have been very successful there have been failures. Some teams staggered because of unclear problem and goal definitions. It is very important that this definition be done when the team is first organized. An overall lack of experience by our team members and facilitators was another problem. This was overcome largely by the robustness of the QFD process and the creativity and cooperation of the team members. We also discovered that the size of the matrices and the number of team members must be manageable. As the matrices get larger, and the participants increase, evaluating all the relationships becomes overwhelming. Although the size of the matrix varies and must not be reduced only to reduce the workload, a 25 by 25 matrix is very manageable and capable of yielding excellent results.

As mentioned earlier another failure was the lack of integration of all the quality tools into the engineer's daily job. Only further training and time will tell if we fail at this task. However, successful and full integration of all the quality tools into the engineer's daily job will assure quality is being built into the' product the first time.

The program is still in the development stage with a lot of paper being scrapped, but no losses of materials and no unusable production equipment purchased. The manufacturing matrices are being developed with the controls and verification matrices roughed out. By the middle of 1990, most of the QFD matrices should be complete, with only continuous maintenance needed to sustain their impact on the program. The first production part should be coming off the new process and an evaluation of the benefits of the whole TQM system can start. We anticipate that a few adjustments to both the process and the QFD matrices will be made to assure our efforts, and product, reflect our customers needs.

BIBLIOGRAPHY

1.) American Supplier Institute, Quality Function DeDloY!l1ent OFD. Dearborn,MI. ASI Press 1988

2.) King, Bob, Better Designs In Half the Time. First Ed. Methuen,MA: GOAL/QPC, 1987

3.) Re Velle, Jack B., The New Quality Technology. Los Angeles,CA: Hughes Aircraft Company, 1988

LCS; 600-10-991

Well there you have my use of Quality Fuction Deployment and the House of Quaoity. If, you have questions or comments please feel free to contact me by leaving a comment below, emailing me, calling me, or leaving a comment on my website.

Bersbach Consulting
Peter Bersbach
Six Sigma Master Black Belt
http://sixsigmatrainingconsulting.com
peter@bersbach.com
1.520.829.0090