Originally Published MDDI February 2005
Product Development Insight
Design for R&R: A Product Quality Methodology
Design engineers need a quality system that enables them to develop a systematic development process to meet increasing product requirements.
Today's mechanical design engineers are tasked with much more than conceiving and designing a functional product with structural integrity. Well-designed products are a culmination of many integrated factors, including nonengineering disciplines such as industrial design and human factors. Methodologies and techniques have evolved to guide engineers through the development process in pursuit of designing the perfect product.
One emerging methodology is design for repeatability and reproducibility (DFRR). Using DFRR, product quality and production yields can be optimized and product development cycles minimized. But perhaps more importantly, when used correctly, DFRR can enlighten engineers about the principles behind achieving a true quality mind-set.
In the late 1980s, a small medical device manufacturer instilled a mind-set into its employees that formed the basis for DFRR. The manufacturer also implemented practices that resulted in dramatic dividends. At the time, the selling price for a skin stapler (a handheld wound closure device) was about $36 per unit. But this innovative company focused on quality, driving the price down significantly and capturing 20% of the market share. Today, the selling price for a skin stapler hovers near $6 per unit. How did this happen? DFRR evolved from the fertile environment that produced these results. This article discusses the methodology and its potential implications for the design process.
DFRR is a methodology that aids in establishing a systematic development process. However, DFRR holds greater implications and benefits than initially meet the eye. The basic goal of DFRR is to understand and establish process control on the component level before moving on to the assembly level. This is done using statistically proven, accurate inspection data to calculate process capability.
Establishing Process Control. Process capability and process control are the backbone of product quality. This is especially true for the development and production of complex medical instrumentation, such as handheld surgical instruments with multiple mechanisms (see the sidebar, “Multiple Mechanisms”).
DFRR enables design engineers to focus on optimizing process capability and maintaining process control. Every process inherently has some degree of variation. For example, the design-for-manufacturability methodology emphasizes that the design engineer's role is to produce component designs that enhance process stabilization.
Capability Index. The process capability index is a statistical means to numerically quantify process stabilization. An inspection procedure is needed to generate the data used to calculate the index. The procedure will involve inspectors, measurement equipment, and a measurement method. DFRR takes the designer a step back, prior to the inspection, by emphasizing design elements to enhance the stability of the inspection procedure. As the adage goes, misleading data are worse than no data at all. Process capability studies done using inaccurate data will lead engineers down a trail of frustration and result in an extended development cycle.
The foundation of DFRR is the Gauge Repeatability and Reproducibility study (Gauge R&R). The term gauge refers to any device or equipment used to make a measurement. Quality engineers routinely use Gauge R&R studies to determine process capability. However, design engineers can aid quality engineers by gaining practical experience performing such studies. Facing the challenges of obtaining an acceptable Gauge R&R outcome will influence designers to be aware of the inspection process when designing individual components. Many good articles and manuals have been written to detail the intricacies of performing and using Gauge R&R studies. In addition, sophisticated software is available to aid in the analysis of the interrelationships among dimensional tolerancing, process capability, and Gauge R&R.
Process Variation. Gauge R&R studies analyze the measurement process variation by separating it into two specific areas:
• Repeatability, or the variation of measurement of a gauge.
• Reproducibility, or the variation of measurements by operators.
Every set of inspection data for a particular process contains both actual production process variation and measurement process variation. Separating out the measurement variation is the purpose of conducting a Gauge R&R study.
|Figure 1. A Gauge R&R of 60% means that the results from two different inspections could vary as much as 60% of the tolerance (click to enlarge).|
A study may involve either two or three operators who inspect one feature dimension on each of 10 individual sample parts. The sample parts are identified before the inspection process starts. The operator must be unaware of the identity of the randomly chosen sample at the time of each inspection. Inspection data are recorded on a worksheet for later statistical analysis. The output of the statistical analysis is typically reported as a percentage of measurement variation in relation to the total dimension tolerance. Thus, a Gauge R&R study result of 60% means the process used to measure the particular feature dimension had a variation of 5.15 s (5.15 s equals 99.0% of the bell curve), which encompassed 60% of the total tolerance range (see Figure 1).
Therefore, if the exact same part were inspected and then reinspected, the measurements from the first inspection and the second inspection could vary as much as 60% of the tolerance. It would be unwise for an engineer to make decisions concerning acceptance or rejection of component lots if the inspection method could vary as much as 60% of the tolerance range.
Misleading Data. A high percentage Gauge R&R result can lead to misleading data. If the same part is inspected twice, the first data point may be within specification and the second data point outside of specification. For example, imagine Part A has a feature dimension of 0.350, ±0.005. The tolerance range is 0.010, meaning that the acceptable tolerance falls between 0.345 and 0.355. A Gauge R&R study result of 60% means that the variation due to the inspection method is 0.006. Using these numbers, the following inspection scenarios are possible (see Figure 2):
|Figure 2. An inspection process with a Gauge R&R of 60% can result in misleading data, enabling one measurement to be within specification while another is outside specification (click to enlarge).|
• The first inspection of Part A finds a measurement of 0.348, which is within the part's specification.
• The second inspection of Part A finds a measurement of 0.342, which is outside the part's specification range.
As a general rule of thumb, an acceptable Gauge R&R result should be less than 20%. However, this is only a guideline. Each individual dimension should be evaluated on its own merits. Engineers who have an intimate understanding of the components they design will be extremely helpful in determining an acceptable Gauge R&R result. In addition, it will quickly become apparent that as the dimension tolerance becomes tighter, achieving an acceptable result becomes more difficult. Plus, with a tighter tolerance, the measurement equipment must possess higher resolution. The common recommendation for equipment resolution is that it should be at least 10 times finer than the tolerance range. So, for a tolerance range of 0.010, the equipment resolution should be at 0.001 or finer.
The results of a design engineer's first Gauge R&R study are usually quite revealing. Many design engineers are routinely involved in deciding the disposition of component lots that have been rejected at incoming inspection. How reliable are those data? It is not uncommon for a novice to obtain a Gauge R&R result higher than 100%, which would mean the measurement variation is greater than the tolerance range. It's a chilling experience to realize that decisions concerning rejected incoming materials were predicated in faulty data. It's no wonder that some organizations operate at process yields of about 80%. How many times have engineers reviewed so-called out-of-spec data from a first-article inspection report, only to track down the inspection technician and get different results upon reinspection? Actual component parts don't have straight or perpendicular edges like a CAD model's portrait does. The inspector's interpretation largely affects a first-article inspection.
Gauge R&R brings a new perspective to the validity of first-article inspection reports as reflections of meaningful data. In fact, there is a good case to be made for design engineers to perform the first-article inspections. Firsthand inspections made by witnessing the true part on a comparator or under a toolmaker's scope are helpful in understanding the actual condition of part features and, ultimately, guiding a product successfully through the development process.
Tips and Tools. Hands-on experience in performing Gauge R&R studies will enhance a design engineer's understanding of the inspection techniques and skills required to implement the DFRR methodology. For DFRR to work, it is also important to understand the components' manufacturing processes and typical physical feature attributes. For example, an edge on an injection-molded part where a shut-off is located would not be a good candidate for an inspection edge. As a tool wears, inconsequential flash can occur at such an edge, producing an unreliable inspection point. Another design trick is to add bosses or ribs specifically to help in providing datums to repeatably position parts for inspection. Adding reference marks or features is another handy design element used in DFRR.
Quality Is Key
The immediate and tangible advantages of implementing DFRR are apparent. But the major advantage to DFRR is hidden below the surface. Integrating quality into the fabric of an organization has become a major focus of American industry in recent decades. In the 1970s and 1980s, Japan surprised the automotive world by capturing a market share never imagined by American automakers. Japan again overtook other American-stronghold markets by supplying superior quality at lower cost. In response, total quality management and the six-sigma initiative arose. These systems are intended to be more than methodologies and emphasize more than just a strong quality manual backed by a quality department. Rather, they aim to instill a philosophy of quality that permeates the entire organization and involves all aspects of the business.
This has often been a difficult undertaking for American manufacturers to implement. The rewards of quality are not easily quantifiable, especially in the near term when many management decisions are made. Quality has been paid much lip service in the form of adherence to the quality manual to obtain ISO certification. But true quality encompasses more than adherence to a book of rules. It involves cultivating the values, attitudes, and character that come with an intimate understanding and tangible experience of what actually provides quality. A major key to quality is optimizing a process for minimal variation. This is achieved by the simple, yet extremely powerful, principle of the standardized process. The standardized process defines and executes a process to where it can be repeated and reproduced. This sounds simple, but it is most profound.
A working knowledge of inspection procedures and equipment, a good understanding of component manufacturing processes, and creativity in design are all major contributors to successfully implementing DFRR. The advantages of DFRR are maximized when the methodology is implemented early in the development cycle. Engineers with the awareness of DFRR, even in the conceptual stages, can assist their development teams in minimizing the development cycle, ensuring product quality, affecting the bottom line, and most importantly, bringing a quality mind-set to the organizations they serve.
Copyright ©2005 Medical Device & Diagnostic Industry