Daniel Schaeffler Daniel Schaeffler
President

Golf Shots and Manufacturing Variability

August 1, 2018
0
Comments
Fig. 1
Every stamped part contains at least some sources of variation that will never be eliminated—perhaps from the sheetmetal, lubrication, stamping process or any one of the many other aspects of the forming system. It is important to understand the sources of variation so that we can design our forming system to be more tolerant of inevitable variability inherent in the inputs.

All input process settings revolve around target value and a tolerance—the basis for one of the largest sources of variation. Consistency of sheetmetal strength is affected by variation in alloy melt chemistry, thickness reductions during rolling and annealing temperature, to name just a few of the inputs. In turn, consistency in strength is one of many variables that affects the dimensional precision of a formed part. If comparing the measured strength over many coils against how often those values occur, they would start to resemble a bell-shaped curve, or what statisticians call a normal distribution. A tight bell curve indicates little variation in measured values, with most readings concentrated near the average value. A wider curve indicates a greater deviation from average.

Let’s put this into different terms. The median driving distance off the tee for a (non-pro) male golfer is 220 yards. This does not mean that all shots will travel 220 yards, but, instead, half of his shots will travel longer and half will fall short. For a consistent golfer, more shots will travel closer to this average rather than spread out over a greater distance.

Now imagine sand traps at 170 and 270 yards. A more consistent golfer able to hit all shots within 50 yards of his average won’t need to change his approach. However, the golfer whose skills would otherwise result in a driving distance of, say, between 160 and 280 yards either will need to adjust his process controls or accept that some shots will fall in the bunkers.
Fig. 2
Similarly, consider sheetmetal production of a grade intended for a part requiring strength between 170 and 270 MPa, representing the lower and upper specification limits. That does not mean that production cannot result in strengths falling outside of this range—just that the product shipped to the customer must meet the range constraints. Fig. 1 depicts a supplier with a production capability that takes up the full range of allowable strength levels. This supplier’s approach satisfies the application needs, at least in terms of strength.

Each company uses different equipment and different process-control capabilities, which affect the property variability of shipped coils. Fig. 2 shows the property variability of three suppliers. All can satisfy the strength requirement. The average strength of the product shipped from mill 2 is lower than that from mill 3. Remember that lower strength commonly is associated with a more formable product. However, the parts manufacturing company still may prefer to receive sheetmetal from mill 3. Even though the material is stronger and probably less formable, it comes in a tighter range. With this reduction in incoming variability, the stamping plant may better control the parts it produces.
Fig. 3
A mathematical term, standard deviation, describes this variability. This is represented by the lowercase Greek letter sigma, or s. Knowing the standard deviation lets you estimate the likelihood that any one test result will fall within tolerance, and its proximity to the average. Two-thirds of all test results will fall within one standard deviation of the average (Fig. 3), meaning that one-third of all test results will fall outside of this range. However, more than 95 percent of all results will fall within two standard deviations of the average value (average ±2sσ). A range of ±6sσ (six standard deviations) encompasses 99.73 percent of all results. This corresponds to a Process Capability Index, or Cpk, of 1.0. A Cpk of 1.33 is associated with readings within ±4sσ of average (63 out-of-compliance results for every one-million tests), while a Cpk of 1.67 corresponds to all readings falling within ±5sσ of average (one defect per one-million tests).

Tighter process controls will decrease the standard deviation. There may be valid reasons why your supplier is not implementing all steps to reduce variability when it already can meet the specified tolerances. For example, the alloying-element range in the melt specification likely is wide to enable production of several grades from that one set of chemistries. This allows for production efficiencies and provides for steady-state processing conditions that will improve process quality. Similarly, adjustments in the annealing temperature cannot be realized instantaneously in the furnace. Rapid changes bring a greater risk of undershooting or overshooting the targeted value, leading to a wider property variation in the finished product. MF

Learn about high-strength-steel grades and formability, tool steels and coatings, presses and die design, and effective lubricant strategies at PMA’s two-day Stamping Higher-Strength Steels Seminar in Nashville, TN, September 12-13. Visit www.pma.org for details and registration, or contact Marianne Sichi at msichi@pma.org for information.

Industry-Related Terms: Die, Forming, Thickness, Tolerance
View Glossary of Metalforming Terms

 

See also: Engineering Quality Solutions, Inc., 4M Partners, LLC

Technologies: Quality Control

Comments

Must be logged in to post a comment.
There are no comments posted.

Related Content

Subscribe to the Newsletter

Start receiving newsletters.