MT475 Unit 2 Seven Quality Tools Exercise Example

Count Says Scratches, Cost Says Wrong Size: An MT475 Unit 2 Quality Tools Exercise on a Skylight Plant's Remake Log

[Student Name]

Purdue University Global

MT475: Quality Management

Unit 2 Quality Tools Exercise

[Instructor Name]

[Date]

Overlook Skylight Company, its remake log and every figure are composite; the sources cited are real.

The Data

Overlook Skylight Company, a composite Lancaster, Pennsylvania maker of residential skylights, logs every unit it must remake before shipping, with the cause, the shift and the cost of the remake. This exercise uses thirty production days of the spring peak, when the plant built 19,200 skylights and remade 806 of them, a remake rate of 4.2 percent. The tools appear in the order the analysis needs them, not the order a textbook lists them.

The Check Sheet

A check sheet tallies the 806 remakes by cause and shift. It comes first because every later chart is built from it. Tague (2005) describes the check sheet as the simplest way to turn observations into data that can be analyzed, and the plant's remake tickets were already a check sheet in all but name. Six causes account for every remake: scratched glass, 231; cracked frame corner welds, 172; wrong size, 139; missing flashing kit parts, 98; voids in the glazing sealant, 87; and hardware faults on venting units, 79. The shift tally shows weld cracks concentrated on the second shift, 108 of 172, while the other causes are spread fairly evenly.

The Count Pareto

Ranked by count, scratched glass leads at 231, weld corner cracks follow at 172 and wrong size is third at 139. The top two causes together account for 403 remakes, 50 percent of the total. A team reading only this chart would attack scratched glass first.

The Cost Pareto

Ranked by cost, the order changes. A wrong-size skylight must be rebuilt from new extrusions and a new insulated glass unit, about $186 each, so the 139 wrong-size units cost $25,854. A cracked weld is cut out and rewelded, about $64 each, so 172 cracks cost $11,008. A scratched glass unit is reglazed, about $38 each, so 231 scratches cost $8,778. Sealant voids cost about $3,915, missing flashing parts about $2,156 and hardware faults about $1,422.

The cause that is third by count is first by cost, by a wide margin.

Juran's principle of the vital few and the trivial many, which the Pareto chart puts into practice, tells a team to work where most of the problem lies (Juran & Godfrey, 1999). The two charts show that "most of the problem" depends on whether problems are counted or priced. For a plant manager with a budget, the cost Pareto is the better guide, and it points to wrong size.

The Histogram

A histogram of corner weld break strength, from 120 destructive tests during the month, shows most welds breaking between 1,100 and 1,400 pounds of force, well above the 800-pound minimum, but with a long lower tail: eleven welds broke between 700 and 900 pounds. A symmetric distribution would suggest random variation around a good average. A long lower tail suggests a specific cause that sometimes weakens welds.

The Scatter Plot

To look for that cause, a scatter plot sets hours since the welder's heating platens were last cleaned against weld cracks per shift. The points rise together: shifts that ran longer after cleaning had more cracks, with a correlation near 0.6. Second shift cleans the platens only at the start of the day, so its welds are made later in the cleaning cycle, which may explain its share of cracks.

The Fishbone

A fishbone diagram for wrong-size skylights runs six branches: methods, machines, materials, people, measurement and environment. Its causes include dealers' rough-opening measurements entered as outside frame dimensions, a saw stop tape that has never been calibrated, order entry staff retyping sizes from faxed forms and similar part numbers for 22-inch and 27-inch frames. Three are circled for testing: unit confusion at order entry, the saw stop calibration and retyped sizes. Montgomery (2020) describes the cause-and-effect diagram as a way to organize hypotheses for investigation, not to establish which is true.

Reading the Charts Together

No single chart tells the whole story. The check sheet located weld cracks on second shift; the histogram showed that weak welds were a distinct group rather than ordinary spread; the scatter plot offered a reason, platen cleaning, that fits both. For wrong size, the cost Pareto set the priority and the fishbone set the questions. Used in sequence, the tools narrow the search from 806 remakes to two testable hypotheses, which is all a month of data can honestly do.

What the Tools Cannot Prove

The tools describe and suggest; they do not prove. The scatter plot shows an association between platen cleaning and cracks; it cannot rule out a third factor, and second shift also has newer operators. The fishbone lists possible causes for wrong size but does not say which matter. And thirty days in the spring peak may not represent the whole year. The next step is a controlled test for each circled cause, such as cleaning platens every four hours on second shift for two weeks and recounting cracks.

References

Juran, J. M., & Godfrey, A. B. (Eds.). (1999). Juran's quality handbook (5th ed.). McGraw-Hill.

Montgomery, D. C. (2020). Introduction to statistical quality control (8th ed.). Wiley.

Tague, N. R. (2005). The quality toolbox (2nd ed.). ASQ Quality Press.