Anyone who has faced a production problem with a need to solve it by using production data can relate to the notion of a brain teaser. The brain teasers presented here are based on real-world situations encountered by workers in manufacturing environments. The brain teasers have three parts: (1) the situation, (2) available data or other supporting information and (3) questions that various workers need answered for continual improvement. Recommended solutions follow in the next issue and on the Web at Quality Online (www.qualitymag.com).

Situation

Nelson is a senior engineer for a company that produces process pumps for a variety of industries. After these pumps are assembled, they are tested for performance. During the performance test, the pumps are tested for leaks. At a recent meeting about the quality of pumps, the customer complaints summary listed “leaky pumps” as the top complaint for a recently introduced pump model, PWX350. This model started shipping three months ago and replaced an earlier model (PVX300). There were no complaints about leaks for the earlier model. Because the leaky pump problem is new, Nelson wants to put a team together to find the cause of the leaks and determine solutions. Before the first team meeting, Nelson requests data about the leak test results for the new pump as well as for the earlier model pump.

Available Data

Enlarge this picture

Data on pump model PWX350 for the first 13 weeks of production are summarized by week in the table, “Pump Model PWX350 Leak Test Results” and data on pump model PVX300 for 22 weeks prior to the introduction of the new model are summarized in the table, “Pump Model PVX300 Leak Test Results.” Nelson took these data to the first meeting, so the team could start working on the leak problem.

Questions

1. What is the behavior of pumps with respect to leaks for both models?

2. Do these data give credence to the recent customer complaints for leaky pumps?

3. If Nelson and his team want to find out how to solve the problem with leaky pumps in the new model, what additional data should they request?

4. What monitoring plan for leaky pumps should be put in place?

Answers to January Brain Teaser

Enlarge this picture

As the new production manager for a company that manufactures overhead conveyor systems, Meredith recognizes that the assembly of these systems depends on all components meeting specifications on several critical characteristics. During her orientation she was told that they rarely have difficulty meeting specifications, but there have been discussions about finding components out of specification but no mention of available process control information.

Q: Make an appropriate average and range chart with the data provided; use the limits in the table provided. What can be interpreted from the chart with the limits provided?
A: The average and range chart with a subgroup size of n=3 using the limits provided in the table shows one average below the lower control limit and all of the ranges below the average range. From this chart, it is clear that the process variation is less than the variation used to determine the limits. See the graph, “Thickness for Component A Using Limits Provided in the Table.” An investigation of the specifications of 1.2 ±0.009 quickly reveals that these limits are based on the specifications and not on the data. Limits calculated from the specifications should not be used for analyzing process behavior.

Enlarge this picture

Q: Using limits calculated from the data, is thickness operating predictably for Component A?
A: The average and range chart in the graph, “Thickness for Component A Using Limits Calculated from the Data,” shows that the process is operating unpredictably on the average. There are two average values above the upper control limit and one average below.

Q: Is it true that all of the values for thickness for Component A meet the stated specifications?
A: Yes, the “Histogram of Thickness for Component A” shows that all data values meet specification.

Q: What is the issue with the original graph that Kevin showed Meredith?
A: The original graph used control limits calculated from the specifications. This practice is not correct because the process behavior chart is designed to determine process behavior independent of specifications. The calculated limits use the amount of routine variation that exists in the process itself and not the amount of variation that is considered acceptable from the specifications.