Understanding control charts
Q: I have about 10 paper testing instruments that my organization uses to measure paper standards. Every different location (there are eight total) uses the same type of instrument, test procedure and the same paper standard out of the same lot. This reduces the variability as much as possible.
The labs collect 10 test data points and send me the results. I am new and hate to admit that I am lost. I need to know which statistical process control chart to use that I can dump the data into and, at the same time, save time. I would like to see the center line, upper and lower limit lines, and when a test or instrument result is out of the limits. The last person also showed the limit lines of the 10 data points from each instrument. This was his designed program. When he left, he took the program with him. I also have a low budget and don’t want to be reprimanded for both problems. I would appreciate all the help you could possibly provide.
A: Control charts would be an excellent tool to determine when a test or instrument result is outside limits. One approach would be to chart the data in three different ways:
- For each instrument, use an individual values/moving range (I-MR) chart to quickly determine if any individual data point is outside limits.
- For each instrument, use an average/standard deviation (X-bar-S) chart with the subgroup size of 10 test data points to determine if each particular instrument is performing as it should on that particular day.
- Take the averages from each subgroup from each instrument on a given day, and use another X-bar-S chart to determine if the instruments as a whole are performing as usual.
To demonstrate, electrical sensitivity data in millivolts were obtained on a device and analyzed as an example of the proposed techniques.
1. I-MR chart on each measurement.
Plot all the points obtained on each instrument on separate charts for each instrument. Figure 1 shows a couple slightly outside-of-limits points that were discovered for instrument No. 10.
2. X-bar-S chart on each subgroup of 10
When plotted in subgroups, however, the process for instrument No. 10 appears to be in control. Sample 1 in Figure 2 represents an average of the first 10 observations; sample 2 is the second 10 observations and so on.
3. X-bar-S chart on averages from all
instruments over time
In Figure 3, the point for test day one represents the average of all the average values from each instrument on the first day. This chart could be used to determine if there are any gross changes over time in the paper standard itself, or if there is generally a change in variation over time between all the instruments’ results taken collectively.
Through the use of this set of control charts, an analyst will be able to see data points and data sets outside historical limits. In addition, control charts are a nice way to detect trends even before going outside limits.
Senior manager, quality engineering and risk management
ISO 9001:2015 impact
Q: With the upcoming revision to ISO 9001, how will ISO/TS 16949, the technical specification containing quality management system requirements for the application of ISO 9001:2008 in the automotive sector, be affected? Will it also be revised in 2015 because it includes all of the ISO 9001 clauses?
San Luis Obispo, CA
Q: The International Automotive Task Force (IATF) had requested and received a waiver in 2012 from the ISO Technical Management Board allowing it to continue to use the current version of the ISO 9001 standard in ISO/TS 16949 indefinitely after the release of ISO 9001:2015.
However, in December 2014, the IATF announced it has formed a team to develop a design specification for the revision of ISO/TS 16949 to align with the ISO 9001:2015 structure and requirements.
Director of consulting
Omnex Engineering and Management
First delegation leader of IATF
Ann Arbor, MI