In a competitive world environment, where there are many vendors selling products and services that appear to be the same, quality, both real and perceived, is often the critical factor determining which product wins in the marketplace. Products that have a reputation for higher quality command a premium, resulting in greater market share and profit margins for the manufacturer. Low quality products not only take a big margin hit at the time of sale, but also taint the manufacturer with a reputation that will hurt future sales, regardless of the quality of future products. Users have a short memory. A company’s quality reputation is only as good as the quality of its most recent product.
The measurement, control and gradual improvement of quality is the goal of all quality systems, no matter what the name. Some of the more common systems are known as SQC (Statistical Quality Control) Quality Engineering, Six-Sigma, TQM (Total Quality Management), TQC (Total Quality Control), TQA (Total Quality Assurance) and CWQC (Company- Wide Quality Control). These systems work on the principle that management must integrate quality into the basic structure of the company, and not relegate it to a Quality Control group within the company. Historically, most of the innovations in quality systems took place in the 20th century, with pioneering work carried out by Frederick W. Taylor (Principles of Scientific Management), Henry Ford (Ford Motor), W. A. Shewhart (Bell Labs), W. E. Deming (Department of Agriculture, War department, Census Bureau), Dr. Joseph M. Juran (Bell Labs), and Dr. Armand V. Feigenbaum among others. Most quality control engineers are familiar with the story of how the statistical quality control pioneer, W. E. Deming, frustrated that US manufactures only gave lip service to quality, took the quality system he developed to Japan, where it was embraced with almost religious zeal. Japanese industry considers Deming a national hero, where his quality system played a major role in the postwar expansion of the Japanese economy. Twenty to thirty years after Japan embraced his methods, Deming found a new audience for his ideas at US companies that wanted to learn Japanese methods of quality control.
All quality systems use Statistical Process Control (SPC) to one degree or another. SPC is a family of statistical techniques used to track and adjust the manufacturing process in order to produce gradual improvements in quality. While it is based on sophisticated mathematical analysis involving sampling theory, probability distributions, and statistical inferences, SPC results can usually be summarized using simple charts that even management can understand. SPC charts can show how product quality varies with respect to critical factors that include things like batch number, time of day, work shift personal, production machine, and input materials. These charts have odd names like XBar-R, Median-Range, Individual-Range, Fraction Number Non-Conforming, and NP. The charts plot some critical process variable that is a measurement of product quality and compares it to predetermined limits that signify whether or not the process is working properly.
Initially, quality control engineers created all SPC charts by hand. Data points were painstakingly gathered, massaged, summed, averaged and plotted by hand on graph paper. It is still done this way in many cases. Often times it is done by the same factory floor personal who control the process being measured, allowing them to “close the loop” as quickly as possible, correcting potential problems in the process before it goes out of control. Just as important, SPC charts tell the operator when to leave the process alone. Trying to micro-adjust a process, when the process is just exhibiting normal random fluctuations in quality, will often drive the process out of control faster than leaving it alone.
The modern tendency is to automate as much of the SPC chart creation process as possible. Electronic measuring devices can often measure quality in real-time, as items are coming off the line. Usually some form of sampling will be used, where one of every N items is measured. The sampled values form the raw the data used in the SPC chart making process. The values can be entered by hand into a SPC chart making program, or they can be entered directly from a file or database connection, removing the potential for transcription errors. The program displays the sampled data in a SPC chart and/table where the operator or quality engineer can make a judgment about whether or not the process is operating in or out of control.
Usually the SPC engineer tasked with automating an existing SPC charting application has to make a decision about the amount of programming he wants to do. Does he purchase an application package that implements standard SPC charts and then go about defining the charts using some sort of menu driven interface or wizard. This is probably the most expensive in terms of up front costs, and the least flexible, but the cheapest in development costs since a programmer does not have to get involved creating the displays. Another choice is to use a general purpose spreadsheet package with charting capability to record, calculate, and display the charts. This is probably a good choice if your charting needs are simple, and you are prepared to write complicated formulas as spreadsheet entries, and your data input is not automated. Another choice is writing the software from scratch, using a charting toolkit software as the base, and creating custom SPC charts using the primitives in the toolkit. This is cheaper up front, but may be expensive in terms of development costs. Often times the third option is the only one available because the end-user has some unique requirement that the pre-packaged software can’t handle, hence everything needs to programmed from scratch.
Another approach is to use a web-based application, which allows you to plot and customize SPC charts, while remembering your setup conditions and data. That is what this web site is all about. It can be considered a tutorial about the most common of the standard SPC charts, and contain working examples of each chart that you can upload your data to and run.
If you are interested in a more complete history of SPC Charting, you are referred to some of the milestone books on the subject.
Shewhart, Walter A[ndrew]. (1931). Economic control of quality of manufactured product. New York: D. Van Nostrand Company. p. ISBN 0-87389-076-0. LCCN 31032090. OCLC 1045408. LCC TS155 .S47. ASQ. Online version found here: https://pqm-online.com/assets/files/lib/books/shewhart1.pdf
Donald J. Wheeler, David Smith Chambers. (1992). Understanding Statistical Process Control. SPC Press, 1992. ISBN 0945320132, 9780945320135.
Montgomery, Douglas C. (2012), Introduction to Statistical Quality Control (7 ed.). Hoboken, New Jersey: John Wiley & Sons, ISBN-10: 9781118146811, ISBN-13: 978-1118146811, ASIN: 1118146816. An expensive book, but you can find used, earlier versions for a small fraction of the current edition.
Western Electric Company (1956), Statistical Quality Control Handbook. (1 ed.), Indianapolis, Indiana: Western Electric Co., p. v, OCLC 33858387. Online version found here: https://www.westernelectric.com/support/statistical-quality-control-handbook.html
- Shewhart X-bar and R and S Control Charts. NIST/Sematech Engineering Statistics Handbook]. National Institute of Standards and Technology. Retrieved 2009-01-13.