It was Walter Shewhart who developed the theory of Statistical Process Control in 1924 while at Bell Labs. His theory and lectures were published in a book called Economic Control of Quality Manufactured Product. He was the first to apply statistical techniques to manufacturing processes.
Post WWII W. Edwards Deming took Shewhart’s SPC techniques to Japan and trained 100’s of Japanese engineers. Deming promised that Japanese quality would be known world-wide within in five years. It only took four!
Statistical Process Control, commonly referred to as SPC, is a form of Statistical Quality Control. It uses graphical displays known as control charts to determine when a process should be adjusted.
SPC is Prevention! Acceptance Sampling is Detection!
A control chart is a form of time series plot. It allows you to monitor process performance over time. More specifically, with a control chart you can detect:
In the time series chart below a shift in the process average can clearly be seen beginning at index value 20.. But, some meaningful changes are not so easily seen by eye. SPC will detect these small meaningful process changes.
Change in process average beginning at Index value 20.
Changes in process variation can also be seen in graphical charts.
The time series plot below shows an increase in process variation at Index values 23 and 24. Like the process average above, SPC can detect much smaller meaningful changes that are not clearly evident by eye.
Increase in process variation at Index values 23 and 24.
Special-cause Variation:
Significant “one-of”, type events. The plot below shows a sharp spike in performance. Some “one-of” type event. And yes, once again, with SPC you can detect much smaller meaningful changes.
Special-cause variation (one-of type event)
Variations Impact
It can be said that the perfect process is one with no variation. A target is set and is hit every time. Unfortunately, no perfect process exists. yet. All processes exhibit variation!
At some point variation can hurt! The symptoms of its existence in business are repair, rework, late deliveries, loss of customer goodwill, high prices, the list goes on and on. For these reasons variation must be controlled.
Controlling process variation minimizes the time, cost and “pain” associated with not producing to requirements! Variation is the enemy of all processes.
Using simple statistics and simple graphical control charts, Shewhart developed statistical process control “weapons” to combat variation.
He showed that variation acted upon processes in two very distinct ways!
When only common-cause variation exists in a process the process is highly predictable.
Processes that are experiencing the second type of variation, special-cause variation, are not predictable. A part that breaks in a production machine or an untrained associate performing a task are examples of potential sources of special-cause variation.
Bill Smith, a Quality Engineer at Motorola in the 1980’s and the father of Six Sigma, gets my credit for redefining the application of SPC.
Smith’s Six Sigma drove home the point that the output is a function of the inputs. Statistical Process Control books were re-written because of Smith’s work.
Statistical Process Control is best applied to process inputs. You can control the inputs but only monitor the outputs.
Process Model - Sources of Variation
Statistical Process Control Basics |
Statistical Process Control Control Charts Xbar & s (Standard Deviation) Chart |
From Statistical Process Control to Statistical Quality Control.
May 10, 16 09:24 PM
A Quality Control Plan is a documented description of the activities needed to control a process or product. The objective of a QCP is to minimize variation.
May 10, 16 08:49 PM
The Largest Collection of Free Six Sigma Tools and Training on the Web!
May 10, 16 07:28 PM
The Weibull distribution is applicable to make population predictions around a wide variety of patterns of variation.