Shainin Methods

Share Button

Previously we discussed Taguchi Methods as an alternative to classical Designed Experiments. Another approach to experimentation is Shainin Methods, as developed by Mr. Dorian Shainin and strongly advocated by Mr. Keki Bhote.

The late Dorian Shainin began introducing a series of experimentation techniques in the 1930’s and continued his work for nearly 50 years. Keki Bhote enhanced and honed these techniques at Motorola, where he helped launch the Six Sigma process. In his book World Class Quality – Second Edition, Mr. Bhote listed the following advantages of Shainin Methods:

  1. Root causes are determined by “talking to the parts” rather than by hunches or opinions.
  2. The techniques are statistically powerful, resulting in the separation of and quantification of main effects and interaction effects.
  3. There are 12 different techniques that provide a wide range of versatility.
  4. The techniques are easy to learn and low cost to use, making it easier to involve the entire workforce.

The 12 techniques may be broken down into the following four categories:

Clue Generation

  1. Multi-vari Analysis: A filtering technique used to eliminate non-contributing causes of variation until the root cause can be distilled.
  1. Concentration Chart or Pictograph: Used to plot the location of a defect on a part or on a layout or grid. The result will either be a random pattern, or a concentration in a particular area(s).
  1. Components Search: Parts and sub-assemblies are swapped between a good product and a bad product until the root cause of the problem is located.
  1. Paired Comparisons: Compare best and worst examples of products, and isolate the characteristics or parameters that are different between the best and the worst.
  1. Product/Process Search: List and measure the process variables that impact the quality of a product. Determine which of these process variables is causing a defect by comparing the measurements from a process that produces good parts with measurements from a process that produces bad parts.

DOE Optimization

  1. Scatterplot (or scatter diagram): Used to visually describe the relationship between two variables by plotting one against the other.
  1. Response Surface Methodology (RSM): A DOE technique typically used after the critical factors in a process have been isolated and we wish to optimize the settings of these factors. RSM designs are used when we know or suspect that there is curvature (i.e. non-linearity) in the response variable.

Formal DOE Techniques

  1. Variables Search: A binary search technique that isolates important from unimportant process variables by experimenting with best and marginal settings for each variable.
  1. Full Factorial Experiments: These experiments address all combinations of factors including all of their interactions, and are best employed for just a few factors that significantly impact the response variable. They are more time consuming and costly to conduct than screening techniques.
  1. B vs. C: B represents the better or improved process while C represents the current process. We run a test using six samples, three B’s and three C’s. If all three B’s outrank all three C’s, there is only one chance in 20 (based on the Law of Combinations) that this is due to chance, giving us 95% confidence that the improved process is better than the current process.

Transition from DOE to SPC

  1. Positrol (or precontrol): Units are classified as red, yellow or green based on where they fall in relation to the specification or tolerance. Green is the middle half of the tolerance, yellow is the remaining half of the tolerance and red is beyond tolerance. Sampling frequency is determined based on how often the process needs adjustment. If a sample is green, continue running. If a sample is yellow, select a second sample. If the second sample is yellow, stop and adjust or correct the process. If either sample is red, stop and adjust or correct the process.
  1. Process Certification (Process Control, and Management Plan): Defines the who, how, where and when of controls that will ensure that the important variables or factors are kept under tight control.

Note: Response Surface Methodology and Full Factorial Experiments are classical DOE techniques. What Mr. Shainin refers to as Process Certification is what we know in Six Sigma as the Process Control and Management Plan. The scatterplot or XY plot is typically included in the Six Sigma body of knowledge.

Your comments or questions about this article are welcome, as are suggestions for future articles. Feel free to contact me by email at roger@keyperformance.com.

About the author: Mr. Roger C. Ellis is an industrial engineer by training and profession. He is a Six Sigma Master Black Belt with over 45 years of business experience in a wide range of fields. Mr. Ellis develops and instructs Six Sigma professional certification courses for Key Performance LLC.   For a more detailed biography, please refer to www.keyperformance.com.

 

 

Share Button

Leave a Reply