Seeing Variation!

Variation in a product or process is sometimes difficult to detect.  Quality Window provides a number of ways to discover variation in an individual variable as well as across multiple variables or applications.  Let’s take a look at a few ways.

First, looking at the Logsheet and seeing a number of colors displayed for a single variable.  Remember the colors are set based on the limits fixed or calculated for a variable.

Another way to see variation is to review a variable on a Control Chart screen.  The chart portion of the screen shows the performance of a variable against the backdrop of the limits for the variable.   The histogram in the lower left of the screen as the Capability Bar (+-3 Std. Dev. around Avg.) located on the left side of the histogram also indicates the volume of the variation against those same limits.

        

The statistics box in the lower right of the screen also has indicators for variation if they are selected (Statistics are selected from the Option menu in Quality Window 5).  Statistics like Standard Deviation and the statistic designed to highlight variation Cr – capability ratio.  If this statistic is highlighted it indicates the variation for this variable has exceeded .75 (a warning limit for variation).

Turn on the Range chart option (click Range On button on toolbar) on the Control Chart screen is another way to see variation.  A spike in the chart indicates a significant change form one value to another.

In the Process Monitor Example  v16-Weight Profile All Lanes is a special variable called an @Pop calculated variable.  This type of variable collects data for 3 weight fields and displays it in a special Control Chart format.  The vertical lines on the screen indicate the amount of variation detected in the 3 weight samples taken.  The blue dot is the average of these variables.

The Quick Report, click Report tab on bottom of the screen, displays a report of summary statistics for all numeric variables in the application.  Any of the statistics that are hilited indicate a problem for that variable and statistic.  The Cr or Variation statistic is important here and the added benefit is you get to see the performance of all variables at the same time.  If you are investigating variation then seeing what other variables are experiencing the same issue can be quite informative.

A very powerful tool for “minimizing variation and running to optimum target” is the Prioritize Hatchart screen.  This screen is unique to Quality Window and is a very useful and intuitive screen.  Once explained, understanding is instant and allows you to quickly focus on what are the issues for which variables.  Seeing groupings of variables with similar problems is also very useful in solving problems.

Three statistics are automatically selected – Cr – capability ratio, Tz – Target z and Cpk – Clearance – without these selection the Hatchart does not display.  The three statistics are displayed at the top of the screen in three Pareto charts – ranking highest to lowest.  The Paretos are always sorted in the left most selected statistic.

To access the chart, press the Prioritize tab at the bottom of the screen and then go to the Options tab at the top of the screen and select Prioritize and then Hatchart.  Only numeric variables with fixed specification limits will be displayed in this chart.

We will focus our attention on the Hatchart in the lower left corner of the screen.

Simply, the left axis is Variation – Capability Ratio and a good Cr is .75 or less. The bottom axis is Target Z and any value of -.5 to +.5 is good.  Therefore everything in the green box is fine.  Any Variable that is outside the box has a Variation or Targeting problem or both.  It is very easy at a single glance to know where the problems are and thus what you need to work on.

 

 

Unlock your data with QW6 SPC Software

Smart teams know the key to improved results is unlocking data that will show you where to look. Give QW6 a try for free and gain the visuals and statistics that can lead your path to success.

Translate »