More Control Charts Module 6
Why? There are many probability distributions in our world
The Binomial Distribution, The distribution of coin tosses.
Two types of data Variables----Continuous Attributes—Discrete, Countable –Two types of attributes data You can count occurrences and non-occurrences. You can only count occurences. Examples?????
Some Variables Shewart Charts X-mR aka i-Chart, Individuals Chart X-bar-range X-bar-sigma
Some Attributes Shewart Charts p-Chart np-Chart u-Chart c-Chart
Decide on type of data Continuo us (Variables ) Data Discrete (Attributes) Data More than one observation per subgroup? < 10 observations per subgroup? Can both occurrences & non- occurrences be counted? Are there equal area of opportunity ? Are the subgroup sizes equal? Ye s No –R–sXmRc-chart u-chartp-chartnp-chart
Example Individuals Chart
Example X-Range Chart
Example X-Sigma Chart
How did they do that?
The basic pattern…. Plot observed measurements over time. –Measurements, counts, rates Plot Centerline –Average measurement or count, pooled rate. Plot Control Limits –Centerline +/- Multiplier X “Standard Deviation”
Multiplier does 3 Things Determines the number of sigmas – usually 3 Converts standard deviations to standard errors (variables data). Can include factor to adjust for unusually small or large number of subgroups or time intervals. Note: How multiplier is constructed and used varies by author.
“Standard Deviation” Based on sample estimate of population standard deviation. Based on moving ranges. Based on ranges.
The i-Chart or XmR Chart Calculate average of all individual values = x Calculate all the moving ranges (MRi) MRi = |x i -x i-1 | Calculate the average MR = Rbar Calculate control limits = xbar +/- 2.66Rbar Plot xbar Plot control limits Plot individual values, points
The Xbar-Range Chart
The Xbar-Sigma Chart
The Xbar-Sigma Chart (Part II)
Is “3” always OK? Notice 3 is multiplied by the SD. This gives +/- 3 Sigma Control Limits. Designed for 25 observations. When you have only 7 observations –β risk is too high When you have 200 observations –α risk is too high Can use T-Sigma Limits
T-Sigma Limits No. of Plotted PointsT
How to use T-Sigma Limits Substitute the T-Sigma limit from the table for the “3” in A3, B3, and B4 above. For attributes charts, simply substitute the T-Sigma Limits for the multiplier in front of the standard error.
The attributes Shewart Charts p-Chart np-Chart u-Chart c-Chart
Example p-Chart
Example np-Chart
Example u-Chart
Example c-Chart
How did they do that?
The p-Chart
The np-Chart Pooled over all subgroups
The c-Chart
The u-Chart
Choosing Charts 1.Continuous A. Only 1 observation per subgroup—use iChart B. More than 1 observation/subgroup i) Less than 10 observations/subgroup—use Xbar-R ii) 10 or more observations/subgroup--use Xbar-Sigma 2.Attributes A. Occurrences (heads) and non-occurrences (tails) can be counted. i) Subgroups of equal size—use np-Chart Ii) Subgroups of unequal size—use p-Chart B. Only occurrences can be counted. i) Equal area of opportunity (denominators)—use c-Chart ii) Unequal area of opportunity– use u-Chart
Decide on type of data Continuous (Variables) Data Discrete (Attributes) Data More than one observation per subgroup? < 10 observations per subgroup? Can both occurrences & non-occurrences be counted? Are there equal area of opportunity? Are the subgroup sizes equal? Yes No –R–sXmRc-chart u-chartp-chartnp-chart See Flow Chart on page 72 of Carey and Lloyd