 |
Carillon Technologies Limited |
Control Chart Discussion and Intrepretation |
|
|
Purpose:The control chart is a method used to understand process averages and variablity.
Control charts accomplish this by showing control limits for data calculated using measurements of samples.
If these limits are exceeded, it indicates that an assignable cause, or special cause, of variation is present. Note that control charts do not indicate what caused this variation, only that this variation is the result of a special cause.
When the sample measurements plot within the control limits, the variation is attributed to small-chance causes, or common causes, and the process is considered to be in statistical control.
Variation due to small-chance cause is usually due to the manner in which the process functions, and requires a long term process improvement effort. In contrast, large assignable causes of variation usually require more immediate correction.
Control charts can be created for any process characteristic that requires control. The most common control charts are for averages (x bar), individuals (x), medians (m), ranges (r), % defective (p), and number of defects per unit (u).
Discussion:Variation, which is caused by methods, environment, materials, machines, and men and women, is inherent in all processes.
This variation must be minimized to assure product consistency.
The goal of management and staff is to understand what part of the process variability is due to small chance causes of variation and what part is due to large assignable causes. In turn, these causes must be analyzed and minimized or, if possible, eliminated.
An effective tool for understanding the sources of process variability is a control chart.
A key initial purpose of a control chart is to detect the presence of assignable causes of variation.
This detection is achieved by taking samples from the process and comparing the measured characteristics of these samples (averages, range, % defective, number of defective) with control limits.
Sample characteristics are plotted against control limits to provide a graphic representation of the process and to highlight trends in the process.
When sample characteristics exceed these control limits, it indicates that the process has gone out of statistical control and that management and staff must take corrective action.
Control chart components include:Upper and lower control limits are calculated in a manner that prevents unnecessary operator or manager adjustment or correction of the process. These limits fall at 3 standard deviations above and below the average value of the measured sample characteristic. A center line that indicates where the process is aimed or should be aimed. |
 |
Because upper and lower control limits encompass 99.7% of all data it is highly unlikely that the control limits will be exceeded by chance alone. Some special, assignable cause would appear to cause process control limits to be exceded.
Therefore, when the control limits are exceeded, managers and operators know with certainty that they must take corrective action on the process's operation.
In a similar manner, when data falls within control limits, managers and operators know with certainity that they should not take any corrective action on the process's operation.
How to use control charts:The main purpose of control charts is to understand process variation in order to take steps to minimize this variation.
As noted previously, two types of variation exist:
1. Common Cause, or inherent, random variation, which is:
Difficult to control or change without making major changes to the process itself.
Should be randomly distributed around the center line of the chart and between the control limits.
2. Special Cause, or assignable variation, which is:
Caused by some change in the process that allows process output to be produced in a non-random, detectable manner.
Correctable in the short-term after it is discovered and studied.
A control chart indicates special cause variation in three ways:
1. When the data point falls outside the upper or lower control limits.
2. When the data points show a shift.
3. When the data points show a trend.
4. When the data points show a cycle
Interpretation of Control Charts:Common Causes The remaining 85% of quality problems are created by the process and can be corrected only by management. If the process is in control, but the results are not acceptable, management must take action to change the process. |
 |
Special Causes A point that falls outside an upper or lower control limit on a chart indicates a special cause. One can assign a reason for the occurance of this point. The special, or assignable, cause for this point and the corrective action, if known, should be noted on the chart. Special causes account for about 6% of all quality problems. |
 |
| Point out of Control Limits | |
| A shift occurs if you have 7 points in a row above the average or center line or seven points in a row below the average or center line |  |
| A trend occurs if you have seven points in a row going up or sevenpoints in a row going down |  |
| A cycle occurs if a repeative pattern is observed (i.e tool wear, temperature variation, pressure etc.) |  |
For out of control conditions
apply the system alignment process:
followed by the system improvement
process:
