Carillon Technologies Limited Analysis of a Completed Histogram

---

Ideal Process:

The simple histogram will show us the variation in the process: This is the ideal situation; the process spread is within the specification limits and is centered. We have plenty of "wiggle" room. We would call this process "highly capable"

Effect of a shift in the mean or average for our "ideal process":

The process has experienced a large shift in it's process average and is now causing many of the parts or services to exceed the requirements. In this case the process average needs to be adjusted lower in order to meet the customer requirements.

A centered "marginal process" with too much variation:

A centered process with a spread approximately the same as the limits, but with no "wiggle" room. We would call this process marginally capable. Emphasis on reducing the inherent variability of this process is needed in order to insure compliance to the customer requirements.

The effects of a small shift in the process average for the "marginal process":

A process with a spread approximately the same as the limits, but with the process off center. With no "wiggle" room with only small shifts in the process average a significant amount of nonconforming materials can be produced.

A centered incapable process that produces nonconforming products or services:

A process with a spread greater than the limits, with nonconformance on both sides of the specifications. This process is "not capable" of meeting customer requirements. In order to meet customer requirements immediate efforts must be made "to sort" the good from the bad and to institute an improvement process aimed at reducing the process variation.

Effects of "assignable causes" on the output of the process:

This example indicates that some assignable cause is acting on the process to produce a non-symmetric distribution. Significant shifts in the mean during the sampling can be indicated by the following two drawings:
A double distribution, suggesting that two different processes are at work or a shift in the mean occured in a single process. Called a Bimodal Distribution. Possible causes could be: change in raw materials, tooling changes, operator readjusted the process etc.
A well-centered main distribution with a second small distribution that is causing major problems. This could be caused by the causes listed above or could indicate two different cavities or the mixing of output from two sets of tooling etc.

Limitations:

As you can see the histogram can convey a lot of information about the process, but it does have limitations:

• It requires many measurements.
• It does not take time into consideration.
• It does not show trends.

Next Steps:

If the processes do not meet the "ideal process" model corrective action will be necessary. Typical next steps could included the following:

• Collecting additonal data sequentially to produce a line graph (either run chart or preferable a control chart) to determine if the problem is time focused.
• The data could also be used to identify if the problem is positional in nature (cavitiy, tooling, measurement points)
• Analysis whether or not the measurement system is capable.
• Follow the System Alignment model in order to pinpoint the problem.

Conclusions

The histogram is an easy but effective analysis tools to pinpoint variation in the process.

Return to Histogram Page