Recalculating Control Limits: Ensuring Accuracy and Relevance in Statistical Process Control

Statistical process control (SPC) plays a crucial role in maintaining and improving product quality by monitoring and analyzing process data to identify variations and potential issues. Control charts, a fundamental tool in SPC, utilize control limits to determine whether a process is operating within acceptable limits of variation. However, there are instances when recalculating control limits becomes necessary to ensure their accuracy and relevance. This article explores the circumstances that warrant recalculation of control limits, drawing insights from reputable sources such as Advantive, SPC for Excel, and Elsmar Cove.

Key Facts

Recalculating control limits: Control limits are typically calculated based on historical data to determine the acceptable range of variation in a process. However, there are instances when it is necessary to recalculate control limits.
Impact of process changes: It is recommended to extend the existing control limits over new data until there is evidence that a change in the process has had an impact on the data, such as shifting or out-of-control evidence.
Initial calculation of control limits: In the past, it was common practice to calculate control limits after a certain number of data points, such as 20 or 100. However, with the availability of computer software, control limits can be calculated after as few as 5 data points. Recalculating control limits after each point until reaching 20 can provide a better estimate of the average and sigma. Once the process is fairly stable, the control limits can be locked and used for future analysis.
Frequency of data collection: The frequency of data collection can also influence when control limits should be recalculated. For data collected monthly or less frequently, it may be appropriate to use a longer period of data, such as one year, to calculate control limits and lock them based on this baseline data.
Impact of out-of-control points: Control limits are designed to identify out-of-control points in a process. Leaving out-of-control points in the calculations may result in wider control limits. However, the control limits are still valid for the data, and the impact of out-of-control points can be minimized by using the right method to calculate control limits and having a sufficient amount of data.

Impact of Process Changes

Process changes, whether intentional or unintentional, can significantly affect the stability and performance of a manufacturing process. When a process change occurs, it is essential to extend the existing control limits over the new data until there is evidence that the change has impacted the data. This approach allows for the observation of any shifts or out-of-control evidence that may arise due to the process change. Only after such evidence is observed should the control limits be recalculated using data subgroups collected after the change was implemented.

Initial Calculation of Control Limits

Traditionally, control limits were calculated after a specific number of data points were collected, often 20 or 100. However, with the advent of computer software, control limits can be calculated with as few as 5 data points. Recalculating control limits after each data point until reaching 20 provides a more accurate estimate of the process average and standard deviation. Once the process exhibits stability, the control limits can be locked and utilized for future analysis.

Frequency of Data Collection

The frequency of data collection can influence the timing of control limit recalculation. For processes where data is collected monthly or less frequently, it may be appropriate to use a longer period of data, such as one year, to calculate control limits. These limits can then be locked based on this baseline data. This approach ensures that the control limits are representative of the long-term process behavior and are not unduly influenced by short-term variations.

Impact of Out-of-Control Points

Control limits are designed to identify out-of-control points in a process, indicating potential issues that require investigation. Leaving out-of-control points in the calculations may result in wider control limits. However, the control limits remain valid for the data, and the impact of out-of-control points can be minimized by employing the appropriate method for calculating control limits and having a sufficient amount of data.

Conclusion

Recalculating control limits is a crucial aspect of SPC, ensuring that control charts remain accurate and relevant in monitoring process performance. By considering factors such as process changes, initial calculation of control limits, frequency of data collection, and the impact of out-of-control points, manufacturers can effectively recalculate control limits to maintain process stability and improve product quality.

References

FAQs

When should I recalculate control limits?

Recalculate control limits when there is evidence of a process change, such as a shift in the process average or an increase in variation.
Recalculate control limits after a significant amount of new data has been collected, especially if the process is evolving or improving.
Recalculate control limits if there are a large number of out-of-control points, as this may indicate that the current control limits are not appropriate.

How often should I recalculate control limits?

The frequency of recalculating control limits depends on the stability of the process and the frequency of data collection.
For stable processes with frequent data collection, control limits may only need to be recalculated annually or semi-annually.
For unstable processes or processes with infrequent data collection, control limits may need to be recalculated more frequently, such as monthly or quarterly.

What are the consequences of not recalculating control limits?

Not recalculating control limits can lead to inaccurate process monitoring, as the control limits may not reflect the current process behavior.
This can result in an increased risk of false alarms (indicating an out-of-control condition when the process is actually in control) or missed alarms (failing to detect an out-of-control condition).
Both scenarios can lead to poor decision-making and a lack of effective process control.

How do I recalculate control limits?

The method for recalculating control limits depends on the type of control chart being used.
Generally, the new control limits are calculated using the same method as the original control limits, but with the new data.
It is important to use all available data when recalculating control limits, including both in-control and out-of-control data points.

What should I do if I have a lot of out-of-control points?

If there are a large number of out-of-control points, it is important to investigate the underlying cause of the variation.
Once the cause has been identified and corrected, the control limits can be recalculated using only the data from after the correction was implemented.
This will ensure that the new control limits are representative of the improved process.

Can I recalculate control limits manually?

While it is possible to recalculate control limits manually, it is generally not recommended.
Manual calculations are time-consuming and prone to error.
Statistical software packages can easily recalculate control limits with just a few clicks, saving time and reducing the risk of errors.

What are some common mistakes to avoid when recalculating control limits?

Avoid recalculating control limits too frequently, as this can lead to unstable control limits and an increased risk of false alarms.
Avoid recalculating control limits based on a small amount of data, as this can result in inaccurate limits.
Avoid recalculating control limits after making a process change without first investigating the impact of the change on the process.

Where can I learn more about recalculating control limits?

There are many resources available online and in libraries that can provide more information on recalculating control limits.
Some reputable sources include the American Society for Quality (ASQ), the International Organization for Standardization (ISO), and the Statistical Process Control Society (SPCS).