Insufficient Data for 25 Subgroups, No Process Control? — Five Steps to Implement Pre-Control Charts

In a hydraulic valve body workshop, the bore diameter is a customer-specified special characteristic with a tolerance of only 0.021 mm. This process involves five different specifications, with a monthly production of 800 pieces, averaging only two to three batches per specification. The quality engineer, following textbook requirements, attempted to create an average-range chart but found after three months that: it was impossible to gather 25 subgroups for each specification, and the control limits were either uncalculable or became obsolete immediately upon a change in specification. As a result, he took a step back and simply drew a few lines based on the drawing tolerances and posted them next to the machine—operators would inspect five pieces at the start of the process, and if all five pieces fell within the middle zone, the process would be released. Over the next three months, the bore diameter of this process had zero nonconformities, and the inspection frequency decreased by about one-third.

This method is not called a control chart but a pre-control chart (Pre-Control).

1. Pre-Control Charts Focus on "Maintaining Conformance," Not "Statistical Variation"

The logic of a conventional control chart is: first, estimate the common cause variation from process data, then calculate the control limits, and use these limits to determine if the process is in control. Pre-control charts, from the outset, do not aim to estimate variation—they directly use the drawing tolerances, dividing the tolerance band into four equal parts:

• The middle two quartiles (totaling 50% of the tolerance band) are called the green zone;
• The two outer quartiles (each 25% of the tolerance band) are called the yellow zone;
• The area beyond the tolerance limits is called the red zone.

The judgment rules are straightforward: if five consecutive pieces fall within the green zone, the process is released and can be relaxed to skip inspection; if two consecutive pieces fall within the same yellow zone, it indicates a process shift, and the machine should be stopped for adjustment; if one piece falls within the red zone, the machine should be stopped immediately, and all products from that time period should be isolated and subjected to more stringent inspection.

Since it involves "counting zones" rather than "calculating statistics," operators do not need to understand standard deviation, and there is no need to accumulate data beforehand. The trade-off is a reduction in information: it is sensitive only to shifts in the process position, not to changes in the size distribution, and it does not provide quantitative conclusions such as Cpk. Therefore, pre-control charts and control charts are not substitutes but complementary tools—control charts answer whether the process is stable and capable, while pre-control charts determine whether the current batch can be released.

There are clear prerequisites for using pre-control charts, all of which must be met: the process capability is already good (Pp at least 1.0, and 1.33 or higher is more stable); the measuring instrument's resolution is sufficient (recommended to be no more than 1/10 of the tolerance band); and the variation is primarily due to position drift (such as tool wear, thermal deformation, and compensation drift) rather than distributional variation caused by multiple cavities or workstations.

2. Five Steps to Implement Pre-Control Charts

Step One: Conduct an Applicability Assessment. Collect data from 30 to 50 recent consecutive pieces to calculate the capability once. Processes with a Pp below 1.0 should not use pre-control but should undergo improvement. Verify that the measuring instrument's resolution meets the T/10 requirement (this is often overlooked in machining processes with tight tolerances and small machining allowances). Finally, assess the nature of the variation; processes with position drift are the most suitable.

Step Two: Divide the Tolerance Band into Four Equal Parts and Mark It at the Site. Calculate the three boundary values and write them down along with the upper and lower tolerance limits as specific numerical values. Create color cards next to the machine, color markers on the measuring tools, or display the zones directly in the inspection system. Colors are just aids; the numerical values are the criteria—this avoids subjective judgments like "it looks about right in the middle."

Step Three: Define Sampling and Judgment Rules. At the start (after a change in specification, tool, or cold machine startup), inspect five consecutive pieces. During normal production, conduct fixed-frequency sampling inspections, with each group consisting of five pieces (e.g., every 20 pieces, every two hours, or twice per shift). If all five pieces are in the green zone, proceed to the next round and reduce the inspection frequency; if two pieces are in the same yellow zone, stop the machine for adjustment; if one piece is in the red zone, stop the machine and trace and isolate all products since the last full green inspection. These rules should be documented in the reaction plan section of the control plan, not just memorized by the team leader.

Step Four: Define Response Authorities and Escalation Paths. Adjustments for the yellow zone should be made by operators according to the compensation amounts specified in the work instructions, followed by a new start of five consecutive pieces. Red zone events must be reported to quality and engineering, and all products since the last full green inspection should be traced. The authority to stop the machine should not be reserved for supervisors only—night shifts often hesitate to stop the machine, which is the most common reason for pre-control chart failure at the site.

Step Five: Regularly Review and Timely Switch. Pre-control charts are transitional tools. Once enough data is accumulated for the same specification (e.g., 20 to 25 consecutive groups), switch to an average-range chart or a single-value-moving-range chart to truly quantify process variation. Additionally, review the adjustment records quarterly: if most "yellow zone adjustments" do not result in significant changes in the values, it suggests that the yellow zone judgment is too strict for the process. At this point, the sampling frequency and capability assessment should be adjusted, not the criteria.

3. An Example

For a precision boring hole φ20H7 (+0.021/0), the tolerance band is 0.021 mm, and the quartiles are 0.00525 mm:

• Green Zone: 20.00525 to 20.01575;
• Lower Yellow Zone: 20.000 to 20.00525, Upper Yellow Zone: 20.01575 to 20.021.

A digital internal micrometer with a resolution of 0.001 mm is used, meeting the requirement of being no more than T/10 (0.0021 mm). After a change in specification, the first five pieces are inspected consecutively, and if all are in the green zone, the process is released. During a certain shift, two consecutive pieces fell into the upper yellow zone (20.016), and the operator, following the work instructions, adjusted the tool compensation downward by 0.003 mm. The subsequent five pieces were all in the green zone, and the products from that time period were confirmed to have no nonconformities after a more stringent inspection. For that month, the bore diameter had zero nonconformities, and the inspection frequency was adjusted from one piece every 20 pieces to skip inspection, reducing inspection labor by about one-third.

4. Three Common Pitfalls

Pitfall One: Implementing Pre-Control on Processes with Insufficient Capability. Processes with a Pp of only 0.8 will frequently enter the yellow and red zones, turning "occasional line stops" into "daily line stops." Eventually, everyone will loosen the criteria, rendering the tool ineffective. A capability assessment must be conducted before implementing pre-control.

Pitfall Two: Treating Pre-Control Charts as Substitutes for Control Charts. Pre-control charts do not estimate variation, calculate Cpk, or detect changes in the size distribution. Using them in situations where capability prediction is needed can miss processes that are truly deteriorating. They are complementary tools, not a binary choice.

Pitfall Three: Having the Chart Without a Response. Posting color cards without documenting the reaction plan and stop authority means that yellow zone alarms will go unaddressed. The value of pre-control charts lies not in the lines next to the machine but in the "inevitable action" that follows an alarm.

5. One-Sentence Summary

Pre-control charts are not a simplified version of control charts but a set of release rules tailored for small batches and transitional periods—using the quartiles of the tolerance band to make immediate judgments at the site, ensuring a response to maintain the baseline, and switching to statistical control charts once sufficient data is accumulated.

Insufficient data is not a reason to abandon monitoring. Use the quartiles of the tolerance band to manage the site initially, and switch to control charts once enough data points are available.

Knowledge code: 6.3.1

Version: v20260926

Author: QTank QTank is dedicated to providing systematic professional knowledge, methodologies, and practical tools for quality management practitioners, helping companies continuously improve their quality capabilities.