Does a CC on the Drawing Mean Cpk Must Reach 1.67? —— Grading Requirements for Special Characteristics and the Five-Step Method for Alternative Control

1. The Question at the Audit Table: "What is the Cpk for This CC?"

A process audit was conducted at an automotive parts company. The auditor flipped to a drawing where a hole dimension was marked as CC and asked the on-site quality engineer: What is the process capability for this characteristic now? The engineer replied confidently: We inspect this dimension 100%, and no nonconforming product has ever slipped through. The auditor followed up: I see the 100% inspection, but where are the capability data? The engineer produced a stack of inspection records, all marked "合格" (conforming), with no control charts or capability indices for the entire year.

The audit conclusion listed three nonconformities:

1. No process capability assessment established for special characteristics;
2. No defined alternative control measures and verification of their effectiveness when capability is insufficient;
3. Inconsistency between the characteristic grades in the control plan and those annotated on the drawing.

This scenario highlights a common cognitive gap: many companies interpret "special characteristics" as "stricter inspection," but they overlook the true meaning—special characteristics are a set of characteristics that must be proven to be "capable of being produced by the process," not a set of "more frequent inspections." The presence of a CC or SC symbol on the drawing implies a comprehensive set of requirements, including capability targets, monitoring methods, response times for anomalies, and document consistency. Simplifying it to "100% inspection" is like answering a math problem with a physical effort.

2. What Exactly Are Special Characteristics Required to Do?

First, let's clarify the grading. Characteristics are typically divided into three categories based on the consequences of failure:

• Safety and Regulatory Characteristics (often marked as CC or with a customer-defined symbol): Failure can lead to personal injury, regulatory violations, or recalls.
• Key Characteristics (SC): Failure can significantly impact customer functionality, assembly, usage experience, or subsequent processes.
• Important and General Characteristics: Impact secondary functions, appearance, or internal processes, and usually do not have independent capability thresholds.

The value of grading is not in determining which is more expensive, but in differential design for three key aspects: differentiated capability targets, monitoring intensity, and response speed for anomalies. If any of these three aspects are missing, the grading is just a symbol.

Why use Cpk instead of the pass rate? Because the pass rate answers whether the current batch is within tolerance, while Cpk answers how much margin the process has to absorb daily variations. Even if a process produces a fully conforming batch, if the mean is close to the upper tolerance limit and the process variation is large, the next batch may have a significant number of nonconforming products. Cpk calculates the multiple of the tolerance margin relative to the process variation—this is why customers insist on seeing capability data: they want evidence that the process will not fail in the future, not just a record that it did not fail in the past.

It's important to remember the division of labor between the two indices. Cpk is estimated using within-subgroup variation and reflects short-term capability; Ppk is estimated using overall variation and reflects long-term actual performance. During the initial submission phase, Ppk is often reviewed, while during mass production, Cpk is closely monitored. A significant difference between the two (e.g., Cpk 1.5 and Ppk 1.0) usually indicates large variation between subgroups—shift changes, mold changes, material changes, or supplier changes have introduced special causes. The focus should be on investigating between-subgroup variations rather than adjusting parameters.

Common industry capability targets (used as internal benchmarks when there are no customer-specific requirements):

The target values are not arbitrarily set. There is a clear priority order for their sources: customer drawings and CSR (Customer Special Requirements) > industry standards or regulations > internal risk assessments. Once determined, these targets must be reflected in three places: the product/process characteristics section of the control plan, the criteria in the inspection work instructions, and the supplier drawings and agreements. Inconsistencies in these three areas are the source of the "inconsistent documents" nonconformity.

3. Five-Step Method: From List to Capability Closure

Step 1: Create a Controlled List of Special Characteristics. Compile a list of special characteristics from the drawing annotations, DFMEA/PFMEA failure consequences, customer requirements, and regulatory requirements. Standardize the symbol system, grading, and numbering, and assign a single responsible person to maintain it. This list serves as the "single authoritative source," and all drawings, control plans, inspection documents, and supplier drawings should align with it. If different customers use different symbols (e.g., CC is called CTQ by Customer A and KC by Customer B), include a symbol cross-reference table in the list to avoid changing the terminology for each customer.

Step 2: Set Capability Targets for Each Characteristic. Clearly specify the target values and criteria for judgment (referencing customer CSR clauses or internal risk assessment records). For safety-related characteristics, also document the "transition strategy for capability insufficiency" to avoid ad-hoc solutions when issues arise.

Step 3: Verify Data Reliability Before Calculating Capability. The sequence is crucial. First, validate the measurement system—repeatability and reproducibility of the measuring instruments used for capability studies should generally be ≤10%, and at most ≤30%, otherwise the calculated index may reflect the capability of the measuring instrument rather than the process. Next, confirm process control—use control charts to eliminate obvious special cause points; calculating Cpk for an uncontrolled process is meaningless. Finally, ensure sufficient data volume—typically, 25 to 30 subgroups or more are needed, covering representative shifts, machines, and batches, rather than repeatedly stacking data from the same hour. Sampling should also cover sources of variation: subgroups from continuous production only reflect within-subgroup variation, and to assess long-term performance, subgroups from different shifts, mold cavities, and batches must be included. A capability study that only collects 30 subgroups from a single shift and machine can only represent that specific shift—this is often where customers require a redo.

Step 4: Define Alternative Control Measures When Capability is Insufficient. This is the step most often skipped and the one auditors love to check. The correct approach is to select one or a combination of measures from the strongest to the weakest, and verify the effectiveness of each layer:

1. Poka-yoke Devices: Intercept at the source, making it physically impossible to install incorrectly, miss, or misalign.
2. 100% Inspection: Use go/no-go gauges for attribute data and automatic measurement for variable data, while verifying the inspection error rate and the R&R of the gauges.
3. Enhanced Monitoring: Shorten sampling intervals and increase rules for identifying anomalies, minimizing the time to detect issues.
4. Corrective Action Plan: Conduct in parallel, find parameter windows (DOE), improve tooling, and initiate design change requests. Each plan must have a responsible person, completion time, and verification points.

Remember: Alternative control measures are transitional, not final solutions. When required by the customer, a deviation permit (deviation request) must be submitted, specifying the scope, duration, and exit conditions. Treating "100% inspection" as a permanent solution essentially shifts process quality issues to inspection costs.

Step 5: Document and Regularly Review. The conclusions of the capability study should be documented in the PPAP submission, control plan, and work instructions. Recalculate process capability after mold changes, material changes, shift changes, equipment maintenance, or process parameter adjustments. Review target values annually or whenever customer CSR updates. This step is often treated as a "wrap-up task," but it is crucial for maintaining capability management—without it, capability data can be lost when personnel changes.

4. A Real-World Example of Improving Capability from 0.90 to 1.71

A die-casting company supplying a shell for a vehicle manufacturer had a position tolerance marked as CC on the drawing, with a tolerance of φ0.5, and the customer CSR explicitly required Cpk ≥ 1.67. The company's monthly reports showed that the Cpk for this characteristic consistently hovered around 0.90, and the handling method was a simple statement: "This dimension is 100% inspected using gauges, with no risk of nonconforming product flowing out."

The first capability study identified two fundamental issues. First, the measurement system: position tolerance was measured using a coordinate measuring machine (CMM), and the clamping method relied on manual positioning by the operator, leading to a high R&R of 22%. This meant that a significant portion of the "capability index" was due to clamping differences. After modifying the clamping fixture, the R&R dropped to 8%. Second, the process was uncontrolled: when the data was stratified by mold cavity, it was found that the offset in cavity 3 was significantly larger than the others. When mixed together, the control chart only showed "slightly larger variation," but stratification immediately revealed the issue. It turned out that the cooling water line for cavity 3 of the die-casting machine was long-term blocked, a maintenance issue rather than a process parameter issue.

After clearly identifying the problems, the company took three parallel actions. Short-term alternative control measures were implemented: the position tolerance was changed to 100% inspection using a dedicated gauge and an automatic sorting poka-yoke was installed to prevent nonconforming parts from entering the next process. Each shift verified the effectiveness of the gauge and poka-yoke using a set of standard parts, and the inspection error rate was tracked monthly. Mid-term corrective actions were also initiated: all mold cavity cooling water lines were cleaned and included in the inspection checklist, a design of experiments (DOE) was conducted to find a more stable parameter window for mold temperature and holding pressure, and the mold cavity was fixed as a stratification variable in the monitoring.

Three months later, the Ppk increased to 1.42, indicating the right direction but still not reaching 1.67. Instead of using "100% inspection" to gloss over the report, the company submitted a deviation permit application to the customer, including the corrective data, control charts, gauge verification records, and subsequent plans. After review, the customer agreed to continue supplying under intensified inspection conditions, with a six-month observation period. Another six months later, the Cpk stabilized at 1.71, the deviation permit was withdrawn, 100% inspection was canceled, and the annual inspection cost decreased. Customer complaints related to this model dropped from three per quarter to zero.

In this case, no high-tech methods were used. The true turning point was two things: first, proving the data's reliability before discussing capability; second, providing a time-limited, responsible plan for capability insufficiency, rather than sealing the problem with 100% inspection.

5. Six Common Misconceptions

1. One-Size-Fits-All Target Setting: Setting all dimensions to 1.67, spreading resources thinly over less critical appearance dimensions, or setting all to 1.33, thereby overlooking safety characteristics that require higher thresholds. Targets should align with the grading.
2. Using 100% Inspection to Replace Capability Proof: 100% inspection only proves that the current batch is conforming, not that the process is under control. Auditors look at capability studies, not the thickness of inspection records.
3. Calculating Cpk for an Uncontrolled Process: Calculating the index when there are still outliers or obvious trends on the control chart results in an inflated number that mixes special causes. It may be useful for reporting but is dangerous for decision-making.
4. Conducting Capability Studies Without Validating the Measurement System: When the gauge R&R exceeds 30%, the Cpk study is actually measuring the gauge's resolution. This is the most common reason for "good-looking but useless" capability data.
5. Forcing Cpk on Non-Normal Characteristics: Characteristics like position, flatness, roundness, and concentricity often have a skewed distribution. Directly applying the normal distribution assumption to calculate Cpk can mislead judgments. Use Ppk, transformation methods, or attribute data handling instead.
6. Closing Capability Insufficiency with 100% Inspection, Without Corrective Actions, Time Limits, or Reviews: The result is the same nonconformity reappearing with a different customer three years later.

6. In Summary

CC and SC are not labels for "stricter inspection," but commitments that "capability must be proven"—targets should be graded, data should speak, and when capability is insufficient, both alternative controls and corrective actions are necessary.

Special characteristics require proof of capability, not just more inspections.

Knowledge code: 8.2.2

Version: v20260915

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