Appearance
A Cpk of 1.0 means 0.27% of holes will be out of tolerance — 2,700 defective parts per million. At Cpk 1.67, that drops to 0.57 parts per million. In deep hole drilling, where a single out-of-tolerance bore can scrap a component worth thousands of dollars, the difference between Cp 1.0 and Cp 2.0 is the difference between gambling and knowing.
What Are Cp and Cpk?
Process capability indices compare the natural variation of a manufacturing process to its specified tolerance limits.
| Index | Formula | What It Measures |
|---|---|---|
| Cp | (USL − LSL) / 6σ | Potential capability — whether the process spread fits within the tolerance, assuming perfect centering |
| Cpk | min[(USL − μ) / 3σ, (μ − LSL) / 3σ] | Actual capability — whether the process is centred within the tolerance |
| Cpm | (USL − LSL) / 6√(σ² + (μ − T)²) | Capability adjusted for distance from target (Taguchi loss) |
| Ppk | Same formula as Cpk, using sample σ instead of estimated σ | Process performance — includes short-term and long-term variation |
Interpretation Guide
| Cpk Range | Process Quality | Defect Rate (ppm) | Action Required |
|---|---|---|---|
| < 1.00 | Not capable | > 2,700 | Immediate process improvement required |
| 1.00–1.33 | Marginally capable | 64–2,700 | Close monitoring, planned improvement |
| 1.33–1.67 | Capable | 0.57–64 | Standard SPC monitoring |
| 1.67–2.00 | Good | 0.00006–0.57 | Reduced inspection frequency |
| > 2.00 | Excellent | < 0.00006 | High-volume production capable |
Note: Many automotive and aerospace customers require Cpk ≥ 1.67 for critical bore dimensions. Internal deep hole drilling operations targeting Cp ≥ 2.0 account for tool wear drift that reduces Cpk during the production run.
Key Quality Characteristics in Deep Hole Drilling
| Characteristic | Typical Tolerance | Measurement Method | Typical Cp | Typical Cpk |
|---|---|---|---|---|
| Bore diameter (BTA, gun drilling) | IT7–IT10 (±0.01–0.05 mm) | Air gauge, plug gauge, CMM | 1.5–2.5 | 1.3–2.0 |
| Bore straightness | 0.1–0.5 mm/m | Laser alignment, straightness gauge | — | — |
| Surface finish (drilled) | Ra 0.4–6.3 μm | Profilometer | — | — |
| Hole position (circular TZ) | Ø0.1–0.5 mm | CMM | 1.5–2.5 (multivariate) | 1.3–2.0 |
| Wall thickness uniformity | ±0.1–0.5 mm | Ultrasonic | 1.3–2.0 | 1.0–1.7 |
Warning: Straightness and surface finish do not have Cp/Cpk indices in the conventional sense because they are unilateral characteristics (one-sided tolerance). For these, use Pp/Ppk calculated against the maximum allowable value, or analyse the distribution of measured values against specification limits directly.
Bore Diameter Capability
Bore diameter is the most commonly monitored quality characteristic in deep hole drilling, and the one with the most established capability data.
Published Results
2025 Study — Comparative Analysis of Drilling Process Capability
A production study of two drilling processes (Process B and Process C) for hole diameter in cast products:
| Index | Process B | Process C |
|---|---|---|
| Cp | > 2.0 | > 2.0 |
| Cpk | > 2.0 | > 2.0 |
| Method | Clements percentile (non-normal) | Clements percentile (non-normal) |
| Data transformation | Johnson transformation | Johnson transformation |
Both processes demonstrated negligible risk of out-of-tolerance parts. Process B showed lower short-term variability and nearly perfect centering. Process C showed better alignment with the nominal target dimension.
GB/T 40681.6-2021 (ISO 22514-6) — Hole Position Example
| Parameter | Value |
|---|---|
| Characteristic | Hole centre coordinates (x, y pairs) |
| Tolerance zone | Ø0.5 mm circular zone |
| Sample size | 100 holes |
| Ĉp (multivariate) | 2.43 |
| P̂pk (multivariate) | 1.48 |
| 95% CI for P̂pk | [1.19, 1.88] |
Typical Capability by Diameter Range
| Diameter Range | Method | Typical Capability | Comments |
|---|---|---|---|
| 1–6 mm | Gun drilling | Cp 1.5–2.0 | Small drills sensitive to runout |
| 6–20 mm | Gun drilling | Cp 1.7–2.5 | Most stable range for gun drills |
| 20–80 mm | BTA drilling | Cp 1.5–2.2 | Chip evacuation affects consistency |
| 80–200 mm | BTA drilling | Cp 1.3–2.0 | Guide pad wear is the limiting factor |
Data Characteristics and Statistical Methods
Skewness in Deep Hole Drilling Data
Deep hole drilling bore diameter data has a characteristic distribution that violates the normality assumption of standard Cp/Cpk:
| Characteristic | Description | Root Cause |
|---|---|---|
| Left truncation | Diameters cannot be smaller than the drill nominal | Drill always cuts at least to nominal size |
| Right skew | Occasional larger diameters are possible | Tool wear, runout, vibration increase bore size |
| Increasing skew with tool wear | Distribution shifts and spreads as tool wears | Progressive insert and guide pad wear |
Recommended Methods for Non-Normal Data
| Method | Application | Advantages |
|---|---|---|
| Johnson transformation | General non-normal data | Transforms to approximate normality; widely implemented in Minitab |
| Clements percentile method | Skewed data with known shape | Does not require transformation; uses Pearson curves |
| Cs index (Pearson, 1995) | Specifically for hole-drilling data | Designed for left-truncated, right-skewed distributions |
| Box-Cox transformation | Moderate non-normality | Simpler than Johnson; effective for mild skew |
Tip: For production deep hole drilling, use Johnson transformation with X̄-R control charts as the default approach. The Clements method is preferred when historical data confirms a stable non-normal distribution shape. The Cs index is most appropriate for academic research or where the specific skewness characteristic of drilling data is the focus of study.
Control Chart Implementation
Recommended Chart Types
| Quality Characteristic | Chart Type | Sampling Frequency | Rationale |
|---|---|---|---|
| Bore diameter | X̄-R chart | 5 parts every hour | Standard variables control |
| Bore diameter (small batch) | I-MR chart | Every part | No subgrouping possible |
| Tool wear trend | X̄ chart with tool offset | Per tool change | Monitor drift, trigger insert change |
| Chatter/spiralling detection | Residual control chart | Continuous sensor data | Detects dynamic instability |
| Hole position (multivariate) | T² Hotelling chart | Per setup + periodic | Multiple coordinates per hole |
| Surface finish | I-MR chart | Every 10th part | One-sided specification |
Residual Control Charts for Chatter Detection
Deep hole drilling is subject to dynamic disturbances (chatter vibration and spiralling) that cannot be detected by conventional X̄-R charts until scrap is produced. Residual control charts address this:
- Fit a time-series model (ARIMA) to the bore diameter signal
- Calculate residuals: actual − predicted values
- Plot residuals on a control chart
- A shift in residual mean or variance indicates the onset of chatter — often 10–20 parts before conventional charts would detect a problem
This approach, validated in Quality Engineering (2008), enables proactive intervention before spiralling produces an out-of-tolerance bore.
Process Improvement Strategies
When Cpk Is Below Target
| Cpk Range | Likely Root Causes | Improvement Actions |
|---|---|---|
| < 1.0 | Tool wear, coolant pressure variation, machine misalignment, guide bushing wear | Baseline machine alignment; stabilise coolant system; standardise tool change intervals |
| 1.0–1.33 | Insert grade mismatched, feed variation, workpiece material variation | Optimise cutting parameters; implement pre-control for material hardness; upgrade insert grade |
| 1.33–1.67 | Tool wear drift, temperature effects, operator variability | Implement automatic tool offset; install coolant temperature control; standardize setup procedures |
| > 1.67 | Normal process variation | Maintain current practices; consider reducing inspection frequency |
Typical Sources of Variation in Deep Hole Drilling
| Source of Variation | Contribution to Total Variance | Control Method |
|---|---|---|
| Tool wear (within insert life) | 30–50% | Scheduled tool change, Cpk monitoring |
| Coolant pressure fluctuation | 10–20% | Regulator check, pressure transducer monitoring |
| Workpiece material hardness variation | 10–15% | Incoming material inspection, Brinell testing |
| Guide bushing wear | 5–15% | Weekly ID measurement, scheduled replacement |
| Machine alignment drift | 5–10% | Quarterly laser alignment check |
| Operator setup variation | 5–10% | Standardised setup procedure, training |
| Temperature effects (coolant, ambient) | 3–8% | Coolant chiller, ambient temperature control |
Process Capability Study Procedure
Step-by-Step Protocol
- Define the quality characteristic — bore diameter, position, surface finish, or straightness
- Establish the tolerance — customer specification or internal standard (IT grade)
- Verify measurement system capability — GR&R must be < 10% of tolerance, or < 30% if marginal
- Select sampling plan — 25+ subgroups of 3–5 parts each, representing full production variation
- Collect data under stable conditions — ensure the process is in statistical control before calculating capability
- Test for normality — Anderson-Darling or Shapiro-Wilk test
- Apply transformation if needed — Johnson, Box-Cox, or Clements method
- Calculate Cp and Cpk — using appropriate formulas for the data type
- Interpret and report — compare against the target Cpk
- Implement ongoing monitoring — establish control limits and reaction plan
Tip: The most common mistake in deep hole drilling capability studies is collecting data without verifying the process is in statistical control. If the X̄-R chart shows out-of-control points (tool wear trend, coolant pressure shift), the calculated Cp/Cpk will be misleading. Stabilise the process first, then measure capability.
FAQ
What Cp/Cpk values should deep hole drilling achieve?
For production deep hole drilling, target Cp ≥ 2.0 (potential capability) and Cpk ≥ 1.67 (actual capability). Published studies confirm that stable BTA and gun drilling processes for bore diameter consistently achieve Cp > 2.0.
Why does bore diameter data from deep hole drilling violate normality?
Bore diameter data is left-truncated (holes cannot be smaller than the drill nominal size) and right-skewed (tool wear and vibration only increase bore size). This requires Johnson transformation, Clements method, or specialised indices like the Cs index.
Can Cp/Cpk be calculated for bore straightness?
Straightness is a unilateral tolerance (maximum allowable deviation), not a bilateral tolerance. Use Pp/Ppk calculated against the maximum allowable value, or compare the distribution directly against the specification limit using the percentage of nonconforming parts.
What control chart is best for deep hole drilling?
X̄-R charts for bore diameter (hourly sampling of 5 parts). For dynamic disturbances like chatter, residual control charts based on ARIMA modelling are more effective at detecting the transition from stable drilling to instability.
How does tool wear affect Cp/Cpk?
Tool wear causes a gradual upward drift in bore diameter (and sometimes an increase in spread). This reduces Cpk over the tool's life even though Cp may remain constant. The solution is scheduled tool changes based on Cpk monitoring — replace inserts when Cpk approaches the minimum acceptable threshold.
How many parts are needed for a capability study?
Minimum 100 parts (25 subgroups of 4 parts each) for a reliable Cp/Cpk estimate. For deep hole drilling specifically, ensure the data spans at least 50% of a complete tool life cycle to capture wear-related variation.
What is the Cs index for hole-drilling data?
The Cs index, proposed by Pearson (1995), is a process capability index specifically designed for the left-truncated, right-skewed distribution characteristic of drilling processes. It extends the Cpmk index by incorporating a skewness adjustment. It is primarily of academic interest; industry practice favours Johnson transformation with standard Cp/Cpk.
What is a residual control chart?
A residual control chart fits a time-series model (ARIMA) to the bore diameter signal, then plots the residuals (actual − predicted) on a standard control chart. Shifts in the residual mean or variance indicate the onset of dynamic instability (chatter, spiralling) 10–20 parts before conventional charts detect a problem.
How is hole position capability assessed?
Hole position within a circular tolerance zone uses multivariate capability analysis per ISO 22514-6 (GB/T 40681.6). The x and y coordinates are treated as a vector, and the tolerance zone is a circle. The multivariate Ĉp compares the process spread to the tolerance circle diameter, while P̂pk accounts for centring within the tolerance zone.
What is the first step to improve Cpk from 1.0 to 1.67?
Reduce variation (increase Cp) before attempting to adjust centring. If the total process variation is 6σ = USL − LSL, the process fills the entire tolerance and no centring adjustment can achieve Cpk > 1.0. First stabilise coolant pressure, tighten tool change intervals, and verify machine alignment — then adjust the tool offset to centre the process on the nominal dimension.
Conclusion
Process capability analysis for deep hole drilling is well established for bore diameter (Cp/Cpk > 2.0 in stable production) and hole position (multivariate Cp > 2.0), but requires awareness of the characteristic skewness and left truncation of drilling data. Standard control charts (X̄-R) detect gradual tool wear drift, while residual control charts are needed for dynamic disturbances like chatter. Process improvement should focus on reducing variation first — stabilising coolant pressure, standardising tool change intervals, and maintaining machine alignment — before attempting to adjust centring. For any deep hole drilling operation producing critical bores, a regular capability monitoring program with Cpk targets of ≥ 1.67 is not a quality department requirement; it is the primary means of detecting when the process is drifting toward scrap before it produces nonconforming parts.