15.3 Instrumentation, Testing, Inspection & Quality

15.3 Instrumentation, Testing, Inspection & Quality

Designs are only as good as the measurements that prove them. This section covers the standard mechanical tests that generate design data, the measurement-system concepts that make test results trustworthy, and the statistical tools that show whether a production process can hold the drawing tolerances. Weld inspection methods (VT, PT, MT, UT, RT) are covered in Section 8.1 and hardness testing in Section 2.2.

---

Materials Testing & Mechanical Inspection Protocols

ASTM E8 / E8M Standard Tensile Testing

ASTM E8 defines the standard quasi-static tensile test using a standardized dog-bone test specimen:

     +---------+                                           +---------+
     | Grip    |================== Gage Length L_0 ========| Grip    |
     | Section |         |   Reduced Diameter D_0   |      | Section |
     +---------+                                           +---------+

• Engineering Stress ($\sigma$) & Strain ($\epsilon$):

$$\sigma = \frac{P}{A_0}, \qquad \epsilon = \frac{L - L_0}{L_0} = \frac{\Delta L}{L_0}$$

where $P$ is applied tensile force, $A_0 = \frac{\pi}{4} D_0^2$ is initial original cross-sectional area, and $L_0$ is original gage length ($50\,\text{mm}$ or $2.0\,\text{in}$). Standard gage length proportionality is $L_0 = 4 D_0$ or $5 D_0$.

• True Stress ($\sigma_T$) & True Strain ($\epsilon_T$): Prior to localized necking, assuming constant material volume ($A_0 L_0 = A L$):

$$\sigma_T = \frac{P}{A} = \sigma (1 + \epsilon)$$ $$\epsilon_T = \ln\left( \frac{L}{L_0} \right) = \ln(1 + \epsilon)$$

• Mechanical Property Extraction:
  • Modulus of Elasticity ($E$): Slope of the linear elastic region ($E = \Delta \sigma / \Delta \epsilon$).
  • $0.2\%$ Offset Yield Strength ($S_y$): For materials lacking a distinct yield point, a line parallel to the elastic slope is drawn starting at $\epsilon = 0.002$ ($0.2\%$ strain). Its intersection with the stress-strain curve defines $S_y$.
  • Ultimate Tensile Strength ($S_{ut}$): The maximum engineering stress before localized plastic instability (necking) initiates ($d\sigma / d\epsilon = 0$).
  • Ductility Measures:

$$\%\text{EL} = \frac{L_f - L_0}{L_0} \times 100\%, \qquad \%\text{RA} = \frac{A_0 - A_f}{A_0} \times 100\%$$

where $L_f$ is final length at fracture and $A_f$ is necked cross-sectional area at fracture.

ASTM E23 Charpy V-Notch (CVN) Impact Testing

Measures dynamic fracture toughness by releasing a heavy pendulum hammer of mass $m$ from initial height $h_0$ to fracture a notched specimen ($10 \times 10 \times 55\,\text{mm}$, $2\,\text{mm}$ deep $45^\circ$ V-notch with root radius $0.25\,\text{mm}$):

$$E_{\text{absorbed}} = m g (h_0 - h_f) \quad [\text{Joules or ft-lbf}]$$

• Ductile-to-Brittle Transition Temperature (DBTT):
  • BCC Metals (Carbon and Low-Alloy Steels): Exhibit a pronounced S-shaped transition curve. At elevated temperatures, failure is ductile shear rupture with high energy absorption ($> 50-100\,\text{J}$). Below the DBTT, cleavage fracture dominates with catastrophic low-energy brittle failure ($< 15\,\text{J}$).
  • FCC Metals (Aluminum, Copper, Austenitic 300-Series Stainless Steels): Possess 12 slip systems that remain active at cryogenic temperatures; they do not exhibit a DBTT and retain excellent toughness even at liquid nitrogen temperatures ($-196^\circ\text{C}$).

---

Statistical Process Control (SPC) & Capability Analysis

Statistical process control monitors manufacturing stability. When a process is in statistical control (variations follow a stable normal distribution with mean $\mu$ and standard deviation $\sigma$):

                      Lower Spec                Upper Spec
                        Limit                     Limit
                         LSL                       USL
                          |                         |
                          |        /-------\        |
                          |       /    |    \       |
                          |      /     |     \      |
                          |    /       |       \    |
                 ---------+---+--------+--------+---+----------
                                      mu
                          |<------- 6*sigma ------->|

Process Capability Indices ($C_p$ and $C_{pk}$)

• Process Capability Potential ($C_p$): Measures the allowable tolerance spread relative to the inherent process spread ($6\sigma$), assuming the process is perfectly centered between the Upper Specification Limit ($USL$) and Lower Specification Limit ($LSL$):

$$C_p = \frac{USL - LSL}{6\sigma}$$

• Process Capability Index ($C_{pk}$): Accounts for off-center process centering relative to specification limits:

$$C_{pk} = \min\left( \frac{USL - \mu}{3\sigma}, \frac{\mu - LSL}{3\sigma} \right)$$

• Relationship Between $C_p$ and $C_{pk}$:

$$C_{pk} = C_p (1 - k), \qquad k = \frac{|m - \mu|}{(USL - LSL)/2}$$

where $m = (USL + LSL)/2$ is the specification midpoint and $k$ is the off-center factor ($0 \le k \le 1$). If the process is perfectly centered, $k = 0$ and $C_{pk} = C_p$.

---

Measurement Fundamentals

• Accuracy is closeness to the true value; precision (repeatability) is closeness of repeated readings to each other; resolution is the smallest change an instrument can display. A gage can be precise but inaccurate if it is out of calibration.
• Calibration traceability: working gages are calibrated against higher-level standards traceable to national standards (NIST in the United States). A common rule of thumb is a test accuracy ratio of at least 4:1, meaning the calibrating standard should be four times more accurate than the tolerance being checked.
• Measurement uncertainty from independent sources combines by root-sum-square: $u_c = \sqrt{u_1^2 + u_2^2 + \dots}$.
• Gage repeatability and reproducibility (gage R&R) splits measurement variation into equipment variation (repeatability) and appraiser variation (reproducibility). AIAG guidance treats a %GRR (relative to tolerance or process variation) under $10\%$ as acceptable, $10\%$ to $30\%$ as possibly acceptable depending on the application, and over $30\%$ as unacceptable.
• Sensors and signal conditioning (strain gages, LVDTs, encoders, thermocouples) are covered in Section 13.2.

Control Charts

Statistical process control separates common-cause variation (the process's normal noise) from special-cause variation (a real change such as tool wear or a new material lot). For an $\bar{X}$–$R$ chart built from subgroups of size $n$:

$$UCL_{\bar{X}} = \bar{\bar{X}} + A_2 \bar{R}, \qquad LCL_{\bar{X}} = \bar{\bar{X}} - A_2 \bar{R}, \qquad UCL_R = D_4 \bar{R}, \qquad LCL_R = D_3 \bar{R}$$

For $n = 5$, $A_2 = 0.577$, $D_4 = 2.114$ and $D_3 = 0$. A point outside the limits, or a non-random pattern such as a long run on one side of the centerline, signals a special cause to investigate. Control limits describe what the process does; specification limits describe what the drawing requires, and the capability indices below compare the two.

Worked Example: Statistical Process Capability ($C_p$ and $C_{pk}$ Analysis)

Problem Statement

A precision CNC grinding cell produces high-pressure hydraulic piston pins. The engineering drawing specifies the outer diameter as:

$$\varnothing 25.000 \pm 0.015\,\text{mm}$$

Statistical sampling of 1,000 finished pins reveals a normal distribution with:

• Sample mean: $\mu = 25.004\,\text{mm}$
• Sample standard deviation: $\sigma = 0.003\,\text{mm}$

Determine:

1. The Upper and Lower Specification Limits ($USL, LSL$).
2. The process capability potential $C_p$.
3. The process capability index $C_{pk}$.
4. Whether the grinding cell meets the standard automotive/industrial capability requirement of $C_{pk} \ge 1.33$.
5. The estimated non-conformance rate (in parts per million, PPM).

Step-by-Step Solution

Step 1: Establish Specification Limits

$$USL = 25.000 + 0.015 = 25.015\,\text{mm}$$ $$LSL = 25.000 - 0.015 = 24.985\,\text{mm}$$ $$\text{Total Tolerance Spread} = USL - LSL = 0.030\,\text{mm}$$

Step 2: Compute Process Capability Potential ($C_p$)

$$C_p = \frac{USL - LSL}{6\sigma} = \frac{0.030\,\text{mm}}{6 \times 0.003\,\text{mm}} = \frac{0.030}{0.018} = 1.667 \approx 1.67$$

Step 3: Compute Process Capability Index ($C_{pk}$)

Evaluate upper and lower capability ratios:

$$CPU = \frac{USL - \mu}{3\sigma} = \frac{25.015 - 25.004}{3 \times 0.003} = \frac{0.011}{0.009} = 1.222$$ $$CPL = \frac{\mu - LSL}{3\sigma} = \frac{25.004 - 24.985}{3 \times 0.003} = \frac{0.019}{0.009} = 2.111$$

Take the minimum:

$$C_{pk} = \min(CPU, CPL) = \min(1.222, 2.111) = 1.22$$

Step 4: Industrial Capability Conformance Evaluation

Although the inherent process spread is excellent ($C_p = 1.67 > 1.33$), the process mean is shifted high by $+0.004\,\text{mm}$. Because $C_{pk} = 1.22 < 1.33$, the grinding cell fails the standard industrial capability threshold.

Step 5: Estimate PPM Non-Conformance

Compute the standard normal deviate $Z$ for the upper limit:

$$Z_{USL} = \frac{USL - \mu}{\sigma} = \frac{25.015 - 25.004}{0.003} = \frac{0.011}{0.003} = +3.67$$

From standard normal cumulative distribution tables, the probability of exceeding $Z = +3.67$ is approximately $P(Z > 3.67) = 0.000121$, or $121\,\text{PPM}$. Re-centering to $\mu = 25.000$ gives $Z = 5.0$ on each side, reducing total scrap to about $0.6\,\text{PPM}$ ($0.3\,\text{PPM}$ per side).