8.2 Attribute Sampling with ANSI/ASQ Z1.4

8.2 Attribute Sampling with ANSI/ASQ Z1.4

Standard Architecture and Scope of ANSI/ASQ Z1.4

ANSI/ASQ Z1.4 (Sampling Procedures and Tables for Inspection by Attributes) is the predominant national and international standard for attribute acceptance sampling across defense, aerospace, automotive, medical device, and commercial manufacturing sectors. It is the direct civilian successor to the historic U.S. military standard MIL-STD-105E and is technically harmonized with international standard ISO 2859-1.

Core System Principles

• Attribute Inspection: Evaluates units on a binary pass/fail, conforming/nonconforming, or go/no-go basis (such as thread plug gage engagement, visual cosmetic blemish check, or leak/no-leak pressure testing), or counts the total number of nonconformities per 100 units.
• Indexed by AQL: ANSI/ASQ Z1.4 is fundamentally designed around the Acceptable Quality Level (AQL). The standard assumes a continuing series of production lots submitted by a supplier whose long-term process average operates at or better than the agreed AQL.
• Not Intended for Isolated Lots: While Z1.4 contains Limiting Quality (LQ) supplementary tables for isolated lots, its primary mathematical architecture and protection mechanisms rely on the continuous application of dynamic switching rules over successive lots.

---

Inspection Levels: General vs. Special

To determine the sample size for a given lot size $N$, an inspector must first identify the designated Inspection Level specified on the quality plan, blueprint, or purchase order. Z1.4 divides inspection levels into two broad categories:

General Inspection Levels (I, II, III)

General inspection levels govern the vast majority of non-destructive visual, dimensional, and mechanical attribute inspections:

• Level II (Normal / Standard Default): Level II is the mandatory default inspection level specified by ANSI/ASQ Z1.4 unless an engineering drawing or contractual quality requirement explicitly mandates otherwise. It provides a balanced baseline between inspection labor cost and statistical protection.
• Level I (Reduced Discrimination): Requires approximately $40%$ smaller sample sizes than Level II. It is utilized when inspection costs must be minimized, the supplier has a well-proven track record of high quality, or less discrimination between good and bad lots can be tolerated.
• Level III (Heightened Discrimination): Requires approximately $50%$ larger sample sizes than Level II. It is utilized when the component is critical, the supplier has an unstable quality history, or the customer demands the highest statistical discrimination without transitioning into full 100% screening.

Special Inspection Levels (S-1, S-2, S-3, S-4)

Special inspection levels yield substantially smaller sample sizes (Code Letters A through G) and are strictly reserved for situations where:

• The inspection test is destructive (e.g., tensile rupture, ballistic puncture, solder joint shear).
• The inspection test is exceptionally expensive or time-consuming (e.g., 500-hour salt spray corrosion testing, metallurgical microsectioning).
• Large sampling risks (higher consumer risk $\beta$) can be tolerated by engineering agreement.
• S-1 provides the smallest sample size (highest risk); S-4 provides the largest sample size among the special levels.

---

Determining Sample Size Code Letters and Plan Selection

Executing a sampling plan under ANSI/ASQ Z1.4 follows a rigid step-by-step procedural workflow:

  +-------------------------------------------------------------------------+
  | STEP 1: Determine Lot Size (N) and Inspection Level (Default = Level II) |
  +-------------------------------------------------------------------------+
                                       |
                                       v
  +-------------------------------------------------------------------------+
  | STEP 2: Consult Table I to find Sample Size Code Letter (A through R)    |
  +-------------------------------------------------------------------------+
                                       |
                                       v
  +-------------------------------------------------------------------------+
  | STEP 3: Enter Master Sampling Table (Table II-A for Normal Single)      |
  |         Locate Code Letter row for Sample Size (n) and AQL column       |
  +-------------------------------------------------------------------------+
                                       |
                                       v
  +-------------------------------------------------------------------------+
  | STEP 4: Read Acceptance Number (Ac) and Rejection Number (Re)           |
  |         (Follow vertical arrows if indicated; update n accordingly)     |
  +-------------------------------------------------------------------------+

Interpreting Directional Arrows in Z1.4 Tables

When indexing a specific Code Letter and AQL in the master sampling tables, inspectors frequently encounter vertical arrows:

• Downward Arrow ($\downarrow$): Use the first sampling plan below the arrow. Crucial rule: The inspector must adopt both the new acceptance/rejection numbers ($Ac, Re$) AND the new sample size ($n$) corresponding to the new code letter row pointing to the numbers.
• Upward Arrow ($\uparrow$): Use the first sampling plan above the arrow, adopting the new $Ac, Re$ and the corresponding new sample size $n$.
• Exception Rule: If the new sample size exceeds the lot size $N$, perform 100% inspection.

---

Sampling Plan Types: Single, Double, and Multiple

ANSI/ASQ Z1.4 provides three structural sampling plan architectures:

1. Single Sampling Plans (Table II-A, II-B, II-C)

In a single sampling plan, the disposition decision is reached after evaluating a single random sample of size $n$:

• Draw $n$ units at random from the homogeneous lot.
• Count the total number of nonconforming units (or nonconformities), denoted as $d$.
• If $d \le Ac$ (Acceptance number): Accept the lot.
• If $d \ge Re$ (Rejection number): Reject the lot.
• Under single sampling, $Re$ is always equal to $Ac + 1$.

2. Double Sampling Plans (Table III-A, III-B, III-C)

A double sampling plan allows a second sample to be drawn if the results of the initial sample are borderline. It is defined by five parameters: $(n_1, Ac_1, Re_1)$ for the first sample, and $(n_2, Ac_2, Re_2)$ for the second sample:

1. Draw initial sample $n_1$ and count defectives $d_1$:
   • If $d_1 \le Ac_1$: Accept the lot immediately.
   • If $d_1 \ge Re_1$: Reject the lot immediately.
   • If $Ac_1 < d_1 < Re_1$: The lot is indeterminate. Draw second sample $n_2$.
2. Count defectives $d_2$ in the second sample and compute cumulative defectives: $d_{\text{cum}} = d_1 + d_2$:
   • If $d_{\text{cum}} \le Ac_2$: Accept the lot.
   • If $d_{\text{cum}} \ge Re_2$: Reject the lot.

Operational Benefits: Double sampling provides significant psychological benefits to suppliers (giving a borderline lot a "second chance"). Furthermore, because clear-cut good or bad lots are decided on sample $n_1$ (where $n_1$ is typically smaller than the equivalent single sample size), double sampling reduces the Average Sample Number (ASN) by $20%$ to $30%$.

3. Multiple Sampling Plans (Table IV-A, IV-B, IV-C)

Extends double sampling up to seven sequential stages. Each stage draws a small sample. Defect totals accumulate across stages until an accept or reject threshold is crossed. While multiple sampling minimizes ASN, its operational complexity, tracking burden, and storage staging requirements make it less common on high-speed shop floors.

---

The Dynamic Switching Rules

The core engine of ANSI/ASQ Z1.4 is its mandatory switching system between Normal, Tightened, and Reduced inspection. These switching rules protect the customer against deteriorating quality and reward suppliers with sustained excellence.

                     +--------------------------+
                     |    REDUCED INSPECTION    |
                     +--------------------------+
                       ^                      |
      10 lots accepted |                      | 1 lot rejected OR
      Production steady|                      | irregular production OR
      Approved by auth |                      | Ac < d < Re condition
                       |                      v
                     +--------------------------+
          +--------> |    NORMAL INSPECTION     | <-------+
          |          |     (Initial State)      |         |
          |          +--------------------------+         |
          |            |                      ^           |
   5 consecutive lots  |                      |           |
   accepted on         | 2 of 5 (or fewer)    |           |
   tightened           | lots rejected        |           |
          |            v                      |           |
          |          +--------------------------+         |
          +--------- |   TIGHTENED INSPECTION   |         |
                     +--------------------------+         |
                               |                          |
                               | 10 consecutive lots      |
                               | remain on tightened      |
                               v                          |
                     +--------------------------+         |
                     |     DISCONTINUATION      |         |
                     |      OF INSPECTION       |         |
                     |  (Stop shipments until   |         |
                     | supplier action verified)| --------+
                     +--------------------------+

1. Normal to Tightened Inspection

• Mandatory Trigger: When 2 out of 5 (or fewer) consecutive lots have been rejected on original Normal inspection.
• Action: Tightened inspection maintains the same sample size $n$ but lowers the acceptance numbers ($Ac$), shifting the OC curve to the left to severely restrict Consumer's Risk.

2. Tightened to Normal Inspection

• Mandatory Trigger: When 5 consecutive lots have been accepted on original Tightened inspection.
• Action: Return to Normal inspection tables.

3. Normal to Reduced Inspection

• Mandatory Triggers (All four conditions must be satisfied simultaneously):
  1. The preceding 10 consecutive lots have been on Normal inspection and all 10 have been accepted;
  2. The total number of nonconforming units in the samples from the 10 lots is less than or equal to the applicable limit number specified in ANSI/ASQ Z1.4 Table VIII (or switching score criteria in updated revisions);
  3. Production is at a steady, uninterrupted rate; AND
  4. Reduced inspection has been formally approved by the responsible quality authority (e.g., Quality Engineering Manager or Customer Quality Assurance).
• Action: Transition to Table II-C (Reduced Single Sampling), which reduces sample sizes by approximately $60%$, significantly lowering inspection overhead.

4. Reduced to Normal Inspection

• Mandatory Trigger: If ANY ONE of the following four events occurs while operating under Reduced inspection, the inspector must immediately revert to Normal inspection on the very next lot:
  1. A lot is rejected;
  2. A lot is accepted, but the number of nonconforming units falls between the acceptance number ($Ac$) and rejection number ($Re$) in plans where $Re \ge Ac + 2$ (the lot is accepted, but the warning triggers return to Normal);
  3. Production becomes irregular, intermittent, or delayed; OR
  4. Other shop conditions (such as tooling breakdown or raw material supplier changes) warrant return to Normal.

5. Discontinuation of Inspection Rule

• Mandatory Trigger: If 10 consecutive lots remain on Tightened inspection without qualifying to return to Normal.
• Contractual Mandate: Inspection under ANSI/ASQ Z1.4 must be formally discontinued. The quality inspector cannot sentence additional lots. Shipments must be suspended, and the supplier must institute formal corrective actions (e.g., 8D or CAPA) to eliminate process causes. Only after corrective actions are audited and approved by the customer quality authority may inspection resume under Tightened inspection.

---

Zero-Acceptance Sampling: Squeglia c=0 Plans

Traditional ANSI/ASQ Z1.4 plans permit acceptance numbers greater than zero (e.g., Code Letter L, AQL 1.5 yields $n = 200, Ac = 7, Re = 8$). Under this plan, if an inspector uncovers 7 nonconforming parts in the sample, the lot is formally stamped as "Accepted." In modern aerospace (AS9100), automotive (IATF 16949), and medical manufacturing, shipping batches containing verified defective parts creates severe ethical, legal, and warranty liabilities.

Principles of Squeglia $c=0$ Plans

Developed by Nicholas L. Squeglia (Zero Acceptance Number Sampling Plans), $c=0$ plans mandate that:

• Acceptance number is always zero ($Ac = 0$) and Rejection number is always one ($Re = 1$).
• If a single defect is found in the sample, the entire lot is rejected immediately.

---

Real Shop Inspection Scenarios & Common Exam Traps

• Exam Trap: Confusing the '2 out of 5' Switching Rule: A classic CQI exam question asks: "Under ANSI/ASQ Z1.4, when must an inspector switch from Normal to Tightened inspection?" Candidates often choose "When 5 consecutive lots are rejected." That is wrong! The rule is 2 out of 5 (or fewer) consecutive lots rejected on original inspection. In contrast, switching back from Tightened to Normal requires 5 consecutive lots accepted.
• Exam Trap: Following Directional Arrows Without Updating Sample Size: When a Z1.4 table contains a downward arrow, the inspector must follow the arrow to the first available $Ac/Re$ numbers. The critical trap: the inspector must also use the new sample size $n$ from that new row, NOT the sample size from the original code letter!
• Exam Trap: The Reduced Inspection 'Gap' ($Ac < d < Re$): In some reduced sampling plans, $Ac = 1$ and $Re = 3$. If an inspector finds exactly 2 nonconforming units, what happens? The lot is accepted, BUT the inspection plan must immediately revert to Normal inspection for subsequent lots.