Sample size code letters for 3,201 to 10,000
| Inspection level | S-1 | S-2 | S-3 | S-4 | I | II | III |
|---|---|---|---|---|---|---|---|
| Code letter | C | D | F | G | J | L | M |
| Sample size | 5 | 8 | 20 | 32 | 80 | 200 | 315 |
Level II is the default. Level I gives smaller samples and less protection, level III larger samples. The special levels S-1 to S-4 are for destructive or expensive tests. At level II the sample is 2 % to 6.2 % of the lot.
Level II plans and their real risk
| AQL | n / Ac / Re | Lot at the AQL is accepted | LQ: still accepted 1 in 10 | AOQL (lot 10,000) |
|---|---|---|---|---|
| 0.01 | 1,250 / 0 / 1 (letter Q) | 88 % | 0.18 % | 0.03 % |
| 0.015 | 800 / 0 / 1 (letter P) | 89 % | 0.29 % | 0.04 % |
| 0.025 | 500 / 0 / 1 (letter N) | 88 % | 0.46 % | 0.07 % |
| 0.04 | 315 / 0 / 1 (letter M) | 88 % | 0.73 % | 0.11 % |
| 0.065 | 200 / 0 / 1 | 88 % | 1.1 % | 0.18 % |
| 0.1 | 125 / 0 / 1 (letter K) | 88 % | 1.8 % | 0.29 % |
| 0.15 | 315 / 1 / 2 (letter M) | 92 % | 1.2 % | 0.26 % |
| 0.25 | 200 / 1 / 2 | 91 % | 1.9 % | 0.41 % |
| 0.4 | 200 / 2 / 3 | 95 % | 2.6 % | 0.67 % |
| 0.65 | 200 / 3 / 4 | 96 % | 3.3 % | 0.95 % |
| 1.0 | 200 / 5 / 6 | 98 % | 4.6 % | 1.6 % |
| 1.5 | 200 / 7 / 8 | 99 % | 5.8 % | 2.2 % |
| 2.5 | 200 / 10 / 11 | 99 % | 7.6 % | 3.2 % |
| 4.0 | 200 / 14 / 15 | 98 % | 9.9 % | 4.6 % |
| 6.5 | 200 / 21 / 22 | 99 % | 13.8 % | 7.3 % |
| 10 | 125 / 21 / 22 (letter K) | 99 % | 21.9 % | 11.8 % |
n = sample size, Ac = accept with this many or fewer nonconforming units, Re = reject with this many or more. Where the standard table shows an arrow, the plan of the letter in brackets is used, with its sample size. LQ (limiting quality) is the defect level that this plan still accepts in 10 % of lots. AOQL is the worst average outgoing quality when rejected lots are sorted 100 %. Risks use the binomial model.
What this means: with AQL 2.5 a lot that is exactly 2.5 % defective passes 99 % of the time. A lot that is 7.6 % defective still passes one time in ten. The plan separates good and bad lots only roughly: lots several times worse than the AQL can still pass.
OC curves at level II
Each curve shows how likely a lot is to be accepted at a given incoming defect level. The steeper the curve, the better the plan separates good lots from bad ones. Read more in the AQL guide.
Levels I and III for common AQL values
| AQL | Level I: n / Ac / Re | Level III: n / Ac / Re |
|---|---|---|
| 0.65 | 80 / 1 / 2 | 315 / 5 / 6 |
| 1.0 | 80 / 2 / 3 | 315 / 7 / 8 |
| 1.5 | 80 / 3 / 4 | 315 / 10 / 11 |
| 2.5 | 80 / 5 / 6 | 315 / 14 / 15 |
| 4.0 | 80 / 7 / 8 | 315 / 21 / 22 |
| 6.5 | 80 / 10 / 11 | 200 / 21 / 22 |
Single sampling, normal inspection (ISO 2859-1 Table 2-A, ANSI/ASQ Z1.4). Always check plans for contractual use against the current edition of the standard.
XLSFree AQL Sampling Calculator for ExcelSample size, Ac / Re, accept or reject decision and OC curve. No macros, free download.Other lot sizes
← Lot size 1,201–3,200Lot size 10,001–35,000 →
2–89–1516–2526–5051–9091–150151–280281–500501–1,2001,201–3,20010,001–35,00035,001–150,000150,001–500,000500,001 and over
Frequently asked questions
What sample size do I need for a lot of 3,201 to 10,000 units?
With general inspection level II the code letter is L and the sample size is 200 units (level I: J, 80 units; level III: M, 315 units). For some AQL values the table arrows move you to a different sample size, so always read n together with Ac and Re in the table for your AQL.
How many defects are allowed in a lot of 3,201 to 10,000 units at AQL 1.0, 2.5 and 4.0?
At level II, normal inspection: AQL 1.0: inspect 200, accept with 5, reject with 6; AQL 2.5: inspect 200, accept with 10, reject with 11; AQL 4.0: inspect 200, accept with 14, reject with 15.
Does every lot of 3,201 to 10,000 units get the same plan?
Yes. ISO 2859-1 and ANSI/ASQ Z1.4 group lot sizes into ranges, and every lot in the range 3,201 to 10,000 gets the same code letter. A lot of 10,001 moves to the next range and a larger sample. Only where the sample would be larger than the lot do you inspect the whole lot.