Control Chart Limits and Reliability Practice

Free control chart and reliability practice for the ASQ CQE and Six Sigma belts. Work out X-bar, R, p and c chart limits from the NIST factors, or the reliability of series, parallel and r out of n systems, and Check my answer names the slip, such as A₂ used for the R chart or a percent where a proportion belongs.

Practise

   
   
   

Check your answer to see the working.

Factors and formulas follow the NIST/SEMATECH e-Handbook. NIST prints D₄ as 2.575 for n = 3 and 2.115 for n = 5, where some books print 2.574 and 2.114: both are accepted.

Every data set is invented for practice.

Practise the exams

The ASQ CQE body of knowledge lists p, c and X-R control charts (VI.F.5) and MTBF, failure rate and the exponential model (III.E), and the CSSGB and CSSBB bodies of knowledge cover control charts too. Start with the free real exam questions, then get the full pack.

ASQ CQE
160 real exam questionsUS$39
Free: 60 questionsGet the full pack
ASQ CSSBB Black Belt
211 real exam questionsUS$39
Free: 26 questionsGet the full pack
ASQ CSSGB Green Belt
144 real exam questionsUS$39
Free: 48 questionsGet the full pack
IASSC ICBB Black Belt
324 real exam questionsUS$39
Free: 107 questionsGet the full pack

Control chart and reliability FAQ

The formulas and NIST factors in short
  • X-bar chart: grand mean ± A₂ × R̄, centre line the grand mean.
  • R chart: UCL D₄ × R̄, LCL D₃ × R̄, centre line R̄.
  • p chart: p̄ ± 3√(p̄(1 - p̄) ÷ n), with p̄ as a proportion.
  • c chart: c̄ ± 3√c̄. A negative LCL means no LCL.
  • Series: R₁ × R₂ × R₃. Parallel: 1 - F₁ × F₂ × F₃.
  • r out of n: add C(n,i) Rⁱ Fⁿ⁻ⁱ for i = r to n.
  • Exponential: R(t) = e^(-t/MTBF). In series, failure rates add.
n A₂ D₃ D₄
2 1.880 0 3.267
3 1.023 0 2.575
4 0.729 0 2.282
5 0.577 0 2.115
6 0.483 0 2.004
7 0.419 0.076 1.924
8 0.373 0.136 1.864
9 0.337 0.184 1.816
10 0.308 0.223 1.777
How do you calculate X‑bar and R chart control limits?

Use the grand mean (X double bar), the mean range R̄ and the NIST factors for the subgroup size n. X-bar chart: UCL = grand mean + A₂ × R̄ and LCL = grand mean - A₂ × R̄, with the grand mean as the centre line. R chart: UCL = D₄ × R̄ and LCL = D₃ × R̄, with R̄ as the centre line. For n = 5, a grand mean of 50.00 and R̄ = 2.00, the X-bar limits are 51.15 and 48.85 and the R chart runs from 0 to 4.23.

Can I use A₂ for the R chart?

No. NIST defines A₂ = 3 ÷ (d₂√n) to set the limits for subgroup means, so it belongs to the X-bar chart only. The R chart has its own factors: UCL = D₄ × R̄ and LCL = D₃ × R̄.

Why is the R chart LCL 0 for small subgroups?

NIST's table gives D₃ = 0 for subgroups of 2 to 6, so LCL = D₃ × R̄ = 0. D₃ first rises above 0 at n = 7, where it is 0.076.

How do you calculate p chart limits?

For m samples of n units each, p̄ = total nonconforming ÷ (m × n), and the limits are p̄ ± 3√(p̄(1 - p̄) ÷ n). In the NIST example, 30 wafers of 50 chips give p̄ = 0.2313, UCL = 0.4102 and LCL = 0.0524.

Should p be a percent or a proportion?

A proportion. NIST notes that a fraction can be written as a decimal or, multiplied by 100, as a percent, but the limit formula needs the decimal: 12% goes in as 0.12. With 12 in it, 1 - p̄ turns negative.

How do you calculate c chart limits?

c̄ = total defects ÷ number of inspection units, and the limits are c̄ ± 3√c̄. In the NIST example, 400 defects on 25 wafers give c̄ = 16, UCL = 16 + 3 × 4 = 28 and LCL = 16 - 3 × 4 = 4.

What if the LCL comes out negative?

NIST says that if the LCL comes out negative, there is no lower control limit. On a c chart that happens whenever c̄ is below 9, since 3√9 = 9. Many books write that limit as 0, so the drill accepts 0 or none.

How do you find the reliability of series and parallel systems?

In series every part must work, so multiply the reliabilities: 0.95 × 0.90 = 0.855. In parallel the system fails only when every part fails, so multiply the unreliabilities and take the result from 1: 1 - 0.05 × 0.10 = 0.995. Reliabilities are never added.

How do you work out an r out of n system?

For n identical parts of reliability R, add the binomial chances that r, r + 1, and so on up to n parts work: the sum of C(n,i) × Rⁱ × (1 - R)ⁿ⁻ⁱ for i = r to n. For 2 out of 3 parts at 0.90: 3 × 0.81 × 0.10 + 0.729 = 0.972. With r = n it is the series model, and with r = 1 the parallel model.

How many parallel parts reach a target reliability?

Find the smallest n with 1 - Fⁿ at or above the target, where F = 1 - R. For parts of 0.80 and a target of 0.99, two parts give 1 - 0.2² = 0.96 and three give 0.992, so you need three.

How do you get R(t) from the MTBF?

For the exponential model, R(t) = e^(-λt) and λ = 1 ÷ MTBF, so R(t) = e^(-t/MTBF). In the NIST example, λ = 0.01 per hour at 100 hours gives a CDF of 0.6321, so R(100) = 0.3679. In series the failure rates add, so the system MTBF is 1 ÷ (λ₁ + λ₂).

Why does my book print a different D₄?

NIST prints D₄ = 2.575 for n = 3 and 2.115 for n = 5, where some tables print 2.574 and 2.114. The drill uses the NIST table and accepts an answer from either.

Does the drill cover np and u charts or availability?

No. The NIST pages it is built on name the u chart but give no limits for np or u charts and no availability formula, so they are left out.

Sources

Checked 20 November 2026 against the NIST/SEMATECH e-Handbook of Statistical Methods and the ASQ bodies of knowledge.

More free tools

All free tools