Free SPC Software & Control Chart Generator

A quick decision tree:

• Continuous data (measurements): subgroups of 2–10 → X-bar R. Subgroups > 10 → X-bar S. Individuals (n = 1) → I-MR.
• Pass/fail data (defectives): constant sample size → NP chart. Variable sample size → P chart.
• Defect counts (defects per unit, with multiple defects possible per item): constant area of opportunity → C chart. Variable area → U chart.

When in doubt, pick the chart type that matches how you collect samples — that is what the rules and constants were designed for.

Western Electric Rules (1956 handbook) are the original Shewhart-zone tests:
1. One point beyond 3σ
2. Two of three consecutive points in zone A (beyond 2σ)
3. Four of five consecutive points in zone B (beyond 1σ)
4. Eight consecutive points on one side of the centre line

Nelson Rules (1984, Journal of Quality Technology) extend the set to eight rules, adding trend (six rising or falling), 14 alternating, 15 within zone C, 8 outside zone C, and a stratification test. Most modern SPC software defaults to Nelson Rules 1–4 for routine monitoring and adds 5–8 when investigating odd patterns. This tool runs both sets and lets you toggle individual rules so you do not chase false alarms.

Both compare process spread to the tolerance window. The difference is which sigma you use:

Cpk (capability) uses the within-subgroup standard deviation (σ̂ = R̄ / d₂ or pooled S). This represents the process at its best, when only common-cause variation is present. Use Cpk to compare a stable process to spec.

Ppk (performance) uses the overall standard deviation (σ̂ = s of all individual measurements). This represents the process as it actually performed over time, including any drift or shifts. Use Ppk to report what the customer actually experienced.

Rule of thumb: Cpk ≥ 1.33 for ongoing monitoring, 1.67 for safety-critical or PPAP. If Ppk is much lower than Cpk, your process is unstable.

For the common SPC tasks — control charts, capability and Gage R&R — yes. Minitab (~$1,750/user/year) and JMP (~$1,495/user/year) cover much more (DOE, multivariate, time-series, mixed models) but their everyday SPC features overlap heavily with this tool.

Use this for: routine control charts, Cpk reporting for PPAP, Gage R&R, quick capability checks, training.
Use Minitab/JMP for: DOE, regression modelling, reliability analysis, enterprise audit trails, regulated industries with formal validation requirements.

Many quality engineers use both — this for the daily 80 %, Minitab for the deep 20 %.

Standard AIAG MSA-4 setup: 2–3 appraisers, 5–10 parts, 2–3 trials each. Have each appraiser measure every part the right number of times, in random order, without seeing previous results.

Paste the data into the Gage R&R module — the tool runs both ANOVA (recommended) and Average-Range methods and reports:
• %GRR (gauge repeatability + reproducibility as % of total variation)
• %EV (equipment variation), %AV (appraiser), %PV (part)
• NDC (number of distinct categories)

AIAG thresholds: %GRR < 10 % acceptable, 10–30 % marginal (acceptable for low-cost or non-critical), > 30 % unacceptable. NDC should be ≥ 5.

They are subgroup-size-dependent factors from Shewhart's original work, tabulated in every SPC textbook and in the AIAG SPC manual:

• A2, A3 — used to compute X-bar control limits from R-bar or S-bar
• D3, D4 — used for R-chart limits
• B3, B4 — used for S-chart limits
• d2 — converts R-bar to σ̂ (standard deviation estimate)
• E2 — used for I-chart limits (individuals)

The generator picks the right constant for your subgroup size automatically — you never have to look them up. But knowing they exist helps when you cross-check against another tool.

Yes. The data entry grid accepts paste from Excel (and Google Sheets, Numbers, LibreOffice). You can also import a CSV file directly. For variables charts the expected format is one row per subgroup with one column per measurement; for attribute charts, one row per sample with the count and sample size.

The math, rules and reporting follow the AIAG SPC and MSA reference manuals which ISO 9001, IATF 16949 and PPAP all reference. But the standards require documented procedures, audit trails and records on top of the math — that is your organisation's QMS, not a tool. Use this tool to produce the charts and reports, store the outputs in your QMS and you are within the spirit of the standard.

For an X-bar R chart with subgroup size n:
• X-bar UCL/LCL = X̄̄ ± A₂ × R̄
• R UCL = D₄ × R̄, R LCL = D₃ × R̄ (D₃ = 0 for n ≤ 6)

For an I-MR chart:
• I UCL/LCL = X̄ ± 2.66 × MR̄ (this is E₂ for MR of 2)
• MR UCL = 3.267 × MR̄, MR LCL = 0

Sigma estimate from range data: σ̂ = R̄ / d₂. The constants A₂, D₃, D₄, d₂ depend on subgroup size — the tool handles all of this for you.

A process is in statistical control when only common-cause variation is acting on it — the points fall randomly within the control limits and pass all the rules. It is out of control when a special cause has appeared: a tool wore, an operator changed, a batch differed, a sensor drifted. "Out of control" is not the same as "out of spec" — a process can be out of control but inside spec (lucky), or in control but out of spec (capability problem, not stability problem). You fix stability first, then capability.

No. All measurements, projects and studies are stored locally in your browser (IndexedDB via Dexie). Nothing is sent to a server. To back up or move your data, use Import / Export: export a single study, a whole project or your entire workspace to a JSON file and import it into another browser or computer. You can also clear the browser data to wipe everything.

Puida Oy builds free professional tools for engineers and quality teams. This SPC software was developed against the AIAG SPC and MSA reference manuals. Bug reports and feature requests: hello@puida.com.