Statistical process control / X̄–R

Separate normal process variation from a real stability signal.

The Mechatrovich X̄–R SPC Kit converts rational subgroup measurements into readable control charts—so unusual movement is visible before it becomes routine output.

Try the limited chartUnderstand X̄–R first
Both X̄ and R chartsEditable subgroup setupAutomatic control limitsOffline engineering workflow
01 / The problem

A process can meet specification and still be statistically unstable.

Specification limits describe acceptable product. Control limits describe the process behaviour observed over time. Mixing the two answers the wrong question: an in-spec point can still signal a process change, while a stable process can still be incapable of meeting specification.

Use the charts in sequence: check the R chart first to confirm within-subgroup variation is stable, then interpret the X̄ chart for movement in the process average.
Specification limitsCome from design or customer requirements. They define acceptable output.
Control limitsAre calculated from process data. They identify statistically unusual behaviour.
02 / Method

One subgroup produces two signals.

The X̄ chart tracks the average of each subgroup. The R chart tracks the range within each subgroup. For subgroup sizes from 2 to 10, standard SPC constants convert the grand average and average range into control limits.

01Collect rational subgroupsMeasurements should represent a short, meaningful window of process output.
02Check range stabilityAn unstable R chart makes the estimated process variation unreliable.
03Check process averageInterpret X̄ signals only after within-subgroup variation is understood.
03 / What this check decides

Before capability, confirm that the process is behaving consistently.

An X̄–R study answers a different question from Cp or Cpk. Control charts ask whether process behaviour is statistically stable over time. Capability analysis asks whether a stable process can fit inside specification limits.

STEP 01 · MSAMeasurement SystemCan I trust the data?
STEP 02 · SPCProcess StabilityIs variation behaving consistently?
STEP 03 · CAPABILITYProcess CapabilityCan the stable process meet specification?
Can I trust within-subgroup variation?Read the R chart first. A range signal means the short-term variation estimate is not stable yet.
Is the process average consistent?If the R chart is stable, inspect X̄ for evidence that the process center changed.
Can I continue to capability?Only continue when there is reasonable evidence that the process is stable enough for capability to be meaningful.
04 / Read the charts in order

R first. X̄ second. Capability later.

Both charts come from the same subgroups, but they answer different stability questions. Reading them in the wrong order can hide the reason a process changed.

R CHART · WITHIN-SUBGROUP VARIATION

Is short-term variation consistent?

Use the R chart first. If the spread inside a subgroup changes unexpectedly, the variation estimate used by the X̄ limits and later capability analysis becomes questionable.

Example: one range point beyond UCL → investigate variation first
  • measurement or fixture condition
  • setup, material, or tool change
  • subgroup formation or sampling issue
R FIRST →
X̄ CHART · PROCESS AVERAGE

Is the process center remaining consistent?

Once within-subgroup variation is understood, use X̄ to look for movement in the process average. A signal suggests that process location may have changed.

Example: process-location signal → investigate the shift before Cpk
  • machine adjustment or setup change
  • tool wear, replacement, or material lot
  • temperature, operator, or intervention effect
Engineering sequence: R chart → X̄ chart → capability. If either stability check shows a meaningful signal, investigate before treating Cp or Cpk as the process decision.
05 / Working trial

Build both control charts from subgroup data.

Enter measurements in the table, or load the adhesive-dose case below. The trial calculates subgroup averages, ranges, center lines and control limits for your selected subgroup size.

Loads six subgroups of four measurements and replaces the current table. Read the case and investigation.

X̄–R STABILITY CHECK / SPC-02LIVE LIMITED VERSION
3–6 equal-size subgroups, 2–5 samples each.
Trial boundary: this limited check evaluates points beyond control limits. Longer run/trend rules require a fuller study.
STABILITY ASSESSMENT

Ready to calculate

X̄̄—
R̄—
Subgroup n—
Signals—
R CHART / WITHIN-SUBGROUP VARIATIONCheck this chart first
X̄ CHART / PROCESS AVERAGEInterpret after the R chart
What was checked

Run the calculation to generate an interpretation.

A short trial cannot establish long-term statistical control. Use enough rational subgroups and your approved SPC procedure for production decisions.

06 / Interpretation

A control-limit signal changes what you do next.

The chart is not only a pass/fail display. The location of the signal tells you which part of the process model needs attention before you move forward.

R signalWithin-subgroup variation is unstable. Investigate measurement conditions, setup changes, material differences, or subgroup formation before capability.
X̄ signalVariation may be reasonably consistent, but the process center moved. Look for machine adjustments, tool changes, material lots, temperature, or operator intervention.
No limit violationNo immediate special-cause signal was detected in this limited check. Longer run and trend rules may still matter before declaring long-term control.
X̄–R RESULT → WHAT TO DO NEXT
R CHART SIGNALInvestigate variationDo not use capability yet because the process-variation estimate is not stable.
→
R STABLE · X̄ SIGNALInvestigate process shiftFind what changed in the process center before using capability as the next decision.
→
NO IMMEDIATE SIGNALContinue to capabilityNow ask whether the stable process can meet specification using Cp / Cpk.
07 / Why SPC comes before Cpk

A capable-looking process can still be unstable.

Specification limits and control limits answer different questions. A point can be inside specification while still representing an unusual process event. Likewise, a statistically stable process can consistently produce output that does not meet specification.

CONTROL LIMITSWhat is normal for this process?

Calculated from process behaviour. They help distinguish routine variation from statistically unusual behaviour.

SPECIFICATION LIMITSWhat is acceptable to the customer or design?

Defined by requirements. They do not tell you whether the underlying process is statistically stable.

Stable ≠ capableIn specification ≠ statistically stableSPC first → Cpk second

Choose the format for your study.

Free browser trial3–6 subgroups, 2–5 samples per subgroup. Calculates X̄–R charts and checks points beyond limits. Use it to explore this example.
Downloadable SPC KitXbar–R control-chart toolkit for subgroup data and process-stability review. Use the offline file package for your own engineering analysis; review the included files and requirements on Gumroad.

Build the complete stability study.

Offline X̄–R SPC Kit · Complete file package · No subscription.

View SPC Kit details on Gumroad

Check the package contents, requirements and current price before purchase. Purchase and file delivery are handled through Gumroad. Engineering acceptance criteria remain subject to your approved quality procedures.