Home / Lean Six Sigma / SPC & process capability
When defects, breakdowns or accidents are rare, what you monitor is the interval between them: how many parts, days or hours pass until the next one, and whether that interval is getting shorter. Use it when a P or U chart would be almost all zeros (rate below 1%): all you need is the list of event dates, or the parts/hours between them; you get the chart with limits, the signals of worsening or improvement and the yearly cost of the events.
Pre-sizing and estimating tool: results do not replace design, calculations or reports signed by a licensed professional, nor verification against the applicable codes and standards.
Each point is the interval up to the event. Center line = median; limits from the 0.135% and 99.865% percentiles. Red points fail a test; the number above is the test (B = Benneyan).
How to read: horizontal axis = events in order; vertical axis = interval up to each event (parts, days or hours since the previous one). Here up is good: a long interval = a rare event. Center line = median; dashed = limits. Red numbered points = signals; many points in a row below the median = events are getting more frequent.
Event count against cumulative exposure, with the line expected from the estimated rate and the 90% Poisson band. The curve above the band is deterioration; below, improvement.
How to read: horizontal axis = cumulative exposure (parts or time); vertical axis = number of events up to there. The straight line is what the reference rate predicted and the grey band is normal variation. Curve above the band = events sped up (worse); below = they became rarer (better).
Histogram of the intervals with the count expected from the fitted distribution; the vertical lines mark the center line and the UCL.
How to read: bars = how many intervals fell in each bin; red line = count predicted by the model. Vertical lines = median and UCL. Many bars on the left and a long tail to the right are normal. Bars well above the line in some stretch mean the model does not describe the data well.
Points on the line: the chosen distribution describes the intervals well, and the limits can be trusted.
How to read: each point is an observed interval, sorted, against the value the model predicted for that position. Points on the line = adequate model and reliable limits. Points drifting away at the ends = some intervals are shorter or longer than the model explains; read the signals with caution.
Points that fail a test, with the percentile of the interval in the fitted distribution.
Take this page’s results to another calculator with the fields already filled in.
The calculation report gathers this page's inputs, results, equations, assumptions and references, with fields for project, engineer and revision. The full report also includes the charts and interpretations, in color. In the print dialog, choose Save as PDF.
When the event is rare: defects in PPM, machine breakdowns, accidents, contaminations. A P or U chart would have almost every point at zero and meaningless limits; the G or T chart tracks the interval between events, which changes visibly when the frequency changes.
The G chart counts discrete units between events: parts, cycles, lots or whole days. The T chart measures continuous time (hours, fractional days) and uses the Weibull or exponential. With breakdown dates and times, use T; with the number of parts between defects, use G.
Because intervals between rare events are very skewed: many short and a few very long. The mean is pulled up by the long ones; the median shows the typical interval. The limits are also percentiles (0.135% and 99.865%) rather than mean ± 3σ.
No: on these charts, up is good. A point above the UCL is an unusually long interval without events, a sign of improvement worth understanding. Deterioration shows as low points, usually via test 2 (9 in a row below the center line) or the Benneyan test.
It is the G chart’s own test: it signals when several intervals in a row are zero, that is, events on the same part or the same day. The number of zeros needed depends on p̂; with p̂ = 0.1 it is 3 zeros in a row.
Paste the column of dates (and times) in event order; the page accepts 06/01/2026 07:30, 2026-06-01 07:30 or the Excel date serial number. On the G chart with dates, events on the same day become a zero interval.
β near 1: random breakdowns, unrelated to time since the last repair. β above 1: wear-out, risk grows with time, and preventive maintenance helps. β below 1: breakdowns right after repair, a sign of poor repair work or a bad spare part.
Important notice: Indus Metrics tools produce estimates and pre-sizing results from the data you enter and from published engineering methods. They do not replace design, calculations or reports signed by a licensed professional engineer (PE), nor verification against the applicable codes, standards and regulations, such as OSHA rules and the ASME codes. Material and fuel properties vary: check manufacturer data or your own process analyses before deciding. By using this site you agree to the Terms of Use.
© 2026 Indus Metrics. Found an error? Write to contato@seudominio.com.