QM BenchQuality engineering tools

Multi-Vari Chart and Box Plot Online: Find the Largest Family of Variation

See whether variation comes from time, from part to part or from inside the part, before you start searching for causes. Multi-vari chart, families of variation and box plots, with a PDF report.How to use ↓

No data yet

Measure a few consecutive parts at several points, repeat it over time, and see which family of variation is the biggest.

  1. 1Enter or paste dataor click Load sample
  2. 2Read the verdictcharts and key figures
  3. 3Download the PDFwith your logo

Where this tool fits

Multi-Vari Chart and Box Plot
Next step
Fishbone DiagramList causes in the largest familyDOE: 2-Level Factorial DesignTest the suspected factors

Multi-vari chart and box plot online

A multi-vari study shows in one chart whether the variation of a characteristic comes from inside the part, from part to part or from time to time. It is one of the quickest ways to narrow a root cause search before running experiments.

Enter time, part, an optional position and the measured value. The tool draws the multi-vari chart, splits the total variation into its families with F-tests and shows box plots by any factor.

How it works in 4 steps: Multi-Vari Chart and Box Plot

  1. Sample5 × 3 × 3time, part, position
  2. ChartMulti-variall values in one view
  3. SplitWithin part 78 %largest family of variation
  4. FocusTapersearch causes in that family
Example: shaft diameter, 3 positions on 3 parts, every hour for 5 hours.

How the Multi-Vari Chart and Box Plot works

The total sum of squares is split into nested families (time, part within time, position within part and repeats). Each family is shown as a share of the total and tested against the family nested inside it.

Box plots use quartiles by the (n + 1) p rule, whiskers to 1.5 IQR and outliers beyond.

When to use it

Use a multi-vari study at the start of a root cause investigation, when a characteristic varies too much and nobody knows why, to decide whether to look at the machine, the material, the fixture or the measurement, and as the first step of a Shainin-style search.

PDF report in one click

Export the study as a PDF with the verdict, the multi-vari chart, the box plot, the table of families of variation and the group statistics.

Private: everything runs in your browser

All calculations run in your browser. Your measurements are never uploaded.

How to run a multi-vari study

  1. Plan the sampling: 3 to 5 consecutive parts, measured at the same positions, repeated 4 to 6 times over the shift.
  2. Enter one row per measurement: time, part, position (optional) and the value.
  3. Read the largest family of variation and look at the chart to see its pattern.
  4. Search for causes inside that family, for example with a fishbone diagram or 5 Why.

If most of the variation is repeat noise, check the gauge first with a Gage R&R. Once the family is known, test the suspected factors with a DOE.

Frequently asked questions

›What is a multi-vari chart?

A chart that shows the main families of variation at once: within a part (positional), from part to part (cyclical) and over time (temporal). The family with the largest spread is where the root cause search should start.

›How do I collect the data?

Take a few consecutive parts (for example 3), measure each at several positions, and repeat this at regular intervals over a shift or a day. Record the time, the part number, the position and the value.

›How are the percentages calculated?

The total sum of squares is split into nested parts: between time periods, between parts within a period, between positions within a part and the remaining repeat variation. Each family is tested with an F-test against the family nested inside it.

›What does the box plot show?

The middle 50 % of the values as a box, the median as a line, the mean as a cross, whiskers up to 1.5 times the box height and outliers as circles. Quartiles follow the same (n + 1) p rule as Minitab.