Each line = one subgroup (time point / batch / location); columns within a line = several readings taken in order inside that subgroup (e.g. consecutive pieces, cavities of one mold). Start a line with a time/batch label (e.g. 08:00). Top-to-bottom = time order. Which is larger — between-subgroup (time trend) or within-subgroup (random noise) — is obvious at a glance. Typical use: sample 4 pieces each hour to see whether shaft diameter drifts by shift, or mold-cavity differences vs within-cavity noise.
Paste data (one subgroup per line, multiple readings within a line)
🤖 AI Interpretation
AI is analyzing, please wait…
What does a multi-vari chart tell you?
· Any process variation splits into two types: ① systematic variation that changes with time/batch (tool wear, material lot, temperature, shift change) and ② within-subgroup random variation (short-term chance noise). Aim at the wrong one and your improvement is wasted
· How to read it: if the dark-green line (subgroup mean) rises/falls clearly or even crosses several light-gray bands → variation is mainly between subgroups / over time; look for factors that "move with time". If the dark-green line is flat and only the light-gray lines jump → variation is mainly within-subgroup random; work on the equipment/gauge itself
· The tool also gives a quantitative cue: mean shift / average within-range >= 2, or means with >=5 consecutive points on one side, both point to systematic variation worth following up
· To formally test between-group differences → "One-Way ANOVA"; to keep monitoring drift over batches → "SPC control chart" I-MR; if you suspect the gage itself is noisy → "Gage R&R"
About Multi-Vari Chart
Generate multi-vari charts online: break variation into within-group and between-group sources by position, time and batch, with AI interpretation.