In this paper a control chart based on sample variances from two quality characteristics for monitoring bivariate normal processes is proposed. The points plotted on the chart are the maximum of the values of these two statistics. The proposed chart (VMAX chart) detects process disturbances faster than the generalized variance |S| chart and has a better diagnostic feature, that is, with the VMAX chart it is easier to relate an out-of-control signal to the variable whose variability has moved away from its in-control state. When the double sampling scheme is used the proposed chart has also a better performance, except in a few cases in which the size of the samples at the second stage is very large.
VMAX control chart; covariance matrix; bivariate processes; double sampling
