16
Evaluate the control charts for unstable conditions. (Your software
should have the capability to do this for you).
Analysis:
Both of the control charts should appear to be stable. If either of them
is not stable, you will need to determine and fix the root cause of the
out-of-control points and take several more groups of data to prove that
the measurement system is back in control.
An instability in the R chart implies that the process variability (within
subgroup) is not stable. An example of this would be that the parts
were moving in the fixture.
An instability in the
X
control chart implies that some factor has
caused a shift in the measurements. This can be from a change in
environment, improper fixturing, change in gauge, etc.
Maintain Control:
It is important to maintain control in the measurement system. If it ever
turns out that the measurement system is “out-of-control”, any
measurements made with it will have unknown uncertainty in them and
might be far from the true value of the characteristic being measured.
Every time the measurement system is prepared for measuring this
characteristic, a group of measurements should be made on the
same sample standard as above.
The data should then be plotted on the same control chart as above.
The control chart should be evaluated for any unstable conditions.
If an unstable condition exists, the measurement system should not
be used until the instability has been remedied and the stability
reestablished.
