FISCHER FISCHERSCOPE X-RAY 5000 Series, Operator'S Manual

Page: 103 / 188

When and how to carry out corrections? Measurement device monitoring for the Fischerscope X-RAY
FISCHERSCOPE
®
X-RAY 103
performs a calibration service, for example, because no calibration
standards are available, the user only has the option to monitor the
measurement device using reference standards. Thus, establishing
reference samples is an obligatory consequence of using a
calibration service, because continued access to standards by the
user is not possible.
To determine the actual situation of the measurement application at
the beginning of the control chart, the characteristics of the
reference sample or of the calibration standard, respectively, are
determined with the secured measurement application in the form
of a larger random sampling (n > 50) without break. This is called a
pre-run . The reason for measuring the random sample in an as
short as possible period is that the instrument should remain stable
during this time such that systematic changes of the instrument can
be neglected during this time span. The result of the pre-run is the
mean value of all group mean values (
nominal value ) and the mean
value of the standard deviation from the single readings within the
groups. The limits for the mean value and for the standard deviation
that are not to be violated by the random deviations are calculated
from the nominal value in relation to the mean standard deviation
and the number of single readings per group. These limits are
called
statistical control limits . With subsequent measurements,
upper and lower violations of these control limits are therefore
caused by systematic changes to the measurement application.
Thus, the control limits characterize when subsequent
measurement results are random or ”natural” (mean value and/or
standard deviation fall within the limits) or when they deviate
systematically or ”unnaturally” from the nominal values determined
during the pre-run (measurement results are outside the limits) and
in doing so indicate a change in the measurement application that
must be corrected. With subsequent regular checks of the
measurement application, using upper or lower violations of the
control limits, it is then very easy to discover on time deviations of
the measurement application from the nominal values and to make
control corrections.
14.8 When and how to carry out corrections?
We recommend the correction of a measurement application as
soon as the mean value or the standard deviation of a group
violates the control limits that have been determined by the pre-run.
As corrective measures we recommend:

a reference measurement,

a normalization (measurement of the spectra of the pure elements