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

Page: 97 / 188

Random and systematic deviations Measurement device monitoring for the Fischerscope X-RAY
FISCHERSCOPE
®
X-RAY 97
Please note that we expressly speak of the trueness as a property
of the
measurement readings . However, many believe the trueness
to be a property of the instrument! Unprofessionally, this is
expressed in the question for the ”accuracy of the instrument”.
However, the trueness is obviously very dependent on the settings
(e.g., the calibration) and the monitoring of the instrument. These in
turn depend on the competency and dependability of the user. For
this reason, the instrument manufacturer is not able to guarantee
the trueness of the values measured with this instrument or even
apply a sticker to the instrument stating the trueness. Each
measurement result is as correct or wrong as allowed by the user.
Thus, the trueness cannot be a criterion for the quality of an
instrument, at best it can be an indication of the stability (trueness
during long-term operation of the instrument without necessary
interim corrective measures).
In measurement technology, the term ”accuracy” is essentially
meaningless and of little relevance. In contrast to it, the terms
trueness and precision that are unambiguous (contrary to the term
accuracy) are preferred. On the international scene of
standardization, the situation is similar: the term
accuracy can be
explained only qualitatively; it is split into the terms trueness and
precision . The term precision in turn is divided into repeatability
(repeatability precision = precision under repeatable conditions)
and
reproducibility (comparative precision = precision under
comparable conditions).
14.5 Random and systematic deviations
Is a single reading that has been measured on a calibration
standard and that deviates significantly from its nominal value, such
that B – A is much bigger or smaller than zero, correct or rather
incorrect? It obviously has error, as long as the calibration standard
is not labeled incorrectly, that is, A being wrong. The question then
remains: is the deviation of B – A from zero a random or a
systematic (i.e., not a random) one?
For an expert, it is not enough to base the answer to this question
on one single reading.
To differentiate between random and systematic deviations , one
needs to make many (>10) measurements under repeatability
conditions on a calibration standard and then compare the nominal
value with the mean value. The first prerequisite for the comparison
is:
the measurement application applies to this calibration
standard. This is very important and is often disregarded in
practical situations! The comparison is particularly inadmissible if
the expected substrate material of the measurement application
deviates significantly from the substrate material of the calibration