FISCHER MMS - Page 92

Glossary
88 MMS
®
Inspection DPM High
A large standard deviation of a gage means poor repeatability. The smaller the
standard deviation, the better the repeatability. The repeatability depends on
the test method and the characteristics of the gage as well as the characteris-
tics of the test piece. Repeatability can be improved, e.g. by averaging from a
series of measurements (mean value from i readings, forming counted averag-
es).
Repeatability conditions
see under Repeatability
Reproducibility
Reproducibility is a term used to describe the random deviations between var-
ious gages, calibrated to various calibration standards, with repeated measure-
ments taken by different people on the same test piece at the same measuring
position (measurements under reproducibility conditions). When a test piece,
the true measuring result of which is known from a marked measuring posi-
tion, is measured by a different gage and different user at the same measuring
position, a different measuring result will normally occur. This is because even
if each gage has been correctly calibrated and tested, different results would
still be measured, e.g. in the case of a ring test with the same test piece with a
marked reference surface for example. The reason is the random deviations of
each individual measurement and hence each mean value of a measurement
series, the arbitrary deviation with every calibration, and also because every
calibration standard is labeled with a nominal value that has an unavoidable de-
viation from the true value due to the measurement (expressed by the guaran-
tee error).
Reproducibility conditions
see under Reproducibility
Test piece
 Large surface with several measurement locations.
 Part of any size
Trueness
trueness of a measuring result is determined by comparing the mean value x
of a measurement series on a specimen (comparison test piece) with the
Nominal value of the specimen acknowledged as true. Trueness is represented
by the difference between the mean value and the nominal value (x
- nominal