|
Accuracy and Practicality -
First-Hand Measurement of Thermal Output
There is a limit as to just how far it is
practical to go in adjusting the manufacturer's
thermal output data in an attempt to obtain greater
accuracy. In the first place, the thermal output
curve provided on the technical data sheet (or by the
polynomial equation) represents an average, since
there is some variation in thermal output
characteristics from gage to gage within a lot. And
the width of the scatterband increases as the test
temperature departs further and further from the
room-temperature reference. The spreading of the
scatterband is approximately linear with deviation
from room temperature, at least over the temperature
range from +32° to +350° F (0° to
+175° C) for which scatter data are available.
At the 2
(95%) confidence level, the variability for A alloy
can be expressed as
+
0.15 microstrain/° F (
+
0.27 microstrain/° C), and that of K alloy as
+
0.25 microstrain/° F (
+
0.45 microstrain/° C). Thus, at a test
temperature of +275° F (+135° C), the 2
width of the scatterband is
+
30 microstrain for
A alloy
, and
+
50 microstrain for
K alloy
.
Furthermore, the thermal output data given in the
gage package were necessarily measured on a
particular lot of a particular test material (see the
table
). Different materials with the same or closely
similar nominal expansion coefficients, and even
different lots and forms of the same material, may
have significantly different thermal expansion
characteristics.
From the above considerations, it should be
evident that in order to achieve the most accurate
correction for thermal output it is generally
necessary to obtain the thermal output data with the
actual test gage installed on the actual test part.
For this purpose, a thermocouple or resistance
temperature sensor is installed immediately adjacent
to the strain gage. The gage is then connected to the
strain indicator and, with no loads applied to the
test part, the instrument is balanced for zero
strain. Subsequently, the test part is subjected to
the test temperature(s), again with no loads applied,
and the temperature and indicated strain are recorded
under equilibrium conditions. If, throughout this
process, the part is completely free of mechanical
and thermal stresses, the resulting strain indication
at any temperature is the thermal output at that
temperature. If the instrument gage factor setting
during subsequent strain measurement is the same as
that used for thermal output calibration, the
observed thermal output at any test temperature can
be subtracted algebraically from the indicated strain
to arrive at the corrected strain. Otherwise, the
thermal output data should be adjusted for the
difference in gage factor settings prior to
subtraction.
In order to correct for thermal output in the
manner described here, it is necessary, of course, to
measure the temperature at the strain gage
installation each time a strain measurement is made.
The principal disadvantage of this procedure is that
two channels of instrumentation are preempted for
each strain gage -- one for the strain gage proper,
and one for the thermocouple or resistance
temperature sensor.
|
Vishay Measurements Group
