© National Instruments
|
2-17
Temperature Sensor Output and Accuracy
The SCB-68A temperature sensor outputs 10 mV/°C and has an accuracy of ±1 °C.
You also can determine the temperature using the following formulas:
T
C
= 100
×
V
t
T
K
=
T
C
+ 273.15
where
V
t
is the temperature sensor output voltage;
and
T
C
,
T
K
, and
T
F
are the temperature readings in degrees Celsius, Kelvin, and
Fahrenheit, respectively.
Thermocouple Sources of Error
When taking thermocouple measurements with the SCB-68A, the possible sources of error are
as follows:
•
Compensation error
—Can arise from two sources—inaccuracy of the temperature sensor
and temperature differences between the temperature sensor and the screw terminals. The
temperature sensor on the SCB-68A is specified to be accurate to ±1 °C. You can minimize
temperature differences between the temperature sensor and the screw terminals by
keeping the SCB-68A away from drafts, heaters, and warm equipment.
•
Linearization error
—A consequence of the polynomials being approximations of the true
thermocouple output. The linearization error depends upon the degree of polynomial used.
•
Measurement error
—The result of inaccuracies in the DAQ device. These inaccuracies
include gain, offset, and noise. Accuracy can be calculated from the DAQ device
specifications. For best results, you must use a well-calibrated DAQ device. NI
recommends that you run self-calibration on your DAQ device frequently to reduce error.
•
Thermocouple wire error
—The result of inconsistencies in the thermocouple
manufacturing process. These inconsistencies, or nonhomogeneities, are the result of
defects or impurities in the thermocouple wire. The errors vary depending on the
thermocouple type and the gauge of wire used, but an error of ±2 °C is typical. For more
information about thermocouple wire errors and more specific data, consult the
thermocouple manufacturer.
•
Noise error
—Error due to inherent system noise. Use the average of a large number of
samples to obtain the most accurate reading. Noisy environments require averaging more
samples for greater accuracy.
T
F
9
5
---
T
C
×
32
+
=
white noise
number of samples
-------------------------------------------------
resulting noise
=
