temperature due to the voltage measurements is a few hundredths of a degree.

THERMOCOUPLE POLYNOMIALS - Voltage to Temperature Conversion

NBS Monograph 125 gives high order polynomials for computing the output voltage of a given thermocouple type over a broad range of temperatures. In order to speed processing and accommodate the CR10's math and storage capabilities, 4 separate 6th order polynomials are used to convert from volts to temperature over the range covered by each thermocouple type. Table 13.4-2 gives error limits for the thermocouple linearization functions.

TABLE 13.4-2. Limits of Error on CR10 Thermocouple Output Linearization (Relative to NBS Standards)

TC Type Range °C

Limits of Error °C

T-270 to 400

 

-270 to

-200

±18 @ -270

 

-200 to

-100

± 0.08

 

-100 to

100

± 0.001

 

100 to

400

± 0.015

J

-150 to

760

± 0.008

 

-100 to

300

± 0.002

E-240 to 1000

-240 to

-130

± 0.4

-130 to

200

± 0.005

200 to

1000

± 0.02

K-50 to 1372

-50 to

950

± 0.01

950 to

1372

± 0.04

REFERENCE JUNCTION COMPENSATION - Temperature to Voltage

The polynomials used for reference junction compensation (converting reference temperature to equivalent TC output voltage) do not cover the entire thermocouple range. Substantial errors will result if the reference junction temperature is outside of the calibrated range. The ranges covered by these calibrations include the CR10 environmental operating range, so there is no problem when the CR10 is used as the reference junction. External reference junction boxes, however, must also be within these temperature ranges. Temperature difference measurements made outside of the reference temperature range

SECTION 13. CR10 MEASUREMENTS

should be made by obtaining the actual temperatures referenced to a junction within the reference temperature range and subtracting. Table 13.4-3 gives the reference temperature ranges covered and the limits of error in the linearizations within these ranges.

Two sources of error arise when the reference temperature is out of range. The most significant error is in the calculated compensation voltage; however, error is also created in the temperature difference calculated from the thermocouple output. For example, suppose the reference temperature for a measurement on a type T thermocouple is 300°C. The compensation voltage calculated by the CR10 corresponds to a temperature of 272.6°C, a -27.4°C error. The type T thermocouple with the measuring junction at 290°C and reference at 300°C would output -

578.7µV; using the reference temperature of 272.6°C, the CR10 calculates a temperature difference of -10.2°C, a -0.2°C error. The temperature calculated by the CR10 would be 262.4°C, 27.6°C low.

TABLE 13.4-3. Reference Temperature

Compensation Range and Linearization

Error Relative to NBS Standards

TC Type

Range °C

Limits of Error °C

T

-100 to 100

± 0.001

J

-150 to 296

± 0.005

E

-150 to 206

± 0.005

K

-50 to 100

± 0.01

ERROR SUMMARY

The magnitude of the errors described in the previous sections illustrate that the greatest sources of error in a thermocouple temperature measurement are likely to be due to the limits of error on the thermocouple wire and in the reference temperature determined with the built-in thermistor. Errors in the thermocouple and reference temperature polynomials are extremely small, and error in the voltage measurement is negligible.

To illustrate the relative magnitude of these errors in the environmental range, we will take a worst case situation where all errors are maximum and additive. A temperature of 45°C is measured with a type T (copper-constantan) thermocouple, using the ±2.5 mV range. The nominal accuracy on this range is 2.5 µV (0.1% of 2.5 mV), which at 45°C changes the temperature by 0.06oC. The RTD is 25°C but is

13-15