Orton | Industrial Pyrometrics | Temperature Measurement Traceability

Pyrometrics

Temperature Measurement Traceability

In 1980 a team from the National Bureau of Standards-NBS (now National Institute of Standards and Technology-NIST) visited Orton to assist in establishing a temperature measurement capability which would be directly traceable to NBS/NIST. Recommendations concerning equipment and procedures, as suggested by the team, were followed to provide a strong traceable link to NBS/NIST.

The furnaces selected for this work were designed in-house but closely resembled proven commercial models. Hairpin MoSi2 heating elements with fiberboard insulation were used with alumina muffle to provide a hot zone uniform to about ±5°C at 600°C and ±1°C at 1200°C. Electrically the furnaces were designed to operate at a maximum power of 5 KVA (100V at 50 amp) and heating rates up to 300°C/hr to 1600°C could be achieved.

Need more help:



		or call us: (614) 818-1338

Thermocouples (Type S, Type B) in which the reference junction is created within a precision ice point reference where the noble metal wire is welded to copper wire which then extends to the EMF measuring system.

The thermocouple wire (obtained from Engelhard) is premium grade Type S or Type B (depending on use temperature), 0.020" diameter and is used in high purity (>99.7%), 4 bore alumina insulators. Three thermocouples are employed in each furnace (measuring reference, and control). A Kaye Ice Point Reference (K140-4) is used to maintain the reference (cold) junctions of the thermocouples to 0.0±0.1°C. A Ð10 to +10°C mercury-in-glass thermometer with 0.1°C subdivisions and NIST certificate is used to monitor reference temperature to within an accuracy of 0.25°C.

For cones, temperature measurement is made with a Fluke 2400A Measurement and Control Link. For Type S thermocouples the total instrument accuracy is 0.4°C for a 24 hour period with 21-25°C ambient and reading rate of 3 per sec. The accuracy values include all error sources (connector gradient, reference junction, linearization conformity, A to D conversion, temperature coefficient, time drift, and scanner offsets) but excludes sensor error. Since an external reference junction is used, the total accuracy will be slightly better than stated since this error source will not exist here.

For TempCheks, an HP34401 Multimeter (24 hour measurement accuracy at 23±1°C is 0.003% of volt reading + 0.003% of range and 1 year measurement accuracy at 23±5°C is 0.005% of volt reading +0.0035% of range) is used for temperature measurement and an Orton Model 29 Programmer Controller for furnace control. In cone work, all three thermocouples are Pt, Pt-Rh extending to the ice point. In TempChek work, the control thermocouple uses Type R, S lead wire (copper, alloy #11) to reach the ice reference.

An EDC DC Calibrator, Model 520A, was available to calibrate the Fluke 2400A and HP 34401. The 1 year limit of error for this instrument is specified as ±(0.002% of setting + 0.0005% of range) for voltage measurements. Traceability of measurement is provided by a certificate of compliance issued by the Quality Control Department at Electronic Development Corporation (EDC). The preferred practice is to have this instrument re-calibrated at EDC on a yearly basis.

A quantity of Type B and Type S thermocouple wire was ordered sufficient to complete all standardization work for pyrometric products. Each spool of each type was sampled by removing wire at the beginning and end of the spool (Type S) or at the beginning, middle, and end of the spool (Type B) in lengths sufficient to be used as thermocouples.

The thermocouples are fabricated by welding, using a carbon arc with dedicated (noble metal only) electrode which is cleaned prior to use. The thermocouples are of the 3 wire design in which one of the duplicated wires can be grounded during use.

After fabrication, the 5 thermocouples without insulators were sent to NBS/NIST for calibration testing. This procedure generated a temperature/EMF table at 1°C increments for each of the 5 thermocouples in which the EMF values were given to 1 µV (about 0.1°C). The NBS/NIST uncertainties assigned to these thermocouples are as follows:


NIST	Type "B"	0-600°C	0.5°C
Test #	0-1450°C	600-1100°C	0.7°C
32010C		1450°C	2.0°C
.	.	.	.
NIST	Type "S"	800-1100°C	0.7°C
Test #	800-1750°C	1450°C	2.0°C
32031C		1750°C	3.0°C

Based on NBS/NIST recommendations, the results of the Type S and Type B tests are each averaged so that temperature corrections for each set of spools can be inferred with the same degree of uncertainty that was obtained with the wires actually tested. Actual temperature corrections were calculated at Orton by comparison of temperature/EMF data for the thermocouples submitted to NBS/NIST with standard tables. All temperatures are based on IPTS-68. In the near future, new thermocouple tables will be available which are based on ITS-90, which will permit conversion of current correction data to the 1990 temperature scale.

Based on further NBS/NIST recommendations, Type S thermocouples were to be used up to 1200°C. Above that temperature, Type B thermocouples were specified to minimize high temperature thermoelectric drift. A major challenge to Orton is to add minimum temperature error to the starting uncertainty of noble metal thermocouples (e.g. 2°C for Type B at 1450°C). This is justification for the use of high precision measuring equipment and strict operating procedures.

Positioning these two thermocouples in interlocking contact makes provision for calibration of the measuring thermocouple against the reference thermocouple. In cone work, the thermocouples are positioned side by side in the same horizontal plane. In TempChek testing, the arrangement is side by side in the same vertical plane. During calibration the heating rate of the furnace is programmed for 60°C/hr when the furnace reaches within 150°C of the first test temperature. The temperature is held constant (Fluke: ±0.2°C, Model 29: <±1°C) for 1 hr. A minimum of 5 readings each from reference and control thermocouples are taken at 1-minute intervals and averaged. For cones, calibration temperatures are at 50°C intervals over the desired temperature range. For TempCheks the interval is 100°C. In product testing runs, the calibration thermocouple is removed from the hot zone of the furnace.

Generally, a calibration run is performed after every 20-30 product testing runs as experience to date has shown that thermoelectric drift is minimal over this frequency range provided that Type S thermocouples do not go above 1200°C.

At that time the EMF measurement instruments are calibrated against the EDC DC Calibrator. An alternative to in-house calibration of thermocouples is to send the measuring thermocouple at periodic intervals to NIST for re-calibration at that location.