REPORT ON THE INTERCOMPARISON OF THE REALISATION IN CENAM (MEXICO) OF THE ITS-90 BETWEEN THE SILVER POINT AND 1700°C, USING VACUUM TUNGSTEN-STRIP LAMPS AS TRANSFER STANDARDS 1. Introduction The CCT decided during its 19th session in September 1996 to undertake an international Intercomparison of local realisations of the ITS-90 above the silver point, using high-stability vacuum tungsten strip lamps as transfer standards. The intercomparison has been qualified as a key intercomparison with the potential effect of entailing documented bilateral agreements as to the equivalence of local realisations of the ITS-90 in the envisaged temperature range. Here are reported the results obtained at the Centro Nacional de Metrología, of Mexico with the lamps set II of the NPL (C860 and C864 serial numbers), received from the National Research Council of Canada on March 1st, 1998, and returned to the National Physical Laboratory of England on May 1st, 1998. 2. Description of the equipment utilised for the calibration of the lamps • The A. C. thermometric bridge As suggested in the protocol, an ASL F-18 a. c. resistance bridge it was utilised to measure a Rt/Rs ratio to determine the resistance value Ramb of the tungsten strips of each lamp. • The thermalised standard resistor for the measurement of Ramb An 1 ohm Wilkins Resistor, model 5685A, Serial No. 263380 in “Tinsley” thermalised box was utilised as the standard resistor of the Rt/Rs determination for the measurement of Ramb for each lamp. The value of calibration of this standard resistor, at the temperature of the thermalised box (around 36,5°C) is 0,999 970 1 Ω with an estimated uncertainty of 0,8 µΩ/Ω (k=2) • The a. c. thermometric bridge for the measurement of Tamb For the measurement of the temperature Tamb of the bases of the lamps, were employed a type Pt-100 resistance thermometer and an ASL F250 a. c. thermometric bridge. The sensor was calibrated between the Hg and the Ga points, according to the ITS-90. The outer diameter of this thermometer fitted the hole drilled at the base of the lamps, and the thermal contact was aimed by the means of a silicon grease. • The shunt resistor dipped in a bath of controlled temperature for the measurement of Ish, the current supplied to the tungsten strip of the lamps A Yokogawa Standard Resistor, type 2792, serial No. 66VW2059 with a measured d. c. value (by the manufacturer) of 10,000 2 mW at 20°C, was employed as a shunt resistor for the determination of the d. c. current supplied to the lamps, connected in series with the lamp’s socket, from the “current” terminals of the resistor.
377
Embed
2. Description of the equipment utilised for the ...kcdb.bipm.org/appendixb/appbresults/cct-k5/cct-k5_instrumentation... · and for the effective wavelength of the utilised interference
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
REPORT ON THE INTERCOMPARISON OF THE REALISATION IN CENAM (MEXICO) OF THE ITS-90 BETWEEN THE SILVER POINT AND 1700°C, USING VACUUM TUNGSTEN-STRIP LAMPS AS TRANSFER STANDARDS
1. Introduction The CCT decided during its 19th session in September 1996 to undertake an international Intercomparison of local realisations of the ITS-90 above the silver point, using high-stability vacuum tungsten strip lamps as transfer standards. The intercomparison has been qualified as a key intercomparison with the potential effect of entailing documented bilateral agreements as to the equivalence of local realisations of the ITS-90 in the envisaged temperature range. Here are reported the results obtained at the Centro Nacional de Metrología, of Mexico with the lamps set II of the NPL (C860 and C864 serial numbers), received from the National Research Council of Canada on March 1st, 1998, and returned to the National Physical Laboratory of England on May 1st, 1998.
2. Description of the equipment utilised for the calibration of the lamps
• The A. C. thermometric bridge
As suggested in the protocol, an ASL F-18 a. c. resistance bridge it was utilised to measure a Rt/Rs ratio to determine the resistance value Ramb of the tungsten strips of each lamp.
• The thermalised standard resistor for the measurement of Ramb
An 1 ohm Wilkins Resistor, model 5685A, Serial No. 263380 in “Tinsley” thermalised box was utilised as the standard resistor of the Rt/Rs determination for the measurement of Ramb for each lamp.
The value of calibration of this standard resistor, at the temperature of the thermalised box (around 36,5°C) is 0,999 970 1 Ω with an estimated uncertainty of 0,8 µΩ/Ω (k=2)
• The a. c. thermometric bridge for the measurement of Tamb
For the measurement of the temperature Tamb of the bases of the lamps, were employed a type Pt-100 resistance thermometer and an ASL F250 a. c. thermometric bridge. The sensor was calibrated between the Hg and the Ga points, according to the ITS-90. The outer diameter of this thermometer fitted the hole drilled at the base of the lamps, and the thermal contact was aimed by the means of a silicon grease.
• The shunt resistor dipped in a bath of controlled temperature for the measurement of Ish, the current supplied to the tungsten strip of the lamps
A Yokogawa Standard Resistor, type 2792, serial No. 66VW2059 with a measured d. c. value (by the manufacturer) of 10,000 2 mW at 20°C, was employed as a shunt resistor for the determination of the d. c. current supplied to the lamps, connected in series with the lamp’s socket, from the “current” terminals of the resistor.
This resistor was dipped in the oil of a Guildline controlled temperature bath, model 9732VT, set at 28°C. Because to the difference to temperature of the specification of the resistor value, was necessary to carry out a correction to the resistance value to temperature of the bath during the calibrations, according to a chart provided by the manufacturer. The uncertainty of the value of the resistance is ±0,05 mΩ, with assumed k=1.
• The multimeter utilised for the determination of Ish.
A Hewlett-Packard multimeter, model 3458A, provided with an IEEE-488 interface was utilised for the measurement of the current supplied to the shunt resistor connected in series with the tungsten strip of the lamp under study. The calculation of Ish was done by means of the Ohm’s law, from the measured voltage at the resistor terminals.
• The pyrometer
See please paragraph 4.2.1.1
• The multimeter utilised for the measurement of Iph, the photocurrent from the pyrometer
Another Hewlett-Packard multimeter, also model 3458A, was utilised for the measurement of the photocurrent of the pyrometer converted to an analogue voltage at the output terminals marked with the “schreiber” legend in the measurement unit of the instrument. The data of the multimeter, and from the other utilised for the determination of Ish, were acquired with a piece of software developed by CENAM.
• The circulating bath for the water flowing into the lamps bases
A circulating bath, Lauda Brinkmann model RM6, was utilised to supply distilled water to the base of the lamps to maintain its temperature at 20°C ± 0,02°C during the calibrations. The volume capacity of this bath is around 6 litres.
3. Reception of the lamps On March 1st, 1998, CENAM received the lamps set No. II of the NPL (serial numbers C860 and C864) from hands of Mr. C. K. Ma of the National Research Council of Canada (NRC), including an ATA carnet and a pro-forma invoice.
3.1.Measurement of Ramb and Tamb upon reception
As requested in paragraph 3.2.1.1, the measurement of Ramb, the resistance of the lamps elements was performed monitoring the base temperature, utilising the cables supplied with the lamps. The metal case of the lamps was allowed to stabilise to the lab temperature for more hour, and were extrapolated to zero current from 20 mA and 20√2 mA.
• Connections Spade connectors of the supplied cables were connected to the screws at the lamp base from the current terminals of the F18 bridge, and clip alligators were connected from the voltage terminals of the F18 to the small tubes fitted at the base, that are used to supply fresh water.
• Position and grounding
The metal case of the lamps were kept in vertical (normal) position and closed as the wires permitted. A copper wire AWG 12 was connected from the handle of the metallic case to a ground bus installed at the laboratory.
• Results
Values measured for Ramb and Tamb upon the reception of the lamps, were:
Ramb for Lamp C860 = 40,049 milliohms at Tamb of the base = 23,32°C;
Ramb for Lamp C864 = 41,995 milliohms at Tamb of the base = 24,18 °C.
3.2.Measurement of Ramb and Tamb after the calibration of the lamps
• Results
After the lamps were calibrated the values of Ramb and Tamb were the following:
Ramb for Lamp C860 = 40,156 milliohms at Tamb of the base = 23,76°C;
Ramb for Lamp C864 = 41,537 milliohms at Tamb of the base = 24,52 °C.
Figures 1 and 2 show the plots of the measured resistance as function of the base temperature for each lamp.
4.1 Experimental and theoretical procedures
4.1.1 Realisation of the ITS-90 at CENAM
An IKE-LP2 pyrometer, described in brief in 4.2.1.1, and a silver point blackbody manufactured by the National Research Council of Canada, filled with high purity silver, have been the elements to realise the temperature scale as a radiances ratio in the CENAM, as stated in the ITS-90.
From the reference photocurrent, obtained with the radiance of the blackbody at the silver point, a temperature scale is developed from a photocurrents ratio in the pyrometer, that is related to the radiances ratio of the ITS-90, as follows:
R TI T
I Tc Tc TREF
EF REF
EF( , )
( )( )
exp( / )exp( / )
λλλ
= =−
−90 2
2 90
11
(eq. 1)
Corrections for the non linearity of the photodetector for different photocurrents ratios, and for the effective wavelength of the utilised interference filter in the pyrometer, are necessary to apply in order to improve the accuracy of the measurements.
4.1.2 Transference of radiance temperatures to strip lamps
For the lamps of the CENAM, the d. c. current supplied to the strip filament of a standard lamp (previously aged, stabilised and aligned), is adjusted to produce the same photocurrent in the pyrometer, as with the radiance of the blackbody at the silver point (this is reference photocurrent for the selected wavelength). Under this condition, the temperature of the silver point is assigned to the radiance temperature of the lamp at the
effective wavelength of the measurement. Later, a set of different currents supplied to the strip filament is determined for a scale of radiance temperatures of the lamp.
4.2 Report on the results 4.2.1. Local conditions The measurements of Ramb and Tamb requested on paragraph 3.2.1.1 were performed in the Platinum Resistance Laboratory of CENAM, where a F18 thermometric bridge was available. This lab is about 15x15,5x3 cubic meter of volume, and was having some troubles to control its temperature at 22,5°C ± 0,7°C, during the time of the measurements. It has a 15 m long brick wall and a big window, single glass, of the same length in the opposite side.
The calibrations of the lamps were performed at the Pyrometry Lab, that is about 6,3x5x3 cubic meter of volume, with double door to reduce the temperature disturbance when they are open. The walls are masonry built with bricks without windows.
The CENAM facilities are located 23 km far away the city of Queretaro, in the south-east direction, at 1910 m about the sea level. Ambient conditions are given 4.2.1.3.
4.2.1.1 The Reference thermometer
To measure the radiance temperatures of the lamps, was utilised a LP2 pyrometer, manufactured by the IKE of Sttutgart, with serial number 80-32/95.11. The detailed description of this instrument is given in [2]. A short description is given here. Its operation is based on the photocurrent produced in a silicon photodiode by the radiance coming into the optical system of the instrument. The photodiode is inside a thermally controlled box that prevents changes in the sensitivity due to different temperatures of operation. Two wheels holding two sets of filters, allow to select either 650 nm or 912 nm wavelengths, or any of the neutral density filters ND1, ND1,3 or ND2. The instrument also offers the option to transmit the signal free of the action of these filters. When the filters are selected, they are allocated in a position where the incoming beam is collimated. The photocurrent of the detector is converted to a voltage signal by means of a transconductance amplifier built very close to the detector. This voltage is measured in a separated unit of the main body of the pyrometer as if were an absolute photocurrent. The objetive lens of the pyrometer is built with two achromats, for the calibrations were utilised a f200 mm and a f400 mm for a 512 mm/1095 mm target distance range.
• Calibration of the reference photocurrent of the pyrometer The calibration of the reference photocurrent at 650 nm of this pyrometer was performed against a silver point blackbody, whose crucible was manufactured by the National Research Council of Canada from high purity graphite (Ultracarbon America, grade UF-4S) and filled with about 780 g of high purity silver (99,9999 %), Alfa AESAR Stock No. 11357, lot K10E13. The detailed description of the geometry of this blackbody is given in [1]. The opening of its cavity has a diameter of 1,5 mm with an estimated emissivity of 0,99997±0,00003. The analogue signal coming out from the “schreiber” terminals of the measuring unit of the pyrometer, was connected to the input of a digital multimeter described in Section 2 above. On all of the results presented in this report, the photocurrent under darkness conditions is substracted from the measured values.
This darkness photocurrent is measured against a small cavity 12 mm in diameter and 10 mm depth painted with a flat black paint. In Mexico “velvet” finishing paints are not commercially available.
• Results of the calibration of the reference photocurrent For the radiance temperature of the silver point blackbody, it was obtained the
photocurrent, at the effective wavelength λef = 651,82 nm, described below:
IREF = 1,3658 x 10-11 Ampere
TREF = 1 234,93 Kelvin
• Effective wavelength (λλλλe)/local reference wavelength (λλλλr1) The Pyrometry laboratory of CENAM does not have means to measure the effective wavelength of the utilised interference filter. Then, a source to specify this parameter is a table of values provided by the manufacturer, as follows:
Table 1 T1 = T2 + 10K T1 = TAu
T2 / K λ12 / nm λ12 / nm
1000 652,04 651,98
1400 651,90 651,91
1800 651,82 651,87
2200 651,76 651,84
2600 651,72 651,82
3000 651,70 651,81 where λ12 is the mean effective wavelength between the temperatures T1 and T2. The TAu value in the table represents the temperature of the gold point (1337,33 K). In these measurements, the reference temperature is the silver point, not the gold point, then is necessary to determine a value for λ12 for any pair of temperatures T1 and T2 that is in agreement with the following equation:
λ λ12 01 2
1 1= +
+
aT
aT
(eq. 2)
The values that satisfy equation 2 are λ0 =651,527 and a = 0,39964 K. With them, the value of λ12 and those given in the Table 1 are in agreement within ± 0,01 nm
• Aperture ratio; f-number The filament strip of the lamps were located at 63 cm to the objetive lenses of the pyrometer. For this distance, the aperture diameter of the pyrometer as per the specification of the manufacturer is 40 mm, this yields to: f630 mm/40 mm = 15,75
• Target distance As indicated above the target distance between the filament strip and the objetive lenses of the pyrometer was 63 cm.
• Target field dimensions At the distance of 63 cm, and the f200 mm and f400 mm objetive lenses, the measured target field diameter is 0,7 mm.
4.2.1.2 Transfer lamps
• Orientation of the lamps To locate the vertical position of the strip filament, a magnified image of it was produced on a screen and this was compared with the cord of a hanging free plumb bob. The horizontal reference angle was located by m Horizontal spatial radiance distribution
After orientating and performing the photoelectric focusing of the lamp element under study, as indicated in 3.2.1.2, but only for the 650 nm wavelength, the lamp was decentered at both sides of its filament to determine the spatial radiance distribution. The obtained values are given in Tables I and II and plotted in Figures 3 and 4.
Table I
Spatial radiance distribution at the height of the notch for lamp C860 at 650 nm
X (mm) Iph (A) Iph/Imax
21,59 5,903E-11 71,82%
21,72 7,496E-11 91,20%
21,84 8,213E-11 99,93%
21,97 8,217E-11 99,98%
22,10 8,219E-11 100,00%
22,23 8,218E-11 99,99%
22,35 8,219E-11 100,00%
22,48 8,215E-11 99,95%
22,61 8,205E-11 99,83%
22,73 7,957E-11 96,81%
22,86 6,619E-11 80,53%
Table II
Spatial radiance distribution at the height of the notch for lamp C864 at 650 nm
X (mm) Iph (A) Iph/Imax
25,15 4,018E-11 48,29%
25,27 5,930E-11 71,27%
25,40 7,792E-11 93,65%
25,53 8,318E-11 99,98%
25,65 8,319E-11 99,99%
25,78 8,320E-11 100,00%
25,91 8,318E-11 99,98%
26,04 8,318E-11 99,98%
26,16 8,314E-11 99,93%
26,29 8,310E-11 99,88%
26,42 8,181E-11 98,33%
26,54 6,846E-11 82,28%
26,67 4,811E-11 57,82%
Angular distribution of the radiance
After rotating the filament around the vertical axis of each lamp, were obtained the values given in Tables III and IV for the angular distribution at 650 nm, and plotted in Figures 5 and 6:
Table III
Angular distribution of the radiance of the lamp C860 at 650 nm
θ (°) Iph (A) θ (°) Iph (A)
-11,25 8,200E-11 -0,25 8,221E-11
-10,25 8,206E-11 0,75 8,222E-11
-9,25 8,211E-11 1,75 8,222E-11
-8,25 8,214E-11 2,75 8,224E-11
-7,25 8,212E-11 3,75 8,226E-11
-6,25 8,210E-11 4,75 8,230E-11
-5,25 8,211E-11 5,75 8,235E-11
-4,25 8,217E-11 6,75 8,237E-11
-3,25 8,217E-11 7,75 8,220E-11
-2,25 8,219E-11 8,75 8,198E-11
-1,25 8,221E-11
Table IV
Angular distribution of the radiance of lamp C864 at 650 nm
θ (°) Iph (A) θ (°) Iph (A)
-12,67 8,324E-11 -0,67 8,320E-11
-11,67 8,327E-11 0,33 8,319E-11
-10,67 8,330E-11 1,33 8,318E-11
-9,67 8,331E-11 2,33 8,315E-11
-8,67 8,331E-11 3,33 8,313E-11
-7,67 8,331E-11 4,33 8,320E-11
-6,67 8,331E-11 5,33 8,338E-11
-5,67 8,330E-11 6,33 8,349E-11
-4,67 8,329E-11 7,33 8,336E-11
-3,67 8,325E-11 8,33 8,322E-11
-2,67 8,322E-11 9,33 8,319E-11
-1,67 8,319E-11 10,33 8,318E-11
In Figures 5 and 6, the 0° position corresponds to the alignment where the beam of a pilot laser reflects to its origin. This was checked doing the laser’s beam to cross a hole made in a card, to produce a reflected image of something resembling a concentric circle of multiple thin dots around the hole, after striking the filament. An arrow on the figures indicates the position utilised for the calibrations, that was the one obtained for the orientation where the horizontal reference angle is got, i. e., where the notch in the filament and a white dot marked on the rear window are seen in a same line that defines the optical axis of the pyrometer.
For these adjustments, the lamps were set for a radiance temperature of 1100°C, that corresponds to the filament current denoted as I(5).
• Nominal base temperature and its stability Because of a mistake on the utilised parameters to calculate the temperature of the sensor used for the measurement of the base temperature, the adjusted temperature in the circulating bath was not 20°C as indicated during the time of the calibrations. In a check made a posteriori on this sensor, was found that the actual values of those parameters yield to about 1,9°C above the 20°C set at the controller. As the measurements were all finished when the mistake was detected, the only option we found to fix the situation was to report our measurements at the corrected values of the calculated temperature. These are shown in Table V, as follows:
Table V. Corrected values of base temperature
Lamp minimum maximum Remarks
C860 21,87°C 22,23°C 1st run of measurements
C860 21,91°C 22,25° 2nd run of measurements
C864 21,76°C 21,88°C 1st run of measurements
C864 21,71°C 21,81°C 2nd run of measurements
• Total burning time a) Lamp C860, was turned “ON” for a total of 19h 55’, for positioning adjustments and
for performing its calibration. This amount accounts for a total of 148,4 Amperes·hours.
b) Lamp C864, was turned “ON” for a total of 20h 26’, for positioning adjustments and for performing its calibration. This amount accounts for a total of 151,89 Amperes·hours.
4.2.1.3 Ambient conditions
• Tamb The temperature of the Pyrometry Laboratory of CENAM, where the calibrations of the lamps were performed, is controlled to 23°C within ±1°C.
The temperature of the Platinum Resistance Laboratory, where the measurements of Ramb and Tamb were made, was having some troubles to control its temperature. During the measurements, the temperature varied between 20°C and 24°C.
• Relative humidity The relative humidity in these laboratories is not under control and depends on the climate outside, that during the time of the calibrations, was between 22% and 45%, after more than 5 months without rain.
• Maximum and minimum values
Table VI
Description Maximum Minimum
Room temperature 24,0°C 22,1°C
Relative humidity 22,2 % 45,3 %
CSIRO Technical Memorandum TIP-P134
Measurement report on vacuum strip lamps calibrated as intercomparison artefacts as part of the CCT key comparison of local realisations of the ITS-90 between the silver point and
1700oC
Dr Mark Ballico
November 1997
Notes (1) Measurement procedures are as specified in the intercomparison protocol [1] supplied by the pilot
laboratory, NMi (Netherlands) (2) Reporting format in this report is based on the format laid out in section 4 of the intercomparison
protocol [1]. (3) This report is based on measurements performed in September 1997 Artefacts Vacuum strip lamps C564 and C681 supplied by NMi (Netherlands). Both lamps were transported packed in block of foam rubber, inside a hard briefcase. The artefacts were carefully hand carried as cabin baggage.
1. Experimental and theoretical procedures
1.1 Realisation of the ITS-90
1.1.1 Description of equipment The APEP-2 pyrometer was developed at NML during the 1980’s and is described in detail in [2]. A 100 mm diameter compound lens masked to 60 mm is the objective lens, imaging the lamp filament 1:1 onto a 0.6 mm diameter bevelled pinhole with an object distance of 600 mm. Light from the pinhole is collimated and passed through a filter rejecting wavelengths from 850 to 1100nm. The collimated beam then passes through a 50 mm square, 3 cavity interference filter (roughly 10nm bandwidth around 650nm, and with wing transmission lower than 10-5 from UV to IR), before detection by a Hamamatsu S1337-1010BQ large area silicon photo-diode. The signal is amplified by a low noise current amplifier (108V/A) and measured by a HP3458A voltmeter. Measurements are made in a temperature and
2
humidity controlled room (21±0.2oC, RH 50±5%). A rotatable prism is used to reflect radiation from either the reference radiator or the test lamps. The current reference radiator consists of 860g of 99.999% pure gold held in a closed carbon crucible (ultra high purity graphite < 10ppm impurities), with a small, thin walled (1.5mm) graphite tube inserted into the sample. The conical end of the graphite tube has a 60o full angle. Details of the construction of these black-bodies can be found in [3]. The cavity formed by the tube constitutes a very close approximation to a black-body. A double aperture insert (2.0 mm inside diameter) placed inside the crucible at the entrance to the 46.1 mm long, 8 mm I.D. black-body cavity enhances the emissivity. The theory given in [3] for calculation of the emissivity of the black-body fixed point gives an emissivity of 0.99995 , which is included as a correction in the realisation of the temperature scale. The black-body fixed point is held in the centre of a sodium heat-pipe furnace liner (uniformity better than 0.1oC), protected by a silica tube. High purity graphite and alumina annular spacers reduce the heat loss from the front of the fixed point black-body, whilst allowing several mm clearance to the F/10 viewing cone of the APEP2 pyrometer.
1.1.2 Experimental procedures The pyrometer’s linearity, size of source effect and relative spectral response are measured yearly, or whenever a change is made to the pyrometer design. Several melts and freezes of the gold fixed point black-body are made before and after any lamp calibrations to provide a reference against which the lamps are calibrated. These are achieved by cycling the black-body furnace 3-4 oC above and below the nominal freezing point temperature. The melt/freeze duration is normally 30-90 minutes.
1.1.3 Formal derivation of spectral radiance temperature The spectral radiance temperature TR of a strip lamp is defined by solving the following equations [4]:
( )
( ) ( ) ( ) ( )
( ) ( ) ( )
R SOSE L V L V
g s T e T d R s e T d
g T e T e T
L BB
w S S BB BB
w ref S ref S ref R
= +
=
=
∫ ∫
1 ( ) ( )
, , ( ) ,
, , ,
λ ε λ λ λ ε λ λ λ
ε λ λ λ
Where; e(λ,T) is the spectral radiance at λ from a black-body at temperature T (Planck’s law) TBB is the ITS-90 assigned temperature of the fixed point (1064.18oC for gold) εBB is the calculated emissivity of the fixed point black-body (0.99995) λref is the reference wavelength for the pyrometer (650nm) TR is the effective radiance temperature of the strip lamp at the reference wavelength L is the linearity correction function, expressed as polynomial VL is the pyrometer signal when measuring the lamp VBB is the pyrometer signal when viewing the fixed-point black-body SOSE is the size of source effect difference between the lamp and the fixed point black-body s(λ) is the pyrometer’s measured relative spectral response (RSR), g is the transmission of the lamp envelope, normally taken as 0.92, εw(λ,T) is the emissivity of tungsten, bi-linearly interpolated from the values given in reference [5].
1.2 Transfer of radiance temperatures to strip lamps
1.2.1 Description of equipment 1. Positioning: Micrometer translation stages allow reproducible positioning (to 5 µm) and angular
rotation (to 2 minutes of arc) of all sources to be viewed by the pyrometer.
3
2. Current supply: The lamps are supplied by a HP6032A source which is operated as a voltage source. Current stability is achieved using an in-house amplifier using feedback from the difference between the voltage across a current monitoring resistor and that obtained by potentiometric division of a reference voltage source.
3. Current measurement: The lamp current is measured using a 0.01Ω, 100A, circulated oil cooled 4 terminal resistor. The voltage across the resistor is measured using a calibrated HP3458A DVM.
4. Base Temperature: The lamp base temperature is controlled using a flow of cooling water supplied from the laboratory water supply through a heat exchanger coil in a controlled temperature water bath. The lamp temperature is monitored using a 12mm long 4mm OD ceramic encapsulated PRT sensor, inserted into the temperature monitoring hole of the lamp under test. The resistance of the PRTs was measured using a HP3456 DVM in 4 wire resistance mode.
1.2.2 Local conditions
1.2.3 Reference thermometer 1. Effective wavelength: This is not explicitly used in the calibration, as direct integration over the
relative spectral response (RSR) of the pyrometer and the emissivity and tungsten are used. However, the commonly defined “effective wavelength” is provided for reference in Table 7 and Table 8.
2. Local reference wavelength: All spectral radiance temperatures are corrected back to 650.00 nm (in air)
3. Full width at half maximum of the spectral response function: 10.4 nm. 4. Aperture ratio: Objective lens is masked to 60mm, giving f/10 optics 5. Target distance: 600mm 6. Target field dimensions: nominally a 0.6mm diameter disk 7. SOSE: This is measured using blackened disks on an illuminated diffuser [8]. Figure 1 gives the
measured SOSE. 8. SOSE for strip lamp: No “effective source diameter” is explicitly calculated; rather, the measured
SOSE function is appropriately integrated [8] over the nominal dimensions of the target strip. The size of source effect of the fixed point black-body is measured directly using a black-body fixed-point simulator with a view-through hole to outside of the furnace. The computed size of source effect correction for the two lamps used in the present work is given in Table 1.
Table 1: Computed difference in SOSE between the gold fixed-point black-body and two strip lamps, for the APEP-2 pyrometer.
Figure 1: Measured size of source effect for the APEP2 pyrometer.
1.2.4 Transfer lamps 1. Transverse-vertical orientation: The pyrometer target spot was located centrally in the tungsten
strip, adjacent to the small locating notch. 2. Axial location: The lamp filament was brought into focus at the pyrometer aperture by visual
alignment, by viewing the 0.6mm aperture from the detector side using a telescope with a 650nm filter.
3. Rotational and tilt orientation: The lamp orientation was adjusted such that the (roughly 4mm diameter) white paint dot on the rear envelope of the lamp was tangent to the notch in the lamp filament, when viewed through the pyrometer with the lens masked to give a large depth of focus.
4. Nominal base temperature and stability: 21oC within a range ±40mK 5. Total burning time: Each lamp had the burning history given in Table 2 whilst at NML.
5
Table 2: Burning history of lamps used as artefacts in the intercomparison (individual times are ±±±±0.2 hours)
Temperature (deg. C) Duration (hours)1100 (align etc.) 4.00
2.1 Annealing The lamps were annealed for 1 hour at a nominal 650nm radiance temperature of 1700oC. Table 3 gives the change in radiance temperature, ∆T, for a constant filament current giving a radiance temperature of nominally at 1100oC.
2.2 Filament resistance The resistance of the lamp filament was measured using an ASL F-18 AC resistance bridge and a nominally 1Ω reference resistor. A 4 wire resistance technique, as specified in [1] was used. The lamp
Table 3: Change in radiance temperature at a nominal 1100oC radiance temperature, due to 1 hour anneal at 1700oC.
Artefact ∆T Lamp C564 -0.190oC Lamp C681 +0.021oC
6
current terminals were used as the resistance current terminal . The outer brass cooling tubes were used as the resistance voltage terminals, by means of the cylindrical copper terminals supplied with the lamps by NMi. The measurements were made at 50mA and 21/2×50mA, and extrapolated back to zero current, assuming the lamp filament is heated proportionally to the square of the filament current. Table 4 gives the results of the measurements. The uncertainty in the measured base temperature is 5mK at the 95% confidence level, and the uncertainty in the calculated resistance at 0mA is 0.5µΩ (at 95% C.L.).
A polynomial was fitted to the resistivity vs. temperature data for tungsten given in [6] and used to convert the resistances of the “After calibration” data to the “On receipt” temperature, in order to estimate the change in strip resistance due to the effects of calibration. A 95% C.L. uncertainty of 74ppm/oC was estimated for this conversion. Table 5 gives the calculated resistance change and its uncertainty.
2.3 Transverse scans As specified in the intercomparison protocol, the reference pyrometer signal was measured as a function of the transverse position of the strip lamp artefacts, in order to provide confirmation that the pyrometer target size is sufficiently small by comparison with the width of the lamp filament. The results of measurements at a nominal 650nm radiance temperature of 1100oC are presented in Figure 2 and Figure 3. These results indicate that this condition is well satisfied for both lamps, lamp C681 having a 0.73mm uniform region and lamp C564 having a 0.66mm uniform region. The small increase in radiance temperature on the left of C564’s strip is thought to be due to current density peaking at the alignment notch in the lamp filament.
2.4 Rotational scans As specified in the intercomparison protocol, the reference pyrometer signal was measured as a function of the orientation, in the horizontal plane, of the strip lamp artefacts. The results of measurements at a nominal 650nm radiance temperature of 1100oC are presented in Figure 4 and Figure 5. Data is also presented for scans taken with a 10% transmission ND filter between the pyrometer and the lamp, to suppress inter-reflections between them. For lamp C564, no significant difference inter-reflections between the lamp and pyrometer were observed. A broad inter-reflection peak around 0.8o from the specified lamp alignment position was observed. For lamp C681, a significant difference was observed upon interposing the ND filter. Enhancements in radiance due to inter-reflection between the lamp and the pyrometer were observed on either side of specified alignment position. Further examination clearly showed one to be caused by reflection between the pyrometer lens and the lamp envelope, and the other by reflection between the pyrometer lens and the lamp filament. These could be clearly identified by passing a light beam back through the pyrometer optics, and examining the size of the cone of light reflected by the lamp’s strip and envelope. At the specified orientation of the lamp, both of these reflections of the pyrometer’s objective lens by the lamp fell outside the objective aperture, and so should not contribute to the
Table 4: Filament resistance measurements on strip lamps C681 and C564, on receipt of the lamps and just before leaving the laboratory.
Artefact Temperature R0mA R50mA-R0mA Lamp C564 On receipt 22.705oC 40.20021 mΩ 7.48 µΩ “ After calibration 22.769oC 40.20744 mΩ 7.44 µΩ Lamp C681 On receipt 22.782oC 34.31369 mΩ 3.62 µΩ “ After calibration 22.349oC 34.24730 mΩ 3.68 µΩ
Table 5: Change in filament resistance at ca. 23oC over the calibration period at CSIRO.
measured lamp irradiance. The ND filter data indicate a broad inter-reflection peak, intrinsic to the lamp, centred roughly 1.5o from the specified lamp alignment position.
2.5 Calibration Two consecutive calibrations were performed on each lamp, under the conditions specified in [1]. The time taken for the lamps to stabilise at each temperature is given in Table 2. Table 7 and Table 8 give the calibration results for lamps C564 and C681 respectively, in the format specified in [1]. Table 6 gives the definitions of the terms used in these calibration tables.
Both lamps appear to have decreased in radiance temperature by roughly 0.1oC between the two calibrations, which may indicate that the lamps have not been fully stabilised upon receipt by CSIRO
Table 6: Definition of terms used in the calibration results tables.
Term Definition I Index to calibration point (as specified in [1]) R Measured photo-current ratio to the gold point black-body Tλ Lamp radiance temperature at λeff Tλ(SOS) “ “ corrected for SOSE between fixed point and lamp Tλ(SOS,LIN) “ “ also corrected for pyrometer linearity Iset Actual lamp current Iref Lamp current specified in [1] λeff Effective wavelength of the pyrometer T(λref,21C) Lamp radiance temperature corrected to λref (=650nm) dT/dTB Lamp base temperature coefficient (given in [1]) dT/dI Sensitivity of lamp radiance temperature to lamp current
(from a polynomial fitted to calibration data) T(λref,20C,Iref) Lamp radiance temperature at the reference conditions.
Figure 2: Normalised radiance signal from the pyrometer as a function of the transverse position of lamp C564. (The dashed curve is the same data on a expanded scale)
Figure 3: Normalised radiance signal from the pyrometer as a function of the transverse position of lamp C681. (The dashed curve is the same data on a expanded scale)
9
-15 -10 -5 0 5 100.992
0.993
0.994
0.995
0.996
0.997
0.998
0.999
1.000
1.001
1.002
lamp alignment mark
No filter With ND filter
Nor
mali
sed
radi
ance
Rotation angle (degrees)
Figure 5: Normalised radiance signal from the pyrometer as a function of the orientation in the horizontal plane of lamp C681 (angles are clockwise viewed from above).
-10 -8 -6 -4 -2 0 2 4 6 8 10
0.996
0.997
0.998
0.999
1.000
1.001
lamp alignment mark
no filter with ND filter
Nor
mali
sed
radi
ance
Rotation angle (degrees)
Figure 4: Normalised radiance signal from the pyrometer as a function of the orientation in the horizontal plane of lamp C564 (angles are clockwise viewed from above).
10
Table 7: Two calibrations of lamp C564, showing (A) corrections for size of source effect and linearity, assuming it as a black-body radiator and (B) corrections to the reference conditions 650nm, 20oC base temperature and specified lamp currents (lower table).
Table 8: Two calibrations of lamp C681, showing (A) corrections for size of source effect and linearity, assuming it as a black-body radiator and (B) corrections to the reference conditions 650nm, 20oC base temperature and specified lamp currents (lower table).
3. Uncertainties The discussion of uncertainties is broken into two sections: Firstly a description of the physical parameters that have an effect on the calibration. Secondly, a table of the numerical values (Table 9) of these uncertainties, combined in accordance with the ISO guide to the expression of uncertainty in measurement [7]. Note: Details of the assumptions and mathematical expressions used to convert these parameters into equivalent uncertainties in temperature are given in [8].
3.1.1 Reference black-body radiator 1. Emissivity: Uncertainty due to physical dimensions of the black-body, the emissivity of the surface
material, and axial gradients in the crucible. 2. Purity: Estimated from the melt and freeze data using the well known technique of plotting the
temperature vs. 1/F where F is the fraction of metal melted. 3. Gradients: Conduction losses through the wall of the black-body due to radiative heat losses. 4. Reproduceability/other: Estimate of other systematic errors, determined by a variety of changes to
the system, such as purge gas rate, freeze/melt duration, furnace balance etc. .
3.1.2 Detector 1. Short term drift: The measured stability of the sensitivity of the pyrometer over a 12 hour period.
3.1.3 Linearity 1. Random: the uncertainty in the measured linearity, resulting from the detector noise 2. Systematic: The estimate of the inherent uncertainty of a doubling step of the doubler apparatus
used to measure the linearity of the detector+amplifier+DVM.
3.1.4 Size of source effect 1. BB SOSE: The uncertainty in the SOSE of the fixed-point black-body 2. Systematics: The estimated systematic errors inherent in the SOSE measurement apparatus. 3. Integration: Uncertainty in the integrated SOSE for the strip, resulting from the numerical
integration technique used. 4. Strip width: Uncertainty in the SOSE due to uncertainty in the measurement of the width of the lamp
filament 5. Curve fit: Uncertainty in the SOSE resulting from the deviation of the fitted curve to the measured
SOSE points.
3.1.5 Spectral parameters 1. Calibration: The uncertainty in the wavelength calibration of the monochromator. 2. Leakage: The estimated out-of-band sensitivity of the pyrometer 3. Temperature coefficient: Uncertainty due to known temperature coefficient of the interference filter
and the long term (over months) pyrometer temperature stability 4. Reproducibility: The type-A uncertainty derived from repeated measurements of the effective
wavelength of the pyrometer. 5. Stability: The measured long term stability of the interference filter
3.1.6 Lamp 1. Base temp: Uncertainty of the measurement of the lamp base temperature. 2. Horiz. posn. : Uncertainty arising from the ability to set the lamp to the specified transverse
position. 3. Rotation: Uncertainty arising from the ability to set the lamp to the specified angular position.
13
4. Supply stability: Short term current stability of the lamp supply. 5. Thermal EMFs: A 0.01 Ω resistor was used as a current shunt, thus contribution from stray thermo-
voltages may affect the current measurement. 6. DVM calibration: The DVM is calibrated yearly, and this component takes account of drifts over
this period. 7. Shunt calibration: The calibration uncertainty of the shunt resistor used to measure the lamp current. 8. Window transmittance: Changes in transmission due to non-reproduceability of the cleaning of the
lamp window. No account is made for changes in the window transmittance during the calibration due to effects such as deposition of tungsten etc.)
9. Prism reflectance: Change in pyrometer sensitivity between the prism position for viewing the fixed point and that for viewing the lamp.
10. dT/dλ: Reference wavelength: An estimate of 3% in dTR/dλ, and a maximum shift of 0.5nm is allowed for in the error budget in converting to the reference wavelength
3.1.7 Other parameters Several other parameters, for which the uncertainty contribution was considered negligible were: 1. DVM resolution for lamp current measurement. 2. DVM resolution for photo-current measurement. 3. Detector noise: Signals were averaged over a 5 minute period, resulting in negligible contribution to
the uncertainty from the noise in the detector system. 4. Neutral Density filters: Not applicable, since we use the pyrometer to step from the gold point in a
single step. 5. Correction of RH: As absorption due to water vapour is negligible at 650nm, no explicit correction
is included.
14
Table 9: Summary of the numerical values of the uncertainty sub-components in the calibration of strip lamps at CSIRO, and their combination into an overall uncertainty estimate.
leakage 0.3 1.7E-01 ppm B 2 4 0 4 6 6temp. coeff. 0.3 0.2 K B 8 0 0 1 2 2reproducability 0.010 nm A 8 1 0 4 9 14stability 0.01 0.006 nm B 2 1 0 2 5 8
Ref. BB emissivity 5.10E-05 2.9E-05 ratio B 8 2 2 3 4 5purity 10 5.8 mK B 2 5 6 7 8 9wall gradients 1 0.6 mK B 2 1 1 1 1 1reprod/other 4 2.3 mK B 2 2 2 3 3 3
Detect. short term drift 1.00E-04 5.8E-05 ratio A 2 4 5 6 8 10
SOSE BB SOSE 2.0E-05 ratio B 2 1 2 2 3 4systematics 1.0E-05 ratio B 2 1 1 1 1 2integration 1.0E-05 ratio B 2 1 1 1 1 2strip width 1.0E-05 ratio A 2 1 1 1 1 2curve fit 1.0E-05 ratio A 8 1 1 1 1 2
Linearity random 3.0 ppm A 8 0 0 1 1 1systematic 6.0 ppm/step B 2 1 0 2 5 8
Lamp base temp. 40 23.1 mK A 40 2 2 1 0 0horiz. posn. 8.0E-05 4.6E-05 ratio A 2 3 4 5 7 8rotation 2.0E-05 1.2E-05 ratio A 2 1 1 1 2 2supply stab. 3.0 1.7 ppm B 40 1 1 1 1 2thermal EMFs 0.5 0.3 uV B 8 5 4 3 3 3DVM calib. 15 8.7 ppm B 40 6 6 6 7 9shunt calib. 2 1.2 ppm B 40 1 1 1 1 1window trans. 1.00E-04 5.8E-05 ratio A 2 4 5 6 8 10prism refl. 3.00E-05 1.7E-05 ratio B 2 1 1 2 2 3dT/dλ 3.00E-02 1.7E-02 rel. err. B 8 0 0 1 1 1
1. References: [1] “Protocol to the comparison of local realisations of the ITS-90 between the silver point and 1700oC using tungsten strip lamps as transfer standards”, document supplied by NMi (June 1997) [2] “A Precision Photoelectric Pyrometer for the realisation for IPTS-68 above 1064.43oC”, T. P. Jones and J. Tapping: Metrologia 18, pp.23-31 (1982) [3] “The Realization of the IPTS-68 above 1064.43 using the NSL Photoelectric Pyrometer”, T.P. Jones and J. Tapping: Metologia, Vol,8, No.1, pp.4-11(1972) [4] “Recommissioning the NML high precision pyrometer APEP-2”, M. Ballico, Proceedings of 2nd Biennial Conference of the MSA, 1997, Melbourne, Australia [5] “Monochromatic emissivity of tungsten in the temperature range 1200-2600oK and in the wavelength range 0.4-4µm”, Latyev, L.N. et. al., High Temperatures - High Pressures, Vol.2, pp.175-181, 1970 [6] “The Electrical Engineering Handbook”, Editor R. C. Dorf, 1993, CRC Press, London [7] “Guide to the Expression of Uncertainties in Measurement”, International organisation for Standardisation, 1993, ISBN 92-67-10188-9 [8] “Independent Australian realisation of the International temperature scale of 1990 using the recommissioned APEP2 pyrometer”, M. Ballico, CSIRO technical memorandum P-48, November 1997.
Key comparison lamps - IMGC
- 1 -
COMPARISON OF LOCAL REALIZATIONS OF THE ITS-90 BETWEEN THE SILVER POINT AND 1700 °C USING VACUUM TUNGSTEN-STRIP LAMPS AS TRANSFER STANDARDS
Report on measurements performed at IMGC
by T. Ricolfi and M. Battuello
1. Local realization of the ITS-90 1.1 Description of the equipment 1.1.1 Reference thermometers Two different thermometers have been used for this exercise: one working at 655 nm and the other one working at 950 nm. The technical data of these thermometers are reported in Table 1. Table 1. Technical data of the reference thermometers
Item
Thermometer at 655 nm
Thermometer at 950 nm
Focal length of objective 200 mm 200 mm Target distance 600 mm 475 mm Target size 1 mm 1.1 mm Aperture ratio 12:1 11:1 Interference filter - centre wavelength - half width
655 nm 11 nm
950 nm 13.3 nm
Detector Si (Hamamatsu S2386-5K) Si (Hamamatsu S1336-44BQ)Temperature of detector 18 °C 28 °C Amplifier gains 5 gains from 106 to 1010 V/A 3 gains from 5x106 to
5x108 V/A
References in the literature for these thermometers are found in [1] and [2]. 1.1.2 Fixed-point blackbody The silver point is used as reference temperature for realizing the ITS-90. A schematic diagram of the crucible and blackbody assembly is shown in Fig. 1. The metal ingot (about 446 g) is contained into a pure graphite crucible whose available volume is about 48 cm3. The silver sample was supplied by Cominco and its nominal purity is 99.9999%. The blackbody cavity is a cylinder 9.5 mm in diameter and 66 mm in depth which is terminated with a conical bottom. The cavity aperture is delimited by a platinum diaphragm 3 mm in diameter. The effective emissivity of the cavity has been estimated to be 0.99994±0.00001.
Key comparison lamps - IMGC
- 2 -
1.2 Characterization of the reference thermometers 1.2.1 Spectral responsivity The spectral responsivity of the reference thermometers is measured using a monochromator of 500 mm focal length and the experimental setup shown in Fig. 2. The measurement procedure consists of the following operations: (a) Wavelength calibration of the monochromator. This is performed using one or more spectral lamps as radiation sources and a silicon photodiode as detector to localize the spectral lines that are focused onto a circular diaphragm past the exit slit of the monochromator. (b) Measurement of the input spectral curve to the thermometer. A tungsten-strip lamp in front of the entrance slit and a neutral pyroelectric detector with gold-black coating flush with the circular diaphragm are used for this measurement. (c) Measurement of the output spectral curve from the thermometer. After removing the pyroelectric detector, the thermometer is aimed at the diaphragm as shown in Fig. 4. The same spectral distribution previously measured with the neutral detector is then used as input curve to the thermometer and the corresponding output curve is measured. (d) Calculation of the spectral responsivity. The relative spectral responsivity of the thermometer is calculated as the ratio of the measured input and output curves. The standard uncertainty in the effective wavelengths calculated from the measured responsivity curves has been estimated to be 0.05 nm at 655 nm and 0.1 nm at 950 nm. The limiting effective wavelength of the two thermometers has been found to obey to the well known relationship 1/λ = a + b/T where T is expressed in kelvin, λ in nanometers, and a = 1.5272774 x 10-3 b = -7.540588 x 10-4 for the thermometer at 655 nm, and a = 1.053157 x 10-3 b = -3.530246 x 10-4 for the thermometer at 950 nm. 1.2.2 Non-linearity The non-linearity of the signal of the reference thermometers was checked using a flux doubling technique with two lamps and a beam splitter. No non-linearity was found up to
Key comparison lamps - IMGC
- 3 -
the upper limit of the comparison, that is 1700 °C. The standard uncertainty in the non-linearity measurements was estimated to be 1x10-4. 1.2.3 Size-of-source effect The SSE was measured using an integrating sphere as radiance source according to the scheme shown in Fig. 3. To simulate a tungsten strip lamp, measurements were also performed by replacing the central black spot with a black strip 1.5 mm wide. The results are shown in Fig. 4 and Fig. 5. 1.2.4 Gain ratios To cover the interval from the silver point to 1700 °C, the following amplifier gains are used with the two thermometers: - 1010, 109 and 108 V/A for the 655 nm thermometer - 5x108, 5x107 and 5x106 V/A for the 950 nm thermometer The ratios between two adjacent gains were measured using a lamp as radiant source. The results are shown in Table 2. Table 2. Measured gain ratios
Gain ratio
Measured ratio
Thermometer at 655 nm Thermometer at 950 nm 1010/109 9.13711 109/108 11.05242
5x108/5x107 9.98486 5x107/5x106 10.02164
1.3 Procedure for setting up the ITS-90 The scale is established onto a tungsten strip lamp using the thermometers as photoelectric comparators. The steps for setting up the scale are as follows: 1. The thermometer is calibrated at the silver point. Only freezing plateaus are considered in
this operation. The temperature distribution around the blackbody aperture is measured during the freezing plateau.
2. The fixed-point calibration is transferred immediately to a reference lamp and then from this to the lamp under calibration.
3. The lamp is successively brought to temperatures above the silver point. At each temperature the signal ratio between the lamp and the reference lamp maintained at the silver point is measured.
4. After adjusting the signal ratios according to Table 2, the lamp temperatures are calculated using the defining equation of the ITS-90.
Corrections for the SSE are computed as follows:
Key comparison lamps - IMGC
- 4 -
(a) The SSE contribution in the fixed-point furnace is calculated using the curves in Fig. 4 or Fig. 5 relative to the black spot of 1.5 mm and the temperature distributions measured during freezing. Let this contribution be SSEf.
(b) The difference between the two curves in Fig. 4 or Fig. 5 is calculated in correspondence to a source diameter equal to the length of the filament of the lamp. Let ∆SSE be this difference. Note: this procedure implies the assumption of uniform temperature on the lamp filament.
(c) The correction in terms of thermometer signal at the silver point is given by SSEf - ∆SSE. This correction can be converted to temperature through the sensitivity of the thermometer.
(d) The temperature correction at a higher temperature T is given by the product of the correction at TAg by (T/TAg)2.
2. Execution of the intercomparison exercise 2.1 Preliminary operations 2.1.1 Measurement of the room temperature resistance Ramb The room temperature resistance of the lamp element was measured using a AΣL F18 resistance bridge. A reference resistor of 1 Ω was used. 2.1.2 Positioning and checking The guidelines of the protocol were strictly followed for alignment and focussing. Horizontal and angular radiance distributions were measured. 2.1.2 Restabilization of the lamps The restabilization was done according to the protocol guidelines. The thermometer at 655 nm was used. 2.2 Calibration of the lamps The procedure described in Section 1.3 was adopted for calibrating the transfer lamps of the comparison. The calibration at 655 nm and 950 nm was done in parallel, ie, at each current value of the transfer lamp measurements were done in sequence with the two thermometers. For each lamp and for each wavelength two complete runs were done. Note. During the calibration it was noticed that the room light generated a stray radiation component at 655 nm. Measurements at this wavelength were then done in the dark. Similarly, it was noticed that at high temperatures the transfer lamp disturbed the readings on the reference lamp. This problem was overcome by screening the thermometers when observing the reference lamp. 2.3 Total burning time The total burning time was 26 h and 15 minutes for lamp 1 and 24 h and 45 minutes for lamp 2.
Key comparison lamps - IMGC
- 5 -
3. Presentation of results 3.1 Preliminary measurements 3.1.1 Measurements of Ramb The results are reported in Table 3. Table 3 Results relative to Ramb
Lamp Before calibration Ramb (ΩΩΩΩ) tamb(°C)
After calibration Ramb (ΩΩΩΩ) tamb(°C)
1 (C860) 2 (C864)
0.0400998 23.01 0.0418553 22.97
0.0400825 23.00 0.0418636 23.03
3.1.2 Angular distribution of radiance The results of this measurement are reported in Fig. 6 and Fig. 7. 3.1.3 Restabilization The results are shown in Table 4. Table 4 Results of restabilization
Lamp Drift at I(5) after restabilization (°C)
Lamp 1 Lamp 2
-0.022 -0.036
3.2 Results of calibration The results of calibration have been arranged according to the protocol instructions. Those relative to lamp 1 are reported in Tables 5 and 6 and those relative to lamp 2 in Tables 7 and 8. It is to be noted that no corrections for non-linearity have been made. Also the reduction of data at 950 nm to RH =0% has not been applied for lack of information about the effect of humidity on the thermometer.
Key comparison lamps - IMGC
- 6 -
Table 5.A Measurement results at 650 nm for lamp 1 (C860)
4. Uncertainties The uncertainties in the final radiance temperature Tλ originate from the combination of the uncertainties in the measured Tλ(λe;Tb) and the uncertainties in the corrections. A map of the various uncertainty sources is shown in Fig. 8. The standard uncertainty estimates at the various calibration points are reported in Tables 9 and Table 10. Because the uncertainties were the same for the two lamps, only those referring to lamp 1 have been reported. The labels in the tables should be read as follows: - s (TAg): uncertainty in the fixed-point calibration - s (λe): uncertainty in the effective wavelength - s(R): uncertainty in the signal ratio - s(x;y;θ;ϕ): uncertainty in positioning - s(SSE): uncertainty in the size-of-source effect correction - s[Tλ(λ)]: uncertainty in the wavelength correction - s[Tλ(Tb)]: uncertainty in the base temperature correction - s[Tλ(I)]: uncertainty in the current correction
References 1. T. Ricolfi and F. Girard: "Precision radiation thermometer for the realization of the ITS-
90 above 962 °C". In: Proc. TEMPBEIJING '97, B. Zhang, L. Han, X. Zhao (eds.). Standards Press of China, Beijing (1997), pp. 55-60
2. M. Battuello, T. Ricolfi and L. Wang: "Realization of the ITS-90 above 962 °C with a photodiode-array radiation thermometer". Metrologia 32, 371-378 (1995/96)
27/06/2003 1 rap_comp_inter3_INM.doc
Bureau National de Métrologie / Institut National de Métrologie BNM / INM
CNAM, 292, rue Saint Martin 75003 Paris Bernard Rougié rougie@cnam/fr tel : 01 40 27 20 22
Georges Bonnier bonnier@cnam/fr tel : 01 40 27 21 58 Georges Negro negro@cnam/fr tel : 01 40 27 25 96
Fax 01 42 71 37 36
CCT : Radiance temperature comparison results Two transfer lamps N° 860 and 864 from the CCT temperature key comparison loop N°2 have been
measured in BNM-INM in September 1998. They came from NPL and were sent to IMGC. This report fits as well as possible with the proposed design report. Description of measurement apparatus, procedure, uncertainty and radiance temperature results are given.
1. Experimental and theoretical procedures 1.1. Realisation of ITS90
1.1.1. Description of the apparatus
1.1.1.1. Reference black body radiator The reference is a copper fixed point black body. The gold point black body we intended to use failed
during CCT lamps measurements.
The whole radiant part is graphite made. The diaphragm cavity diameter is 8 mm and the length 70 mm. It is included in an horizontal quartz cell opened on the observing side. An Helium gas flow prevents oxidation and reduces optical move of black body’s image.
The heating device is made of 18 ‘Kanthal’ elements. They are separated in 3 zones regulated by three independent regulators. The temperature uniformity is checked to be better than 1 K along the crucible.
The crucible is prepared in liquid phase on a high purity metal filling bench. It contains 800 g of high purity copper(99.995%). After operation, copper and graphite filling pieces samples are picked up for impurity analyse.
1.1.2. Procedure The black body is previously stabilised at a 2°C temperature under or above the melting or freezing point.
Then the regulator set point is changed for 4 °C up or down so that the temperature pass beyond the melting or freezing point. The radiance temperature sampling interval time is 2 mn. To observe the freezing point we must give a set point 15°C under the freezing temperature up to the beginning of the plateau to overrun the copper over cooling effect.
The plateau is defined as the time range where measurements are not more different than 0.05 K away from the average. The temperature plateau uncertainty is 0.25 K.
1.1.3. Spectral radiance temperature This temperature is the black body cavity internal wall one, it will be given as ‘reference temperature’ in
the results table.
1.1.3.1. Temperature difference due to the wall of the cavity The temperature gradient through the graphite thickness as been computed taking in account the radiant
power loss and the graphite wall thermal conductivity. The gradient is very low : 0.004 K. Its uncertainty will be neglected. The value T(FP) is the ITS-90 copper freezing point minus the gradient value.
27/06/2003 2 rap_comp_inter3_INM.doc
1.1.3.2. Emissivity correction The emissivity has been computed by Monte Carlo method : Its value is 0.9994 with an uncertainty of
0.0002.
1.1.3.3. Impurities The copper impurity analyse shows among 20 tested metals a rate no larger than 5 ppm except a 10 ppm
rate for silver. The corresponding change in freezing temperature will be neglected.
1.2. Transfer of radiance temperatures to strip lamps
1.2.1. Thermometer The light sources (reference fixed point black body, lamps, alignment laser and spectral line lamps) are
placed on a fix bench.
The measurement device, pyrometer, is a complicated one with a two mirrors optical entrance, a monochromator and detectors, placed on a single moving stage allowing a 1.6 m range of movement. Additional functions such as linearity measurement, CCD camera for sources alignment and internal alignment laser are available.
The optical entrance is constituted by a concave and a convex gold coated mirror. The image magnification is unity, the focusing distance of the main concave mirror is 1 meter and the frontal distance from source to aperture stop is 500 mm.
The aperture and field stops are respectively 45 mm ( *or 25 mm) and 0.5 mm circular diaphragms. So the geometrical extend is defined by a 0.006 sr solid angle and a 0.25 mm² field area ( *: Solid angle is 0.0019 sr for a few part of our first measurements).
The spectral selection is achieved with a monochromator. It is a Czerny-Turner type with 0.5 focal distance and 1200gr/mm grating. The slit function has a symmetrical trapezoidal shape (middle height width 3.2 nm).
The detector is a silicon photo diode 1337 Hamamatsu type whose output current is collected by a trans-impedance amplifier. The value of impedance is 109 or 1010. Its linearity has been checked by the linearity measurement device included in the optical entrance arrangement.
All the elements, optical entrance, monochromator and detector are taken as a single device we call here pyrometer.
1.2.2. Procedure The temperature measurement of lamps is made by comparing lamps with a copper fixed point black
body. The gold point black body we intended to use failed during CCT lamps measurements. This defect has increased the lamps burning time.
We can consider that lamps are directly compared to the black body, but with a roughly 10 hours comparing time. The stability of the pyrometer is good enough to assume that there is no change in the response pyrometer for 10 hours. Furthermore the freezing and melting plateau have been previously checked not to be more different than 25 mK. According to this, we have made the pyrometer response measurement, said “ pyrometer calibration ”, in front of the fixed point black body, two times a day : The first one in the morning by observing the melting plateau and the second one in the evening with the freezing plateau.
The relative change in response along the day was always lower than 5.10-4. A linear interpolation has been done to compute the equivalent response for each lamp measurement time.
The pyrometer calibration has been made only for the 650 nm wavelength. Along the day the ratio of 950 nm and 650 nm pyrometer response with the black body has been measured so that we could compute the equivalent response at 950 nm as well as 650 nm.
The two lamps of the CCT comparison have been measured together at both wavelengths by the pyrometer for a period between the ‘morning’ and ‘evening’ pyrometer calibrations. A local lamp has been regularly included in the measurement cycle to give an information on pyrometer stability.
27/06/2003 3 rap_comp_inter3_INM.doc
The whole calibration procedure (§ 3.2.2) has been exactly executed. The valid temperature measurements have been done in five runs, each one including current number (1 to 5) or (5 to 11) and one (5,7,9,11 and 11) for particular check.
We didn’t make the radiance temperature measurement at the 950 nm wavelength for the current numbers 10 and 11. In our procedure this measurement generates a range of photo current larger than 1 to 1000 that we can not accept for several reasons. We didn’t make the additional run which was necessary and would have added a delay in the comparison schedule.
2. Results and uncertainties description 2.1. Reference black body radiator
The internal wall radiance temperature of the black body is 1357.766 K. It appears in the results as ‘reference black body radiator (T[FP])’. The emissivity is taken in account by dividing the measured black body flux by its value (0.9996). This correction, included in the photo current ratio, does not appear in results tables. Its uncertainty is 0.0002.
2.2. Reference thermometer
2.2.1. Ratio of photo current The interval time separating lamp and black body photo currents measurements is about 5 hours. We
adopt as photo current ratio’s uncertainty the relative pyrometer response drift during this time. Although we have made a linear interpolation of pyrometer sensitivity we choose its drift, 0.015%, as the 1/3rd of difference between ‘morning’ and ‘evening’ pyrometer calibration.
2.2.2. Source size effect The device designed for this correction did not operate as we hoped at 650 nm and 950 nm. Hence, the
measurement has been made with a device lent for a few days by a sphere manufacturer. We can not do measurements for various radiuses of the sphere but the sphere radius , r=19 mm, corresponds to the hottest part of the black body. The radiance homogeneity is checked to be better than 2% on its entire area. A 5 mm length and r0=2.5 mm radius graphite cavity is placed in the centre of the sphere aperture and SSE value is computed as the flux in the black centre target divided by the flux on an other part of the sphere aperture.
Wavelength SSE(2.5 mm, 19 mm)
650 nm 1.8E-03
950 nm 2.0E-03
Due to the large difference between the black body and the lamp radiation area, we consider that the correction applied can only take in account the black body size of source effect. The uncertainty related to this effect is enlarged by 1/3rd of the correction value itself because of the absence of lamp effect measurement.
2.2.3. Spectral parameters
2.2.3.1. Wavelength selection The central wavelength value of the monochromator spectral transfer function has been calibrated by the
mean of several spectral lines of Mercury and Rubidium. The wavelengths are given in normal air condition. The reproducibility is 0.005 nm but the non linearity of wavelength versus mechanical translation increases uncertainty up to 0.02 nm.
Important note : Due to a mistake in the reading of the wavelength calibration table, the 900 nm value has been chosen in place of the 950 nm one. All the measurements have been done at the 900 nm wavelength.
Because of spectral width the effective wavelength is not exactly equal to the central wavelength, but it does not differ from the central one more than 0.02 nm for all temperatures.
27/06/2003 4 rap_comp_inter3_INM.doc
2.2.3.2. Blocking The transmittance, outside the transfer function, is not exactly null. Hence, the flux is affected for a small
part by the entire spectrum of the source. By using band pass and high pass filters we have evaluated the relative parasite part of flux. This has been done only for the black body. The blocking effect is 4.10-3 for 650 nm. The blocking effect is ten times lower for 950 nm and will be neglected. We estimate that our measurement can not lead to a correction but gives the uncertainty related to this effect.
Assuming that the parasite transmittance is constant in the active spectral range, we have computed the part of flux coming from outside the central transfer function.
Dif T Pl T Se d( ) ( , ) ( )= ⋅ ⋅ ⋅∫α λ λ λ600
1050
Dif(T) : is the parasite part of flux with T the real temperature of the source.
Pl(λ,T) : Planck law
Se(λ) : global spectral sensitivity of pyrometer
α : Coefficient computed for Dif(T) to be equal to the one measured with the black body.
This assumption gives a weight too large to the wavelengths far from 650 nm and maximises the corresponding error.
The relative photo current ratio change due to unperfected blocking is equal to Dif(Tblack-body)-Dif(Tlampe). We give computation result and we consider it as the uncertainty due to this effect.
Temperature 960°C 1300°C 1700°C
Flux ratio uncertainty
4.10-4 9.10-4 1.210-3
Temperature uncertainty
0.026 0.102 0.220
2.3. Transfer lamps
2.3.1. Lamp current correction The current in the lamps is given by measuring the voltage on a resistor placed in the circuit . It is
regularly set at a value very close to the one of current table. The remaining correction is made by using a first order polynomial evaluation of dTλ(I)/dI.
The uncertainty of current measurement is 5.10-5.
2.3.2. Base temperature correction These corrections have been made according to the procedure. They are very low. Their uncertainty is not
very well established. We estimate it to 0.2 °C for ‘Tbase’ giving rise to 0.01 °C for radiance temperature in the worse condition.
2.3.3. Residual parameters
2.3.3.1. Light sources position Various scans of the horizontal transverse radiance of the ribbon have been done at several positions of
the lamps along the optical axis and several height levels up and down around the notch. A scan is a set of eleven values acquired within 1.5 mm.
2.3.3.1.1. Focusing The focusing position along the optical axis has been derived from these scans. The position sensitivity is
about 0.5% of the distance from the lamp to the focusing mirror. The width of the plateau is 0.6 mm for the best position.
27/06/2003 5 rap_comp_inter3_INM.doc
2.3.3.1.2. Orientation The temperature variation versus angle around vertical axis shows a very sensitive change for the 864
lamp. The slope is 0.86 K/°. This value leads to a very important uncertainty and we wish it could result from a wrong measurement. It would be interesting to compare this result to the one which would have been done by other laboratories.
2.3.3.1.3. Horizontal and vertical transverse position We have evaluated the horizontal position as the centre of the graph of a scan. This one fits with the value
given by the CCD camera better than 0.1 mm.
The scan graphs done at different heights around the notch show a loss of signal by the same height and side of the notch as we thought it would be..
The slope of the radiance has been evaluated around the central point :
The alignment uncertainties are evaluated to 0.1 mm for position and 0.2° for orientation from which we can derive radiance and temperature uncertainties by using the position and orientation gradients.
Final uncertainty L860 0.064 0.079 0.134 0.123 0.259 0.193
Final uncertainty L864 0.126 0.170 0.222 0.274 0.380 0.430
Conclusion The reproducibility of our measurements, not reported here, is five times lower than global uncertainty
and influences any values. Both most important parameters are blocking effect and size of source effect. The precise estimation of blocking effect is very difficult and we have, as a precaution, increase its uncertainty value. The size of source effect has not been estimated as well as we hoped to do because our device measurement do not operate correctly yet. Hence in this case the uncertainty is larger than the one we could normally obtain with the apparatus to come..
We wait for the other laboratories report concerning the lamp 864 orientation temperature sensitivity which seems to exceed the normal values. We hope to reduce uncertainty part due to this effect which is bigger than all other ones.
27/06/2003 8 rap_comp_inter3_INM.doc
Annexes
Table A : Lamp C860; lambda = 650 nm Number of current
Lamp current as defined in appendix B
Lamp current ratio i(lamp)/i(T[FP])
ref temp of black body T(FP)
Lamp temperature Temp corrected of source size effect
3.1. Local conditions ...................................................................................................................................8
3.1.2. Transfer lamps ................................................................................................................................9
Report on Calibration of Vacuum Tungsten-Strip Lamps as Transfer Standards for the Inter-comparison of
the Local Realizations of the ITS-90 between the Silver Point and 1700 °C
28 December 1998 21
Table 7. Lamp currents corresponding to the copper point temperature.
Lamp T(FP) I(FP)
without SOS correction
I(FP)
with SOS correction
C564 1084.62 °C 5.30767 A 5.30720 A
C681 1084.62 °C 6.57615 A 6.57472 A
Table 8. Resistance of the lamps before and after the measurements, which were measured by a
ASL F18 Bridge with 50 root 2 mA current, 3 Hz bandwidth and a 1 ΩΩΩΩ reference resistor.
Measurement just after reception of the lamps (1997/10/15),
Resistance Ambient Temp
Lamp 1 (C564) 40.5384 mΩ 24.54 °C
Lamp 2 (C681) 34.7352 mΩ 25.21 °C
Measurement just before transferring lamps to NIM (1997/12/8),
Resistance Ambient Temp
Lamp 1 (C564) 40.4865 mΩ 24.59 °C
Lamp 2 (C681) 34.6972 mΩ 25.39 °C
Report on Calibration of Vacuum Tungsten-Strip Lamps as Transfer Standards for the Inter-comparison of
the Local Realizations of the ITS-90 between the Silver Point and 1700 °C
28 December 1998 22
4. Uncertainties
Table 9. The budgets of uncertainties for the calibration of the lamps.
Type Temperature 964 °C 1066 °C 1086 °C 1703°C
Source of uncertainty Standard deviation
Reference blackbody radiator
B Impurity 0.008 0.010 0.010 0.021
B Temp gradient along cavity wall 0.017 0.019 0.020 0.042
B Emissivity 0.003 0.004 0.004 0.009
B Temp difference across cavity wall 0.001 0.001 0.001 0.002
Reference thermometer
B Ratio of photo current –measurement 0.007 0.008 0.008 0.018
B Ratio of photo current –resolution 0.009 0.003 0.004 0.003
A Non-linearity 0.034 0.040 0.000 0.088
A SSE 0.034 0.040 0.042 0.088
B Spectral response function 0.017 0.003 0.000 0.138
B Blocking at the side band 0.000 0.000 0.000 0.003
B Mean effective wavelength 0.000 0.000 0.000 0.003
Transfer Lamps
A Lamp current, as set 0.002 0.002 0.001 0.000
B Lamp current, as prescribed 0.002 0.002 0.001 0.000
Radiance temperature
B Short term stability 0.002 0.001 0.001 0.000
B Drift 0.017 0.019 0.020 0.042
B Dependence on wavelength 0.001 0.001 0.001 0.002
B Dependence on base temp. 0.006 0.002 0.002 0.000
A Alignment 0.017 0.019 0.020 0.042
B Target field 0.017 0.019 0.020 0.042
B Cleaning of the window 0.004 0.005 0.005 0.011
Lamp-thermometer composite
A Repeatability of radiance temperature 0.019 0.006 0.009 0.007
B Correction for SSE 0.034 0.040 0.042 0.088
B Correction for SSE & non-linearity 0.034 0.040 0.000 0.088
B Conversion to reference wavelength 0.001 0.001 0.001 0.002
B Conversion to reference base temperature 0.000 0.000 0.000 0.000
B Conversion to prescribed lamp current 0.013 0.012 0.012 0.010
Effective standard deviation of type A 0.050 0.061 0.047 0.132
Effective standard deviation of type B 0.060 0.069 0.058 0.202
Report on Calibration of Vacuum Tungsten-Strip Lamps as Transfer Standards for the Inter-comparison of
the Local Realizations of the ITS-90 between the Silver Point and 1700 °C
28 December 1998 23
Overall effective standard deviation 0.085 0.095 0.070 0.245
Report on Calibration of Vacuum Tungsten-Strip Lamps as Transfer Standards for the Inter-comparison of
the Local Realizations of the ITS-90 between the Silver Point and 1700 °C
28 December 1998 24
Table 10. Effective standard deviations of calibrations of the lamps at the specified reference
conditions.
Effective standard
deviation (°C)
Current of C564
Lamp (A)
Current of C681
Lamp (A)
Temp (°C)
0.085 4.480 5.508 964
0.090 4.721 5.822 1002
0.095 5.169 6.399 1066
0.075 5.322 6.594 1086
0.100 5.441 6.745 1102
0.115 6.272 7.795 1201
0.135 7.194 8.948 1301
0.160 8.189 10.183 1402
0.185 9.242 11.487 1502
0.210 10.347 12.851 1602
0.245 11.502 14.273 1703
5. Conclusions
According to the protocol, we have calibrated two sets of high-stability tungsten vacuum lamp (C564 and
C681) supplied by the pilot laboratory, VSL via NML. This experiment includes measurements of lamp
resistance, temperature drift after an 1 hour-stabilization, horizontal temperature distribution and angular
distribution of radiance temperature of the lamps. In this paper, we have reported the experimental and
theoretical procedures for the local realization of the ITS-90 in KRISS, the calibration results of the lamps
which has been corrected to the reference conditions of the protocol, and the uncertainty analysis of the
calibration. The uncertainty of the calibration in terms of the effective standard deviation is varying from
0.085 °C to 0.245 °C as temperature changes from the silver point to 1700 °C.
NIST_CCT.DOC 8/17/1998 Page 1 of 17
CCT Radiance Temperature Intercomparison Report
1. Laboratory designation: National Institute of Standards and Technology (NIST) 2. Address: NIST
Building 221, Room B208 Gaithersburg, MD 20899-0001 USA
3. Contact person: Mr. Charles Gibson Phone: 301 975 2329 Fax: 301 869 5700 E-mail: [email protected]
4. Lamp set # II: C860 & C864 5. Temperature points: 1000 οC to 1700 οC in 100 οC steps and 1064.18 οC 6. The measurement procedures will not be repeated here in this document. NIST followed
the procedures described in the ‘Protocol to the Comparison of Local Realizations of the ITS-90 between the Silver Point and 1700 οC, using Vacuum Tungsten-Strip Lamps as Transfer Standards’ with Appendixes A to D.
7. Realization of the ITS-90
The reference temperature standard, a gold fixed-point blackbody (Au) with a temperature (TAu) of 1064.18 C (1337.33 K), and the Planck radiation law are used to realize and disseminate the 1990 NIST Radiance Temperature Scale. Equation (1) is used to calculate the spectral radiance L8,Au(8, TAu) of the fixed-point blackbody for 8 = 655.3 nm in all the measurements of this calibration facility. Measurements are performed from 800 °C to 2300 °C for lamps, from 800 °C to 2700 °C for radiation thermometers, and extrapolated to 4200 °C for some disappearing filament optical pyrometers.
)1))/((exp( 2
521
−⋅⋅⋅⋅⋅
=Tncn
cL L
λλε
λλ
λλ (1)
Temperature is defined as a function of spectral radiance using the following equation
( ))/()(1ln),( 52
1
2
λλλλλ λελ
λLncn
cLT
L ⋅⋅⋅+⋅⋅= (2)
The NIST photoelectric pyrometer (PEP) is the transfer device used to compare the
spectral radiances of the sources by the direct substitution method. The signals are corrected for size of source, amplifier gain, and linearity. The NIST PEP is a filtered radiometer that uses two interference filters to select the bandpass. The spectral bandwidth is 5 nm with a mean effective wavelength of 655.3 nm. A photomultiplier tube with an S-20 spectral response is used in the DC mode. The measurement spot size is a 0.6 mm by 0.8 mm rectangle.
NIST_CCT.DOC 8/17/1998 Page 2 of 17
A high stability vacuum lamp operated at a single radiance temperature of approximately 1255 °C is the working standard (WS). The spectral radiance ratio is given by Eq. (3).
Au
WS
Au
WS1 )(
)(SS
TLTLr ==
λ
λ (3)
After applying correction factors to the signals in Eq. (3) for amplifier calibration (CA), linearity (CL), and size of source (CS), the spectral radiance of the WS lamp can be written as
( )( )AuSLA
WSSLA1
252
1WS
1expGCCCGCCC
r
Tncn
cL
Au
L
⋅⋅⋅⋅⋅⋅
⋅⋅
−
⋅⋅
⋅⋅
⋅=
λλ
ε
λλ
λ (4)
where G is the amplifier gain.
8. Transfer of radiance temperature to strip lamps
The values of radiance temperature apply when the lamp has been aligned to a specified orientation while operating at a designated radiance temperature and after the lamp has reached thermal stability at each specified operating current. The WS and Test Lamp (TL) are fully aligned at one temperature. At all other temperatures, the TL is translated vertically and/or horizontally so that the target area viewed by the PEP is always centered on the lamp filament at the height of the notch. No additional rotational alignments were performed.
The initial lamp current that corresponds to 1400 °C was selected from Appendix D of the intercomparison protocol. The TL was turned on and set to approximately 1400 °C and aligned after 30 min. The TL was spectrally compared to the WS to determine its radiance temperature. The spectral radiance ratio of the TL to the WS, r2, was measured three times by alternately translating the WS and the TL to the PEP. If the percent standard deviation of the ratio was less then the control limit, then the next calibration point was measured; otherwise, the point was repeated. The control limits were equal to the relative standard uncertainty (k = 1) of the ratio and were determined from previous calibration data. The control limits were the expected uncertainties for the spectral radiance ratio measurement. Next, the TL was aligned and measured at 1000 C after waiting 30 min. The additional calibration temperatures were aligned and measured in increasing order after waiting 30 min between data points. The measurement was repeated the next day starting at 1000 °C. The calibration log file was generated from each day of measurements and the following data was stored in a summary file for each test lamp: nominal temperature, measured temperature, lamp current, and spectral radiance ratio.
9. Description of gold-point blackbody
In the Radiance Temperature Calibration Laboratory (RTCL), a gold fixed-point blackbody with a calculated emissivity of 0.9999, designed and built by the NIST Optical Technology Division, is the primary standard used to realize the 1990 NIST Radiance Temperature Scale. The blackbody, shown in Fig. 1, consists of a graphite cavity, a crucible of gold, and a cylindrical heat-pipe furnace. The cavity, which is 76 mm in length and 6 mm
NIST_CCT.DOC 8/17/1998 Page 3 of 17
in diameter and has a 60 conical end shown, is made from Ultra “F” grade graphite (spectrographic purities of 10 ppm or less). Surrounding this cavity is a crucible containing 0.99999 pure gold. The cavity, along with graphite rings and silica glass spacers, is placed in an alumina tube. The front rings define a solid angle with a f/6 field of view, while the back rings support the thermocouple. A furnace (see Fig. 1), which consists of a sodium heat-pipe heated by two semi-cylindrical ceramic heater elements inside of a mullite tube, is enclosed in a water-cooled housing (631 mL/min) and is operated in an argon environment (37 mL/min with furnace door closed and 235 mL/min with furnace door open).
The duration of a melt or freeze plateau is approximately 40 min, and the time delay between these observation periods is about 45 min. Measurements during the freeze cycle show a negative slope of 20 mK in 30 min. The blackbody is slowly heated over about 8 h before reaching the melting point and is typically ramped up over night so that it is held just below the melting point the next morning. After the initial heating at 8 A, the melt cycle is begun by increasing the current to 8.5 A until the temperature reaches 1071 C. The freeze cycle is begun by lowering the blackbody current to 7.95 A. Then the blackbody current is raised to 8.5 A at 1050 C to begin the melt cycle again.
Figure 1 . Schematic of gold-point blackbody.
NIST_CCT.DOC 8/17/1998 Page 4 of 17
10. Working standard lamp
A vacuum tungsten ribbon filament lamp of the Quinn-Lee type is used in the temperature scale realization as the secondary temperature standard. This lamp maintains the temperature scale between scale realizations with the gold-point blackbody and as the transfer standard for calibration measurements. The temperature of the working standard lamp (serial number SL20) is determined by spectral comparison with the gold-point blackbody. This lamp is operated at a single current (7.7788 A DC) to produce a spectral radiance about eight times higher than that of the gold-point blackbody at 655.3 nm. The radiance temperature of working standard lamps is about 1255 °C. This lamp is stable to better than 0.1 °C over 100 h when operated under these single current conditions. 11. NIST photoelectric pyrometer
The PEP is a NIST-designed transfer radiometer, which uses refractive optics to image the source onto the detector. The schematic of the NIST PEP is shown in Fig. 2. A drawing of the measurement system is shown in Fig. 3. The measurement system is completely automated and controlled by a personal computer, while the laboratory environment is monitored by temperature and relative humidity sensors. Lamps and blackbodies are positioned onto the optical axis of the PEP using a closed-loop motor controller system that allows positioning to within 0.01 mm.
Figure 2. NIST photoelectric pyrometer
NIST_CCT.DOC 8/17/1998 Page 5 of 17
LEGENDAMP Amplifier
EM Environmental Monitor TC Thermoelectric CoolerGPBB Gold-point Blackbody TL Test LampGPL Gold-point Lamp TS Temperature SensorPC Personal Computer VTBB Variable Temperature BlackbodyPMT Photomultiplier Tube WSL Working Standard Lamp
PS Power SupplyDVM Digital Voltmeter RS Room Humidity Sensor
EM
PC
PS
AMP
DVM
VTBB
TL TLTLGPBB
WSL
GPL
Source Table
RS TS
PMT
TC
Figure 3. NIST radiance temperature laboratory measurement system.
12. Local conditions during measurements
Room temperature (Tamb): 24 °C ± 1 °C Relative humidity (RH): 25 % ± 5 %
13. Reference thermometer
Mean effective wavelength = 655.3 nm as defined below
( )
−⋅=−
2112
2 11/ln21 TTLL
cTTλ (5)
where T1 is the temperature of the WS lamp T2 is the temperature of the TL Half-width of spectral response function 5 nm Target distance 640 mm Target field dimensions 0.6 mm wide by 0.8 mm high rectangle Aperture ratio f/13 Detector S-20 PMT
NIST_CCT.DOC 8/17/1998 Page 6 of 17
Detector measuring mode DC Size of source effect
The magnitude of the size of source correction is determined by measuring the spectral radiance of a large uniform diffuse source with various apertures in front of the source. The large uniform diffuse source designed at NIST is used to measure the size of source effect. A 1 kW frosted quartz-halogen lamp is placed in a 20 cm by 23 cm by 20 cm vented housing. A lens focuses the lamp onto an opening in the housing that is covered by a 25.4 mm diameter circular diffuser. The apertures are on an aperture slide for quick positioning and reproducibility. The aperture sizes measured are a 1.4 mm by 25.4 mm slit, a 3 mm by 25.4 mm slit, and a 15 mm diameter hole which approximate the sizes of the WS filament, the TL filament, and the Au blackbody opening. The ratio r1 is the measurement of the ratio of the WS to the AuBB, see Eq. (3). The size of source correction is determined from the ratio of the measurements of the 15 mm diameter hole to the 1.4 mm by 25.4 mm slit. The correction factor is 1.0009 ± 0.0006 and the temperature correction is 0.10 °C. A negligible difference was observed for the measurements of the 1.4 mm by 25.4 mm slit and the 3 mm by 25.4 mm slit; therefore, the correction factor for ratio of the TL to the WS is 1.
14. Transfer lamps Nominal base temperature: C860 19.9 °C at 1000 °C and 20.7 °C at 1700 °C C864 19.7 °C at 1000 °C and 20.2 °C at 1700 °C Base temperatures repeated better than 0.01 °C Total burning time: Date Lamp Task New hours Total hours 11/14/97 C860 Restabilization
1100°C, 1700°C for 1 h, 1100°C 5 h 25 min 5 h 25 min
11/17/97 C860 Measurements 1400°C, 1000°C to 1700°C
11 h 2 min 16 h 27 min
11/18/97 C860 Measurements 1000 °C to 1700°C
8 h 38 min 25 h 5 min
11/23/97 C860 Fixed point 1064.18 °C
4 h 35 min 29 h 40 min
11/14/97 C864 Restabilization
1100°C, 1700°C for 1 h, 1100°C 5 h 18 min 5 h 18 min
11/17/97 C864 Measurements 1400°C, 1000°C to 1700°C
10 h 58 min 16 h 16 min
11/18/97 C864 Measurements 1000 °C to 1700°C
8 h 38 min 24 h 54 min
11/23/97 C864 Fixed point 1064.18 °C
4 h 33 min 29 h 27 min
NIST_CCT.DOC 8/17/1998 Page 7 of 17
15. Ambient conditions
Room temperature (Tamb): Mean: 24.6 °C Max.: 25.0 °C
16. Resistance results C860 11/14/97 Ramb = 39.863 mΩ Tamb = 21.486 °C C860 11/24/97 Ramb = 39.789 mΩ Tamb = 21.077 °C C864 11/14/97 Ramb = 41.574 mΩ Tamb = 21.148 °C C864 11/24/97 Ramb = 41.600 mΩ Tamb = 21.109 °C The lamp resistance was measured on 11/14/97 before removing the lamps from the case. After the electrical connections were made, insulation was placed around the case opening, and the lamps were allowed to reach equilibrium overnight. Calibrated 25 kΩ thermistors were used to measure the base temperature. Bushings and heat sink grease were used to improve thermal contact. 17. Lamp mapping data is in text file: NIST CCT Lamp mapping.txt
-0 .40%
-0.35%
-0.30%
-0.25%
-0.20%
-0.15%
-0.10%
-0.05%
0.00%
0.05%
0.10%
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
Horizontal distance from target [m m ]
Cha
nge
in R
adia
nce
[%]
C860 #1C860 #2C864 #1C864 #2
(- va lues are on the notch side)
Figure 4. T ranslational m apping of the lam p filam ent a long the horizontal axis with a 0.6 m m x 0.8 m m target area.
NIST_CCT.DOC 8/17/1998 Page 9 of 17
18. Table A data is in text file: NIST CCT Table A.txt C860 Table A SSE & Non-lin corr. No. Corr. SSE Corr. SSE & Non-lin corr. Meas #. Itab: I(j) Iset: I(l) I(j)-I(l) STL/SAu TAu Tλ Tλ Tλ
Notes: Numbers 19 thru 21 are measurements using the gold point
blackbody
NIST_CCT.DOC 8/17/1998 Page 14 of 17
λr = 650.0 nm λr - λe = -5.3 nm The following temperatures were corrected for base temperature: 1000 oC, 1064.18 oC, and 1100 oC The wavelength coefficients, dTλ/dλ, were taken from the table in Appendix B The temperature coefficients, dTλ/dTb, were calculated using the polynomial in Appendix D
NIST_CCT.DOC 8/17/1998 Page 15 of 17
20. The Uncertainty analysis is summarized in Tables C, D, & E below: Table C: Working Standard lamp Expanded Uncertainties
Expanded Source of Uncertainty Type Uncertainties (oC)
Refractive index B 0.03 Wavelength B 0.37 Freezing temperature of gold B 0.30 Second radiation constant B 0.03 Emissivity of Au B 0.02 First radiation constant B 0.00 Ratio of WS signal to Au signal A 0.03 WS amplifier calibration correction A 0.01 WS linearity correction A 0.11 WS size of source correction A 0.06 WS amplifier gain B 0.00 Au amplifier calibration correction A 0.01 Au linearity correction A 0.11 Au size of source correction A 0.02 Au amplifier gain B 0.00 Digital voltmeter B 0.00 WS current B 0.11 WS stability B 0.11
Expanded uncertainty U = kuc(T), where k = 2 0.53
NIST_CCT.DOC 8/17/1998 Page 16 of 17
Table D: Test lamp vs. Au Expanded Uncertainties
ExpandedSource of Uncertainty Type Uncertainties (oC)
Refractive index B 0.03Wavelength B 0.37Freezing temperature of gold B 0.30Second radiation constant B 0.03Emissivity of Au B 0.02First radiation constant B 0.00Ratio of TL signal to Au signal A 0.09TL amplifier calibration correction A 0.01TL linearity correction A 0.11TL size of source correction A 0.06TL amplifier gain B 0.00Au amplifier calibration correction A 0.01Au linearity correction A 0.11Au size of source correction A 0.02Au amplifier gain B 0.00Digital voltmeter B 0.00TL current B 0.13TL stability B 0.11
Expanded uncertainty U = kuc(T), where k = 2 0.54
NIST_CCT.DOC 8/17/1998 Page 17 of 17
Table E: Test lamp Expanded Uncertainties
Expanded Uncertainties [ oC]
Source of Uncertainty Type Temperature [ oC]
1000 1100 1200 1300 1400 1500 1600 1700
1. Calibration of the reference radiance temperature lamp relative to the 1990 NIST Radiance Temperature Scale A 0.37 0.43 0.49 0.56 0.64 0.71 0.80 0.882. Test lamp temperature determination A 0.07 0.09 0.10 0.11 0.13 0.14 0.16 0.183. Current measurement B 0.14 0.13 0.12 0.11 0.11 0.11 0.11 0.10
4. Mean effective wavelength measurement for the NIST Photoelectric Pyrometer B 0.06 0.04 0.02 0.01 0.05 0.09 0.13 0.185. Test lamp alignment B 0.12 0.14 0.16 0.19 0.21 0.24 0.27 0.30
6. 1990 NIST Radiance Temperature Scale relative to the Thermodynamic Temperature Scale B 0.21 0.24 0.28 0.32 0.36 0.40 0.45 0.50
Expanded uncertainty U = kuc(T), where k = 2 0.47 0.54 0.61 0.69 0.78 0.88 0.98 1.09
1
Report on the comparison of local realization of the ITS-90 between the silver point and 1700°C using vacuum tungsten-strip lamps as transfer standards
Wang Li
National Measurement Centre, Singapore Productivity and Standards Board
1. Local realization of the ITS-90 1.1. Description of equipment 1.1.1. Reference thermometer The optical arrangement of the thermometer is shown in Figure 1. The thermometer works at a fixed target distance of 500 mm from the objective lens L1 (focal length 250 mm) with unit magnification. The nominal target size is determined by a field stop of 0.85 mm in diameter. The objective lens L1 acts as the aperture stop which determines an aperture ratio of 10.4:1, whilst lens L2 acts as a collimator. Two interference filters with nominal centre wavelength of 649 nm are used. The half-height bandwidth of two filters is 5nm and 10 nm respectively. A silicon detector (Hamamatsu, Type No. 1337-1010 BR) operated in the photovoltaic mode is used. The photocurrent from the detector is converted to voltage by an operational amplifier (Burr-Brown OPA128 LM) with one feedback resistor of 107 Ω. The thermometer operates from the silver point to 1700°C without introducing any neutral filter. The interference filter and the detector are housed in a detachable module which is removed for insertion of a telescope during alignment. The detector and the interference filters are not under temperature control. 1.1.2. Fixed point The silver point is used for the fixed point calibration. A schematic view of the arrangement for realising the Ag freezing point is shown in Figure 2. The design of the crucible and blackbody assembly is the same as IMGC’s blackbody fixed point. The metal ingot is held in pure graphite (5 ppm impurity) crucible. Its available volume is about 48 cm3. The silver sample is from Johnson Matthey with nominal purity better than 99.9999%. The blackbody cavity is a cylinder 9.5 mm in diameter and 66 mm in depth terminated with a conical bottom. The cavity aperture is delimited by a graphite diaphragm 3 mm in diameter. The effective emissivity of the cavity is estimated to be 0.999938. The crucible is mounted in a quartz tube (50 mm in diameter) with a protection cone in front. The quartz tube is then inserted into a three-zone furnace. Argon gas is used to protect the graphite parts from oxidation during the realisation. The temperature distribution of the circular area around the blackbody aperture is measured during the realization of the Ag for SSE correction. 1.1.3. Measurement of relative spectral responsivity and non-linearity The relative spectral responsivity and the non-linearity were measured by NML, Australia, two months before the comparison. The relative spectral responsivity of the thermometer was measured in situ by reference to 3 element (5 reflection) trap detector, the quantum efficiency of which was uniform to better than 99.9% over 500-850 nm. Outside this region the characterised trap response function was used. In the pass-band region of the interference filters, the measurement was done in an interval of 0.5 nm with a monochromator half bandwidth of 0.1 nm. An interval of 4 nm was used with the monochromator half bandwidth of 1 nm outside the pass-band of the filters. The effective wavelength is then calculated at different temperatures according to its definition. The non-linearity measurement was done for a range of input fluxes by a flux addition method using a beam splitter. Detector voltages were incremented in ratios of 2 until the voltage signal equivalent to a temperature of approximately 1800°C was reached. It was found that the error that would be incurred by ignoring the linearity correction is 0.012°C at 1700°C.
2
1.2. The ITS-90 above the silver point The ITS-90 is kept in the reference thermometer itself. No strip lamps are used. The thermometer is calibrated at the Ag point before every usage. The temperatures corresponding to different voltage signals are then calculated according to the ITS-90 definition. Non non-linearity correction is applied. The non-linearity is considered as a part of the uncertainty. The SSE correction is applied to every calculated temperatures. 1.3. Measurement steps of the comparison a. Measurement of positioning effects (spatial and rotational) b. Restabilization of the transfer lamp c. Calibration of the thermometer at the Ag point then calibration of the transfer lamp from I(1) to I(11).
The whole cycle was finished within one day. d. Repetition of step c for 3 times. Three runs of measurement results were thus obtained for each transfer
lamp. 2. Presentation of results 2.1. Local conditions 2.1.1. Reference thermometer - Effective wavelength: 649.175 nm (in vacuum) at the Ag point. - Half-width of spectral response function: 5 nm - Aperture ratio: 10.4:1 - Target distance: 500 mm - Target field dimensions: 0.85 mm - SSE: see Figure 3 2.1.2. Transfer lamps - Orientation of the lamps: reference orientation, rotational effect see Figure 4 - Nominal base temperature: 20°C ± 0.05°C (1 σ level) - Total burning time: lamp 1 = 30 h; lamp 2 = 30 h 20 minutes 2.1.3. Ambient conditions - Tamb: 23 ± 1°C - RH: 55% ± 5% 2.2. Measurement result 2.2.1. Data tabulation Lamp 1 (C564) The measurement results of lamp 1 are listed in Table A1 and Table B1 as required. Due to the total burning time constrain, I(3) and I(4) were not measured in the third run. The final Tλ[λr;I(j)] are listed in the last column of Table B1. They are obtained by taking the means of three run results except at I(3) and I(4). For these two currents, the final Tλ[λr;I(j)] are obtained by taking the means of first two run results. The parameters ∂Tλ/∂I are obtained from a best fit of the final Tλ[λr;I(j)] and I(j) using a 6 order polynomial. Lamp 2 (C681) The measurement results of lamp 2 are listed in Table A2 and Table B2 as required. Due to some alignment problem, the first run results and the results of I(1) and I(9) of the second run have some significant errors and
3
are eliminated. As in lamp 1 case, due to the total burning time constrain, only some currents were measured in the third run. The final Tλ[λr;I(j)] are listed in the last column of Table B2. They are obtained by taking the means of two run results for those currents which have data in both runs. For the rest of the currents, the final Tλ[λr;I(j)] are from single run (either from 2nd run or 3rd run). The parameters ∂Tλ/∂I are obtained from a best fit of the final Tλ[λr;I(j)] and I(j) using a 6 order polynomial. 2.2.2. Currents at fixed point temperatures As described in 1.3, the Ag point was realised every time before measurements on the transfer lamps were taken. The Ag point was not transferred to the transfer lamps. However, the T(FP) vs. I[T(FP)] can be obtained by using ∂Tλ/∂I if required. 2.2.3. Ramb and Tamb Some difficulties were encountered in following the recommended method of measurement due to the instability of the room temperature. To solve this problem, each of the lamps was kept inside a temperature controlled air-bath at 23°C during the measurement. Before After Ramb (Ω) Tamb (°C) Ramb (Ω) Tamb (°C) Lamp 1 0.0402676 22.975 0.0402218 22.937 Lamp 2 0.0343305 22.968 0.0343450 22.970 2.3. Uncertainties 2.3.1. Uncertainty at the Ag temperature The uncertainty components at the Ag temperature (Tag) and their values are listed in Table C. Table C. Uncertainties at Tag. Uncertainty components Standard
uncertainty (mK)
Combined standard uncertainty
s(TAg)(mK) Impurities in Ag sample Emissivity of cavity Temperature drop across cavity bottom Random uncertainty including the noise and the repeatability of plateaus Short term stability of the thermometer Differential sse correction between blackbody and lamp *
5.0
4.3
1.3
52.0
19.0
10
57 * Note: 1. The temperature along the strip is considered as uniform. 2. The 1.5mm strip sse curve (see Figure 3) is used for both lamps although their strips have different width. 3. The errors due to these operations are considered as parts of the uncertainties. 2.3.2. Uncertainties in Tλ [I(j)]
4
The uncertainty components at Tλ [I(j)] and their values are listed in Table D1 and Table D2 for lamp1 and lamp2 respectively. In the Tables, the following terms are used: s(λe): standard uncertainty of the effective wavelength measurement. s(λr): standard uncertainty of reference wavelength due to uncertainty in ∂Tλ/∂λ and uncertainty in
effective wavelength measurement. s(I): standard uncertainty of the current measurement. s(Tb): standard uncertainty of base temperature due to uncertainty in ∂Tλ/∂Tb and uncertainty in Tb
measurement. s(random): random standard uncertainty including the noise and the repeatability of the lamp readings. s(position): standard uncertainty of positioning (spatial and rotational). s(drift of signal): standard uncertainty due to drift of the reference thermometer during measurement. s(drift of lamp): standard uncertainty due to drift of the lamps during measurement. s(Tamb): standard uncertainty due to ambient temperature variation during the measurement. s(non-linearity): standard uncertainty of non-linearity of the photodetector.
5
5
Table A1. Measurement results of lamp 1 (C564).
Number of measurement
I(j) (A)
I(l) (A)
I(j)-I(l) (A)
R Tλ(λe;Tb) (°C)
Tλ(λe;Tb) (°C) corrected for SSE
Tλ(λe;Tb) (°C) corrected for SSE and non-linearity*
CCT Key Comparison: ITS-90 from 962 °°°°C to 1700 °°°°C, NPL Measurements, July to August 1997 H C McEvoy and K M Raven
Restricted-Commercial NPL Report CBTM S7
CONTENTS Page 1 DESCRIPTION OF THE PYROMETER ....................................................... 1 1.1 THE OPTICAL SYSTEM......................................................................... 1 1.2 THE DETECTOR...................................................................................... 2 1.3 THE FIXED POINT BLACKBODY SOURCES .................................... 2 1.4 REALISING THE TEMPERATURE SCALE USING THE
PYROMETER............................................................................................ 2 2 MEASUREMENTS PERFORMED ON THE LAMPS ................................. 3 2.1 AGEING AND STABILISATION ........................................................... 3 2.2 MEASUREMENT OF Ramb, THE ROOM TEMPERATURE
RESISTANCE OF THE LAMP FILAMENT .......................................... 3 2.3 SETTING UP THE LAMPS ..................................................................... 3 2.4 POSITIONAL EFFECT CHECKS ........................................................... 4 2.5 MEASUREMENT OF THE BASE TEMPERATURE
COEFFICIENTS........................................................................................ 4 2.6 CALIBRATION OF THE LAMPS........................................................... 4 2.7 RE-MEASUREMENT OF Ramb ................................................................ 5 2.8 SIZE-OF-SOURCE EFFECT MEASUREMENTS ................................. 5 3. RESULTS ............................................................................................................. 6 3.1 DRIFT RATE OF THE LAMPS............................................................... 6 3.2 RESULTS OF THE MEASUREMENTS OF Ramb .................................. 6 3.3 RESULTS OF THE POSITIONAL EFFECT CHECKS ......................... 6 3.4 BASE TEMPERATURE COEFFICIENTS.............................................. 7 3.5 RESULTS OF THE SSE MEASUREMENTS......................................... 7 3.6 RESULTS OF THE CALIBRATIONS OF THE LAMPS....................... 8 3.7 RESULTS OF THE MEASUREMENTS AGAINST THE FIXED
POINTS..................................................................................................... 11 3.8 POLYNOMIAL FITTING OF THE CALIBRATION DATA ............... 13 4. UNCERTAINTIES............................................................................................. 14 5. REFERENCES ................................................................................................... 16 FIGURES ........................................................................................................................ 17
Restricted-Commercial NPL Report CBTM S7
1
CCT key comparison: ITS-90 from 962 °°°°C to 1700 °°°°C NPL measurements, July-August 1997 The following is a description of the equipment and procedure used at NPL for the initial calibration of the three high-stability lamps for the CCT project 'Comparison of Local Realisations of the ITS-90 between the Silver Point and 1700 °C using Vacuum Tungsten-Strip Lamps as Transfer Standards'. 1 DESCRIPTION OF THE PYROMETER The NPL primary pyrometer, originally built by Quinn and modified by Coates, was used to calibrate the three lamps. This has been described previously [1,2,3] and a description is also given below. The pyrometer has an f/11 aperture ratio. For a source to be correctly focused it must be place approximately 120 cm from the off-axis ellipsoidal mirror. A schematic diagram of the pyrometer is shown in Figure 1. 1.1 THE OPTICAL SYSTEM The radiation from the source is limited by a circular aperture stop, then focused by an off-axis ellipsoidal mirror. The beam is then reflected by a plane mirror to produce an image, of unit magnification, on a 0.75 mm diameter field stop. After passing through the field-stop aperture, the radiation is collimated into a beam approximately 1 cm in diameter by a small plano-convex lens, and then passes through a combination of four neutral-density filters. One or more of these filters can be put into the beam to attenuate the signal to bring it within the dynamic range of the system. The beam then passes through one of a number of interference filters. Those currently in the pyrometer have peak transmission wavelengths of nominally 665, 800 and 906 nm. The full width at half maximum transmission (FWHM) of each of the filters is 22 nm, 24 nm and 8 nm respectively. The 665 and 906 nm filters are each placed in combination with a broad-band filter (FWHM = 100 nm and 32 nm respectively) centred at approximately the same wavelength as the corresponding narrow band filter. This was done to improve the out-of-band blocking, especially above 1150 nm where there was found to be significant transmission. For the work reported here only the 665 nm filter combination was used. The filter/filter combinations are calibrated in-situ using a scanning monochromator and line sources. A small platinum resistance thermometer placed beside the filter wheel monitors the temperature of the filters. The change in transmission wavelength with temperature (temperature coefficient) of each filter is also measured; thus a correction can be applied to allow for changes in filter temperature. The optical alignment of the pyrometer on the source is achieved by moving a small plane mirror into the beam, as shown in Figure 1, and then observing the image produced on a circular graticule using the viewing telescope. With the mirror in this position, the optical path to the detector is blocked, preventing radiation from the source from reaching the detector. This enables the background radiance signal to be determined. 1.2 THE DETECTOR
Restricted-Commercial NPL Report CBTM S7
2
The detector is a Hamamatsu type S1337 - 1010BQ silicon photodiode. It is mounted approximately 25 mm from the interference filter wheel on a number of optical positioning stages. The mounting incorporates a heating element with associated temperature controller to maintain the detector at 25 °C. The detector is connected to a digital voltmeter (DVM) via an amplifier, from which the output of the detector can be read in terms of a voltage. The DVM is connected, via an IEEE interface, to a computer which is used for data acquisition and processing. The linearity of the detector has been verified using both a radiance halving (double aperture) technique and a, slightly different, radiance doubling technique. 1.3 THE FIXED POINT BLACKBODY SOURCES The NPL Ag and Au fixed point blackbody sources were used for this work. These have been described previously [1,4]. They are comprised of a graphite crucible containing an ingot of 99.9999% pure metal within an electrically heated furnace. Each source has a 3.0 mm diameter aperture, defined by a Rh disc placed immediately in front of the blackbody aperture. 1.4 REALISING THE TEMPERATURE SCALE USING THE PYROMETER The NPL primary pyrometer is used as a comparator of radiance temperatures. The radiance from a source at a known temperature, normally a fixed point blackbody at the Ag or Au freezing point or a high-stability lamp, is compared with that of the test source at the chosen wavelength. The temperature of the test source is then determined from Planck's Law (Equation 1) using the effective wavelength method described by Kostkowski and Lee (Equation 2) [5,6,7,8].
where T90(X) corresponds to the kelvin temperature of the Ag, Au (or Cu) freezing point, Lλ[T90] and Lλ[T90(X)] are the spectral radiances of the blackbodies at temperatures T90 and T90(X) respectively, λ is the wavelength in vacuo, c2 is the second radiation constant, 0.014388 m.K and λe is the effective wavelength.
1) -] Tc[ ( / 1) -]
(X)Tc[ ( = (X))T(L / )T(L
90
2
90
29090 λλλλ expexp 1
TB + A = 1
90eλ 2
Restricted-Commercial NPL Report CBTM S7
3
2 MEASUREMENTS PERFORMED ON THE LAMPS 2.1 AGEING AND STABILISATION Three high-stability evacuated tungsten ribbon lamps, manufactured by Polaron Special Lamps Division, UK, were used for the calibration. Each was a type 10/V lamp and had a 1.5 mm wide filament. The lamp numbers were C840, C864 and C860. C864 and C860 were to be circulated as part of the intercomparison; C840 was to be kept in reserve in case of breakage of either of the other lamps. The lamps were aged at 1800 °C, in order to improve their stability, then pre-stabilised and stabilised for 100 hours at 1700 °C prior to the calibration [9]. The drift rate for each lamp was determined during the stabilisation process; the values obtained are shown in Section 3.1. 2.2 MEASUREMENT OF Ramb, THE ROOM TEMPERATURE RESISTANCE OF THE
LAMP FILAMENT Before any measurements were made on the lamps, Ramb was measured. An ASL F18 bridge was used, in conjunction with a 1Ω standard resistor. The latter was held in a bath of stirred oil at 20.5 (±0.3) °C. The temperature of the resistor was measured using a calibrated PRT probe. Leads from the current terminal of the bridge were connected to the current terminals of the lamp; leads from the voltage terminal of the bridge were connected to the water tubes on the lamp base, using crocodile clips. A calibrated type T thermocouple was used to measure tamb, the temperature of the lamp base. The following bridge settings were used: source impedance = 10 Ω; carrier = 20 mA and 20√2 mA; low frequency (25 Hz); gain = 103; bandwidth = 0.5 Hz. Measurements were made with the lamps in their carrying case, the case grounded, and the lid closed as much as possible. The results are shown in Section 3.2 Table 1. 2.3 SETTING UP THE LAMPS Once Ramb had been measured the lamps were set up in front of the primary pyrometer. They were mounted on bases which allowed movement in all three planes, plus angular and tilt adjustment. Each lamp was aligned using a plumbline so that the filament was vertical when viewed from the side and the rear. The angular position was set so that the spot on the rear envelope was positioned directly behind the filament along the optical axis of the pyrometer. The height and horizontal position of the filament were adjusted so that the pyrometer was viewing a circular area of the filament, 0.75 mm in diameter, at the height of the notch and midway between the mouth of the notch and the opposite side of the filament. The lamp current was provided by three power supplies (one for each lamp); two were manufactured by Vinculum Ltd, the other by Hewlett Packard. For each lamp the current was determined by measuring the voltage across a 0.01 Ω standard resistor incorporated into the circuit. The voltage was measured using a calibrated digital voltmeter (DVM). The three standard resistors were held in a stirred oil bath maintained at 25 °C. The temperature was monitored using three calibrated mercury-in-glass thermometers, one placed in the central well of each resistor. The temperature of the lamp bases was controlled at 20.0 (± 0.3) °C using cooled circulating water. The temperature was measured using a calibrated type T thermocouple placed in the hole in each base. Good thermal contact was obtained by packing the hole with thermal conducting paste. A second calibrated DVM was used to measure the output from each thermocouple.
Restricted-Commercial NPL Report CBTM S7
4
Throughout both calibration runs the room temperature was measured using a calibrated thermometer placed close to the pyrometer. The temperature of the interference filter was also measured so a correction could be applied to the wavelength to allow for any drift in the filter temperature. During the second run, the relative humidity was measured using a calibrated meter. 2.4 POSITIONAL EFFECT CHECKS The front window of each lamp was cleaned. The lamps were run for one hour at approximately 1500 °C to relieve strain in the filament caused by moving them. Then, with them still at about 1500 °C, the following checks were performed: i) The filament was rotated about the vertical axis in 1° steps up to ± 10° from the normal
alignment position; ii) The filament was rotated about the horizontal axis perpendicular to the optical axis of
the pyrometer in 0.5° steps over ± 3° from the normal (vertical) position; iii) The pyrometer was scanned across the filament, at the height of the notch, in steps of
0.125 mm to each edge of the filament. This was done to assess the horizontal radiance distribution.
The results of all these checks can be found in Section 3.3. 2.5 MEASUREMENT OF THE BASE TEMPERATURE COEFFICIENTS Base temperature coefficients (BTCs) (i.e. the change in radiance temperature of the lamp filament per °C change in the base temperature) were determined for all three lamps at 900 °C, 1000 °C and 1100 °C. This was done by altering the base temperature from 8 °C to 35 °C and measuring the change in radiance temperature. The coefficients were fitted using a second order polynomial equation: BTC = a + bt +ct2, where BTC is the base temperature coefficient, and t is the radiance temperature in °C. The base temperature coefficients and the polynomial expressions are given in Section 3.4, Tables 2 to 4. 2.6 CALIBRATION OF THE LAMPS The front window of each lamp was cleaned using ethanol and a lens tissue. The lamps were then calibrated over the range 962 °C to 1700 °C. At 962, 1000, 1064 and 1084 °C they were calibrated by direct comparison with either the Ag or Au fixed point blackbody source. Several melts and freezes were performed for each lamp temperature, and the average result was obtained for each lamp. Above 1084 °C, the calibration was carried out using a radiance doubling technique, using the measurements against the gold point as the reference. The front window of each lamp was cleaned again, then the calibration was repeated as above. At each temperature, the measurements for each lamp were corrected to a particular current using a typical current/temperature relationship for this type of lamp. As the current corrections were small, this would not have introduced any significant error. At and below 1100 °C the results were also corrected to a base temperature of 20 °C using the derived polynomial
Restricted-Commercial NPL Report CBTM S7
5
expression. The measurements were made at a wavelength of approximately 664.64 nm. The results were corrected to 650 nm using the polynomial equation provided in the protocol. Throughout both calibrations, the maximum rate of increase or decrease of current was 1 A/minute. Overnight the current to all lamps was turned off. The total burning time of the lamps was 52 hours for the first run and 47 hours for the second run. The results of the measurements are given in Tables 5 to 10. The results of all the measurements made against the Ag and Au fixed-point blackbodies are given in Tables 11 to 16. 2.7 RE-MEASUREMENT OF Ramb Before the lamps were transported to their next destination Ramb was remeasured as before. The results are given in Table 1. 2.8 SIZE-OF-SOURCE EFFECT MEASUREMENTS The size-of-source effect (SSE) of the pyrometer was measured before and after both calibration runs in order to assess the correction to be applied to the results. This was done using a large area heat-pipe blackbody source in conjunction with a set of black aperture plates. The apertures ranged from 0.75 mm to 25 mm diameter and were placed, in turn, onto a water-cooled plate positioned immediately in front of the blackbody source. A thermocouple in the rear of the blackbody allowed the temperature to be monitored so that a correction could be made for any drift during the measurements. In addition, the pyrometer was scanned across the front of both the Ag and Au blackbodies during a melt and freeze in order to assess the thermal profile and to enable the effective source diameter to be determined (see [10]). The SSE versus aperture diameter was fitted using the expression:
where y is the SSE (relative to the 25 mm diameter aperture), x is the aperture diameter (in mm), and a = 9.996985x10-1, b = -3.488035x10-3, c = -3.898927x10-1 This expression was used to calculate the SSE for the fixed points and lamp using the thermal profiles of the furnaces and allowing for the effect of the strip-shaped lamp filament. Hence the effective diameters of all the sources could be calculated. A correction was then applied to the lamp radiance temperature to allow for the SSE, as described in [10]. The results are given in Section 3.5. 3. RESULTS 3.1 DRIFT RATE OF THE LAMPS The drift rates of the lamps during the stability tests were as follows: Lamp C860: -0.03 °C/100 hours at 1700 °C Lamp C864: -0.29 °C/100 hours at 1700 °C
(cx)]b[ + a =y exp 3
Restricted-Commercial NPL Report CBTM S7
6
Lamp C840: -0.08 °C/100 hours at 1700 °C. The uncertainty in measuring the stability, evaluated at an uncertainty of approximately 95% confidence, is 0.6 °C. All the drift rates were therefore within one standard measurement uncertainty, and within the limit of 0.3 °C/100 hours required by the Comparison protocol. 3.2 RESULTS OF THE MEASUREMENTS OF Ramb Table 1.
Lamp number
Pre-calibration Post calibration
Ramb Ω Self-heating
Ω tamb
°C Ramb Ω
Self-heating Ω
tamb
°C
C864 0.041684 -2x10-6 22.0 0.042805 +1x10-6 28.5
C860 0.039923 -3x10-6 21.9 0.040234 +2x10-6 24.5
C840 0.040426 +8x10-6 21.2 0.041520 -1x10-7 28.1 All the self-heating effect values are insignificant (< 10-5 Ω). 3.3 RESULTS OF THE POSITIONAL EFFECT CHECKS i) The results of the rotation of the lamps about a vertical axis are shown in Figures 2 to 4. ii) Rotation of the lamps about the horizontal axis perpendicular to the optical axis of the
pyrometer showed negligible variance in the radiance temperature (<0.04 °C) within ± 3° of the vertical position.
ii) Scanning the pyrometer horizontally across the filament showed a region of ± 0.25 mm
around the normal alignment position where the radiance temperature varied by ≤ 0.2 °C. Thus, the target size of the source is sufficient to fill the field-of-view of the pyrometer.
3.4 BASE TEMPERATURE COEFFICIENTS The results of the measurements of the base temperature coefficients are given in the following tables; a,b and c are the polynomial coefficients derived from the data. Table 2 - Lamp number C864
t(°C) at 650 nm Base temperature coefficient
907.30 0.0697
999.84 0.0280
1103.12 0.0084
Restricted-Commercial NPL Report CBTM S7
7
a = 1.68557; b = -2.98818x10-3; c = 1.33056x10-6 Table 3 - Lamp number C860
t(°C) at 650 nm Base temperature coefficient
907.52 0.0781
1001.61 0.0315
1104.19 0.0112
a = 1.90290; b = -3.38388x10-3; c = 1.51305x10-6 Table 4 - Lamp number C840
t(°C) at 650 nm Base temperature coefficient
905.80 0.0800
1001.56 0.0306
1102.23 0.0099
a = 1.98017; b = -3.52845x10-3; c = 1.57945x10-6 3.5 RESULTS OF THE SSE MEASUREMENTS The effective diameters of the Ag and Au fixed point blackbody sources were found to be 11.10 and 13.59 mm respectively. The effective width of the lamp filament was 2.96 mm. This leads to a correction of +0.11 °C and +0.14 °C for the lamps at the Ag and Au point respectively. The SSE correction at higher temperatures may be found by extrapolation.
Restricted-Commercial NPL Report CBTM S7
8
3.6 RESULTS OF THE CALIBRATIONS OF THE LAMPS Tables 5 to 10 show the results of both calibration runs. Corrections were applied to convert the results to a wavelength of 650 nm (6th column) and to allow for the pyrometer's SSE (7th column). The last column in the Tables gives the final corrected radiance temperatures for each lamp current. These were fitted using a 6th order polynomial equation, to give the current/temperature relationship for each lamp, and the results tabulated. Note that, in the tables, λ is the reference wavelength of the pyrometer from the filter calibration; it is not the effective wavelength λe. The room temperature during both calibration runs was 22.0 (± 1.0) °C. Maximum and minimum values were 24.2 °C and 20.7 °C respectively for the first run, and 24.3 °C and 20.3 °C respectively for the second run. During the second run the relative humidity was 33.0 (± 4.0)%, with maximum and minimum values of 42.6% and 26.4% respectively. Table 5 - Lamp C864, 1st calibration run
3.7 RESULTS OF THE MEASUREMENTS AGAINST THE FIXED POINTS Tables 11-16 show all the results of the measurements against the fixed point blackbody sources. Table 11 - Lamp C864, 1st calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 4.9474 962.227 664.650 963.854 963.964
Ag 4.9474 962.222 664.656 963.849 963.959
Au 5.2805 1003.657 664.647 1005.398 1005.518
Au 5.8076 1064.201 664.652 1066.120 1066.260
Au 5.8076 1064.212 664.658 1066.132 1066.272
Au 5.9967 1084.754 664.629 1086.733 1086.873
Restricted-Commercial NPL Report CBTM S7
14
Table 12 - Lamp C860, 1st calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (reference)
(nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 5.0809 961.666 664.650 963.291 963.401
Ag 5.0809 961.674 664.656 963.300 963.410
Au 5.4294 1004.257 664.647 1006.000 1006.120
Au 5.9615 1064.129 664.652 1066.048 1066.188
Au 5.9615 1064.131 664.658 1066.051 1066.191
Au 6.1586 1085.046 664.629 1087.026 1087.166
Table 13 - Lamp C840, 1st calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (reference)
(nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 5.0149 962.311 664.650 963.938 964.048
Ag 5.0149 962.305 664.656 963.933 964.043
Au 5.3227 1000.315 664.647 1002.046 1002.166
Au 5.8825 1063.999 664.652 1065.917 1066.057
Au 5.8825 1063.998 664.658 1065.917 1066.057
Au 6.0753 1084.685 664.629 1086.664 1086.804
Table 14 - Lamp C864, 2nd calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (reference)
(nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 4.9490 962.477 664.648 964.104 964.214
Ag 4.9490 962.471 664.643 964.098 964.208
Ag 5.2827 1003.941 664.641 1005.682 1005.802
Au 5.8091 1064.403 664.633 1066.320 1066.460
Au 5.9959 1084.705 664.639 1086.685 1086.825
Restricted-Commercial NPL Report CBTM S7
15
Au 5.9959 1084.710 664.648 1086.691 1086.831
Restricted-Commercial NPL Report CBTM S7
16
Table 15 - Lamp C860, 2nd calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (reference)
(nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 5.0824 961.690 664.648 963.315 963.425
Ag 5.0824 961.661 664.643 963.285 963.395
Ag 5.4308 1004.252 664.641 1005.994 1006.114
Au 5.9631 1064.160 664.633 1066.076 1066.216
Au 6.1585 1084.900 664.639 1086.881 1087.021
Au 6.1585 1084.895 664.648 1086.877 1087.017
Table 16 - Lamp C840, 2nd calibration run
Fixed point used
Measured lamp current
(A)
Lamp temperature (°C)
λ (reference)
(nm)
Lamp temperature corrected to
650 nm (°C)
Lamp temperature SSE corrected (°C)
Ag 5.0152 962.037 664.648 963.663 963.773
Ag 5.0152 962.024 664.643 963.649 963.759
Ag 5.3228 1000.074 664.641 1001.804 1001.924
Au 5.8828 1063.905 664.633 1065.821 1065.961
Au 6.0748 1084.511 664.639 1086.490 1086.631
Au 6.0748 1084.508 664.648 1086.489 1086.629
3.8 POLYNOMIAL FITTING OF THE CALIBRATION DATA The Chebyshev polynomial coefficients for the current/temperature relationship of the second calibration run for each lamp are given in Table 17. The fits for all three sets of data were good, the largest residuals being equivalent to a temperature uncertainty of 0.05 °C.
Restricted-Commercial NPL Report CBTM S7
17
Table 17 - Chebyshev coefficients for the fit of the second calibration run
4. UNCERTAINTIES The measurement uncertainties, evaluated at a level of approximately 95% confidence, are given in Tables 18 and 19. Table 18 - Uncertainty in the realisation of the fixed points
Table 19 - Uncertainty in the calibration of the lamps starting from the fixed points
Source of uncertainty Type Uncertainty (°C)
962°C 1064°C 1300°C 1500°C 1700°C
Realisation of fixed point B 0.014 0.016 - - -
Uncertainty from previous measurements (propagated)
- - 0.13 0.20 0.27
Lamp radiance temperature: statistical reproducibility resolution of DVM drift in reference lamp during comparison
A A B B
0.010 0.030 0.001 0.050
0.005 0.010 0.001 0.050
0.005 0.010 0.001 0.050
0.005 0.020 0.001 0.050
0.005 0.020 0.001 0.050
Current measurements: reproducibility current stability calibration of DVM resolution of DVM calibration of standard resistor current correction
A A B B B B
N/A 0.030 0.060 0.010 0.004 0.005
N/A 0.030 0.060 0.010 0.003 0.005
N/A 0.020 0.050 0.010 0.003 0.000
N/A 0.010 0.040 0.010 0.004 0.000
N/A 0.010 0.040 0.010 0.005 0.000
Base temperature: calibration of DVM resolution of DVM calibration of thermocouple measurement of BTC (10%)
B B B B
0.007 0.001 0.007 0.001
0.003 0.001 0.003 0.000
- - - -
- - - -
- - - -
Interference filter/wavelength: calibration of filter temperature coefficient of filter
B B
- -
- -
0.007 0.003
0.010 0.003
0.012 0.003
Alignment of sources B 0.020 0.020 0.020 0.020 0.020
Size-of-source effect Detector linearity
B B
0.03 N/A
0.03 N/A
0.04 N/A
0.05 N/A
0.07 N/A
Quality of polynomial fit A 0.050 0.050 0.050 0.050 0.050
Total 1σ (excluding wavelength conversion)
0.11 0.10 0.17 0.23 0.29
Conversion to 650 nm due to equation B 0.09 0.11 0.15 0.21 0.26
Total 1σ 0.14 0.15 0.22 0.31 0.39
Total 2σ 0.29 0.30 0.45 0.62 0.79
Restricted-Commercial NPL Report CBTM S7
19
5. REFERENCES [1] McEvoy H.C., Chattle M.V., Butler J., Metrologia, 1996, 33, 353-362 [2] Coates P.B., Inst. Phys. Conf. Ser. No. 26, 1975, 238-243 [3] Coates P. B., Andrews J. W., In Temperature: Its Measurement and Control in
Science and Industry, Vol. 5 (Edited by J. F. Schooley), New York, American Institute of Physics, 1982, 109-114
[4] Chu B., McEvoy H. C., Andrews J. W., Meas. Sci. Technol., 1994, 5, 12-19 [5] Kostkowski H. J., Lee R. D., In Temperature: Its Measurement and Control in
Science and Industry, Vol. 3 (Edited by C. M. Herzfeld), New York, Reinhold, 1962, 449-481
[6] Kostkowski H. J., Lee R. D., US Nat. Bur. Stand. (US) Monograph. 41, 1962, [7] Coates P. B., High Temp.-High Press., 1979, 11, 289-300 [8] Coates P. B., Metrologia, 1977, 13, 1-5 [9] Coates P. B., NPL Report QU 62, March 1981 [10] Bloembergen P., Duan Y., Bosma R., Yuan Z., "The characterization of Radiation
Thermometers subject to the size-of-source effect", In Proceedings of TEMPMEKO '96 (Edited by Piero Marcarino), Levrotto & Bella, 1997 IMEKO TC 12, 261-266
RESTRICTED - COMMERCIAL NPL Report CBTM S13
CONTENTS
Page
1. DESCRIPTION OF THE PYROMETER..................................................................................... 1
2. MEASUREMENTS PERFORMED ON THE LAMPS.............................................................. 1 2.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE
OF THE LAMP FILAMENT..................................................................................... 1 2.2 MEASUREMENTS PERFORMED ON THE LAMPS .................................................. 1 2.3 CHECK OF LAMP CALIBRATION AT THE SILVER POINT .................................. 2 2.4 CHECK OF LAMP C840 AT THE GOLD POINT........................................................ 2 2.5 RE-MEASUREMENT OF RAMB ....................................................................................... 2
3. RESULTS OF THE MEASUREMENTS...................................................................................... 2 3.1 RESULTS OF THE MEASUREMENTS OF RAMB .......................................................... 2 3.2 RESTABILISATION OF THE LAMPS ........................................................................... 3 3.3 RESULTS OF THE CALIBRATION CHECKS ON C860 AND C864 ........................ 3 3.4 RESULTS OF THE FIXED POINT CHECKS WITH THE LAMPS............................. 6 3.5 POLYNOMIAL FITTING OF THE CALIBRATION DATA....................................... 6
CCT key comparison: ITS-90 from 962 °°°°C to 1700 °°°°C NPL measurements, June 1998
by
H C McEvoy and K M Raven
1. DESCRIPTION OF THE PYROMETER Since the initial calibration of the lamps during July-August 1997, there have been two changes to the NPL Primary Pyrometer. Firstly, following a leakage of water into the laboratory, the silver coating on the mirrors was found to be badly tarnished. Two new mirrors, one off-axis ellipsoid and one plane mirror, were coated with a protected gold coating, then placed in the pyrometer. Secondly, the interference filters in the pyrometer were re-calibrated in-situ, using a monochromator and line sources. The transmission wavelength of the 665 nm filter was found to have shifted downwards by approximately 0.3 nm from the previous calibration, more than likely due to the filter having been re-angled prior to its calibration to avoid inter-reflections. 2. MEASUREMENTS PERFORMED ON THE LAMPS 2.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE OF THE
LAMP FILAMENT Before any measurements were performed on the lamps, Ramb was measured with the lamps still in the box and the lid closed as much as possible. The measurements were performed using an ASL F18 bridge (see NPL report CBTM S7 for details; however, for these measurements the gain was set to x104). The results are given in Section 3.1 Table 1.
2.2 MEASUREMENTS PERFORMED ON THE LAMPS The lamps were set up in front of the NPL Primary Pyrometer as before. Details can be found in NPL report CBTM S7. The front window of each lamp was cleaned with a dry lens tissue before the first and second measurement runs. They were also cleaned of dust particles regularly throughout both measurement runs. The measurements were performed using lamp C840 as the reference. This had been calibrated at the same time as the other two lamps (during July/August 1997), but had been kept on the shelf as a reserve in case of breakage of C860 or C864 during the intercomparison. Its calibration was assumed not to have drifted. Firstly, C860 and C864 were re-stabilised as described in the protocol. They were measured at 1100 °C using C840 as the reference, re-stabilised at 1700 °C for one hour, then re-measured at 1100 °C. The results of the re-stabilisation are shown in Table 2. The calibrations of lamps C860 and C864 were checked at each of the temperatures specified in the protocol. At each temperature, the lamps were set at the required current taken from the current/temperature table derived from the curve fit of the second calibration run performed during July-August 1997. Using C840 as the reference the radiance temperature of each of the other two lamps was determined. Two complete calibration checks were performed from 962 °C to 1700 °C. The wavelength at which the measurements were made was approximately 664.30 nm. The results of the measurements are shown in Section 3.3 Tables 3 to 6.
RESTRICTED - COMMERCIAL NPL Report CBTM S13
2
2.3 CHECK OF LAMP CALIBRATION AT THE SILVER POINT Additionally, the calibration of all three lamps was checked at 962 °C using the NPL silver point blackbody as the reference source. The size-of-source effect (SSE) of the pyrometer was measured immediately after the lamp measurements using a large area heat-pipe blackbody and a set of apertures, as described in NPL report CBTM S7. The SSE versus aperture diameter was fitted using the expression: y = a + b[exp(cx)] (1) where y is the SSE relative to a 25 mm diameter aperture, x is the aperture diameter in mm, and a = 99.956108, b= -6.694180 x10-1, c = -6.017013 x10-1. This expression was used to calculate the effective diameters for the silver point and lamps in the same way as before, i.e. allowing for the thermal profile of the furnace and the effect of the strip-shaped lamp filament. They were found to be 8.09 mm and 2.71 mm respectively for the silver point and lamp. This leads to a correction in the lamp radiance of +0.17 °C at 962 °C. The results were then corrected to a radiance temperature of 650 nm using the expression provided in the protocol. Table 7details these results.
2.4 CHECK OF LAMP C840 AT THE GOLD POINT As an additional check of the calibration of lamp C840, it was compared with the NPL gold point blackbody source. The other lamps were not checked as they had already been transported to the next laboratory in the circulation. Eight measurements of the lamp were made. The average results are given in Table 7. The effective diameter of the gold point was calculated using Equation (1) above and was found to be 10.58 mm. This leads to a correction in the lamp radiance of +0.22 °C at 1064 °C. 2.5 RE-MEASUREMENT OF RAMB Before the lamps were transported to their next destination, Ramb was measured as before. The results of these measurements are shown in Section 3.1, Table 1. 3. RESULTS OF THE MEASUREMENTS 3.1 RESULTS OF THE MEASUREMENTS OF RAMB Table 1
All the self-heating effect values are insignificant (< 10-5 Ω).
RESTRICTED - COMMERCIAL NPL Report CBTM S13
3
3.2 RESTABILISATION OF THE LAMPS Table 2
Lamp Number Current (A) Radiance temperature before
restabilisation (°C)
Radiance temperature after
restabilisation (°C)
Difference (°C)
C860 6.2845 1097.871 1098.124 0.253*
C864 6.1215 1100.020 1100.011 0.009 *Note: a small mark was seen on the front window of C860, just within the field of view of the pyrometer, when the lamp was turned up to 1700 °C. This was not cleaned off until after the stabilisation as relative measurements were required. It is likely that this mark influenced the repeatability of the measurements because, after cleaning the lamp, its radiance temperature was re-measured and found to be 1099.957 °C. Therefore, the stability of C860 wasn’t demonstrated by these results. The front windows of both lamps were cleaned thoroughly before any further measurements were performed.
3.3 RESULTS OF THE CALIBRATION CHECKS ON C860 AND C864 Tables 3 to 6 give the results of the calibration checks of the lamps using C840 as the reference. For each current, the radiance temperature from the curve fit of the second calibration run performed during 1997 is compared to that determined in June 1998. All radiance temperatures in the second and third columns of the Tables have been corrected to a wavelength of 650 nm. The differences in the two calibrations are given in the last column. The temperature and relative humidity during the first measurement run were 21.4 (± 0.4) °C and 31.1 (± 2.8) % respectively; the values for the second run were 22.0 (± 2.3) °C and 39.0 (± 3.4) % respectively. The total burning time for the lamps, including the restabilisation and silver point measurements, was 48 hours. The burning time during the gold point measurements with C840 was approximately 15 hours.
RESTRICTED - COMMERCIAL NPL Report CBTM S13
4
Table 3 - C860, 1st measurement run
I (A)
Temperature from 1997 2nd calibration run
(°C)
Temperature using C840 as reference (1998)
(°C)
Difference 1998-1997
(°C)
5.0711 961.910 961.435 -0.475
5.3803 1000.155 999.698 -0.457
5.9465 1064.380 1064.162 -0.218
6.1381 1084.853 1084.653 -0.200
6.2847 1100.177 1100.061 -0.116
7.2977 1200.160 1200.082 -0.078
7.2977(rpt) 1200.160 1200.063 -0.097
8.3979 1300.167 1300.140 -0.027
9.5690 1400.167 1400.153 -0.014
10.8046 1500.206 1500.247 +0.041
12.0977 1600.205 1600.171 -0.034
13.4452 1700.234 1700.230 -0.004
Table 4 - C860, 2nd measurement run
I (A)
Temperature from 1997 2nd calibration run
(°C)
Temperature using C840 as reference (1998)
(°C)
Difference 1998-1997
(°C)
5.0707 961.859 961.686 -0.173
5.3803 1000.155 999.965 -0.190
5.3810(rpt) 1000.238 999.971 -0.267
5.9464 1064.370 1064.310 -0.060
6.1379 1084.832 1084.810 -0.022
6.2845 1100.156 1100.168 +0.012
7.2976 1200.151 1200.129 -0.022
8.3983 1300.202 1300.204 +0.002
8.3985(rpt) 1300.219 1300.243 +0.024
9.5692 1400.183 1400.193 +0.010
10.8047 1500.214 1500.356 +0.142
12.0976 1600.197 1600.273 +0.076
13.4451 1700.226 1700.296 +0.070
RESTRICTED - COMMERCIAL NPL Report CBTM S13
5
Table 5 - C864, 1st measurement run
I (A)
Temperature from 1997 2nd calibration run
(°C)
Temperature using C840 as reference (1998)
(°C)
Difference 1998-1997
(°C)
4.9318 961.974 961.335 -0.639
5.2374 1000.293 999.810 -0.483
5.7911 1064.456 1064.156 -0.300
5.9787 1084.946 1084.710 -0.236
6.1216 1100.277 1100.089 -0.188
7.1089 1200.282 1200.224 -0.058
7.1089(rpt) 1200.282 1200.204 -0.078
8.1786 1300.324 1300.284 -0.040
9.3157 1400.316 1400.279 -0.037
10.5145 1500.366 1500.333 -0.033
11.7688 1600.375 1600.266 -0.109
13.0753 1700.398 1700.295 -0.103
Table 6 - C864, 2nd measurement run
I (A)
Temperature from 1997 2nd calibration run
(°C)
Temperature using C840 as reference (1998)
(°C)
Difference 1998-1997
(°C)
4.9314 961.961 961.483 -0.478
5.2373 1000.280 999.964 -0.316
5.2370(rpt) 1000.244 999.982 -0.262
5.7909 1064.433 1064.218 -0.215
5.9787 1084.946 1084.808 -0.138
6.1216 1100.277 1100.149 -0.128
7.1086 1200.252 1200.169 -0.083
8.1786 1300.324 1300.221 -0.103
8.1785(rpt) 1300.315 1300.226 -0.089
9.3156 1400.308 1400.148 -0.160
10.5144 1500.358 1500.308 -0.050
11.7688 1600.375 1600.253 -0.122
13.0752 1700.391 1700.295 -0.096
RESTRICTED - COMMERCIAL NPL Report CBTM S13
6
3.4 RESULTS OF THE FIXED POINT CHECKS WITH THE LAMPS Table 7 gives the average results of the measurements made with all three lamps using the silver point blackbody source as the reference, and the average result of the measurements of C840 using the gold point blackbody source. In the Table, the third column gives the radiance temperature at a pyrometer reference wavelength 664.35 nm; the fourth, the radiance temperature corrected to 650 nm. The sixth column gives the radiance temperature corrected for SSE. The radiance temperature from the curve fit of the second calibration performed in 1997, and the difference between this and the 1998 determination, are given in the last two columns.
Table 7 - Calibration checks using the Ag and Au point blackbody sources
The difference between the 1998 and 1997 calibrations of C840 at the Ag point is consistent with the value obtained at the Au point within the measurement uncertainty. For ease of comparison, the differences between the 1997 and 1998 measurements for each lamp are shown in Figures 1 to 3. These Figures show the differences between (i) the 1997 raw calibration data and the 1997 curve fit; (ii) the 1998 first measurement run and the 1997 curve fit; (iii) the 1998 second measurement run and the 1997 curve fit; and (iv) the 1998 measurements using the fixed point(s) and the 1997 curve fit. Figure 4 shows the difference between the calibration ‘shift’ of C864 and that of C860 for both 1998 measurement runs.
3.5 POLYNOMIAL FITTING OF THE CALIBRATION DATA This was not performed for these measurements, as the aim was to measure the drift in the calibration of the lamps since July/August 1998 rather than to provide a re-calibration.
4. MEASUREMENT UNCERTAINTIES The uncertainties of the measurements are given in tables 8 and 9. They include the calibration uncertainty of the reference lamp, which includes a component to allow for possible drift of the calibration since August 1997. The uncertainties of the correction to 650 nm and the SSE can be ignored for all measurements using the reference lamp: for the former, relative measurements are being made; for the latter the sources being compared are very similar. The uncertainty in the 1997 calibration of C860 and C864 is also given, along with the total 2s combined uncertainty of the 1997 and 1998 measurements.
RESTRICTED - COMMERCIAL NPL Report CBTM S13
7
Table 8 - Uncertainty in the realisation of the silver/gold point
Table 9 - Uncertainty in the 1998 lamp measurements, C840 as reference
Source of uncertainty Uncertainty (°C)
962°C* 962°C 1064°C 1300°C 1500°C 1700°C
Realisation of fixed point 0.023 - (0.023) 4 - - -
Calibration of C840 at 664nm 1 - 0.15 0.15 0.18 0.24 0.30
Lamp radiance temperature: statistical reproducibility resolution of DVM drift in reference lamp during comparison
0.010 0.030 0.001 0.010
0.010 0.030 0.001 0.010
0.005 0.010 0.001 0.010
0.005 0.010 0.001 0.010
0.005 0.020 0.001 0.010
0.005 0.020 0.001 0.010
Reproducibility between the two measurement runs2
0.090 0.090 0.050 - - -
Current measurements: reproducibility current stability calibration of DVM resolution of DVM calibration of standard resistor current interpolation
N/A 0.030 0.060 0.010 0.004 0.010
N/A 0.030 0.060 0.010 0.004 0.010
N/A 0.030 0.060 0.010 0.003 0.010
N/A 0.020 0.050 0.010 0.003 0.010
N/A 0.010 0.040 0.010 0.004 0.010
N/A 0.010 0.040 0.010 0.005 0.010
Base temperature: calibration of DVM resolution of DVM calibration of thermocouple measurement of BTC (10%)
0.007 0.001 0.007 0.001
0.007 0.001 0.007 0.001
0.003 0.001 0.003 0.000
- - - -
- - - -
- - - -
Filter wavelength: calibration of filter temperature coefficient of filter
0.010 0.010
0.010 0.010
0.010 0.010
0.010 0.015
0.015 0.020
0.020 0.025
Alignment of sources 0.020 0.020 0.020 0.020 0.020 0.020
Quality of polynomial fit N/A N/A N/A N/A N/A N/A
Size-of-source effect Detector linearity
0.030 N/A
N/A N/A
(0.030) 4 N/A
N/A N/A
N/A N/A
N/A N/A
Total 1s uncertainty 0.13 0.19 0.17
(0.10)4
0.19 0.25 0.31
1s uncertainty of 1997 calibration of C860 and C864 at 650 nm3
0.14 0.14 0.15 0.22 0.31 0.39
Combined 2s uncertainty of 1997 and 1998 measurements
0.38 0.47 0.45
(0.36) 4
0.58 0.80 0.99
RESTRICTED - COMMERCIAL NPL Report CBTM S13
9
Notes for Table of Uncertainties, Table 9: * Using the NPL silver point blackbody source as the reference;
1 Includes an allowance for drift in the calibration since August 1997. Also note that this is the calibration uncertainty at 664 nm: the uncertainty of the conversion to 650 nm is not included;
2 To allow for the differences between the two 1998 runs at the lower temperatures. Above the gold point, the differences are well within the measurement uncertainties;
3 From Table 19 in NPL Report CBTM S7; 4 Using the NPL gold point blackbody source as the reference.
5. CONCLUSION The results of the calibration check (Tables 3 to 7) show that the calibrations of lamps C860 and C864 have changed since the original calibrations in July/August 1997. The changes are within the 2s total combined uncertainty of the measurements at 1064 °C and above, but well outside the combined uncertainties at 962 °C. The measurements against the Ag point confirm these changes. The Ag and Au point measurements also show that the calibration of lamp C840 has drifted by a relatively large amount at these temperatures, despite it not having been used since it was calibrated. The reason for this is not clear. For C860 and C864, the changes are likely to be due to drift in the calibration of the lamps after a number of hours of use and transport between several laboratories. Calibration drift is usually larger at the higher temperatures, which is not observed here. However, the measurements were performed using C840 as the reference, and, as this appears to have drifted, this would naturally influence the measurement results. When the lamps are returned to NPL following the next stage of the circulation, they will be fully re-calibrated using the radiance doubling technique. Until then it is recommended that the 1997 calibration of the lamps is used as the reference, and all measurements are compared with that.
NPL Report CBTM S31
CCT Key Comparison: ITS-90
from 962 °°°°C to 1700 °°°°C,
NPL Measurements with VSL
Lamps, October 1998
H C McEvoy & K M Raven
December 1998
RESTRICTED COMMERCIAL
RESTRICTED-COMMERCIAL NPL Report CBTM S31
CCT Key Comparison: ITS-90 from 962 °°°°C to 1700 °°°°C,
NPL Measurements with VSL Lamps, October 1998
H C McEvoy and K M Raven
ABSTRACT
This report describes the measurements performed at NPL with the VSL high-stability lamps, numbers C564 and C681. The work was carried out during October 1998. For a detailed description of the NPL Primary Pyrometer, and for further details of the measurement techniques, refer to NPL Reports numbered CBTM S7 and CBTM S13.
National Physical Laboratory Queens Road, Teddington, Middlesex, TW11 0LW
This Report is supplied restricted commercial Extracts from this report may be reproduced provided the source is acknowledged
Approved on behalf of the Managing Director, NPL by Dr D W Robinson, Centre for Basic and Thermal Metrology
RESTRICTED-COMMERCIAL NPL Report CBTM S31
CONTENTS
Page 1. MEASUREMENTS PERFORMED ON THE LAMPS ........................................... 1
1.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE OF THE LAMP FILAMENT............................................................................. 1
1.2 SETTING UP THE LAMPS.............................................................................. 1 1.3 POSITIONAL EFFECT CHECKS .................................................................... 1 1.4 RESTABILISATION OF LAMPS .................................................................... 2 1.5 CALIBRATION OF THE LAMPS.................................................................... 2 1.6 SIZE-OF-SOURCE EFFECT MEASUREMENTS........................................... 3 1.7 RE-MEASUREMENT OF RAMB....................................................................... 3
2. RESULTS OF THE MEASUREMENTS.................................................................. 3
2.1 RESULTS OF THE MEASUREMENTS OF RAMB.......................................... 3 2.2 RESULTS OF THE POSITIONAL EFFECT CHECKS................................... 3 2.3 RESTABILISATION OF THE LAMPS............................................................ 4 2.4 CALIBRATION RESULTS FOR C681 AND C564......................................... 4 2.5 RESULTS OF THE FIXED POINT MEASUREMENTS ................................ 5 2.6 POLYNOMIAL FITTING OF THE CALIBRATION DATA .......................... 5
CCT key comparison: ITS-90 from 962 °C to 1700 °C NPL measurements with VSL lamps, October 1998
by
H C McEvoy and K M Raven
The following describes the measurements performed at NPL with the VSL high-stability lamps, numbers C564 and C681. The work was carried out during October 1998. For a detailed description of the NPL Primary Pyrometer, and for further details of the measurement techniques, refer to NPL Reports numbered CBTM S7 and CBTM S13. 1. MEASUREMENTS PERFORMED ON THE LAMPS 1.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE OF
THE LAMP FILAMENT Before any measurements were performed on the lamps, Ramb was measured with the lamps still in the box and the lid closed as much as possible. The measurements were performed using an ASL F18 bridge (see NPL report CBTM S13 for details), and using the leads supplied with the lamps. The results are given in Section 2.1 Table 1. 1.2 SETTING UP THE LAMPS The lamps were set up in front of the NPL Primary Pyrometer as before (see NPL report CBTM S7, Section 2.3). Before the first and second measurement runs the front window of each lamp was cleaned with a few drops of ethanol, then polished thoroughly with a dry lens tissue. The front windows were cleaned of dust particles regularly throughout both measurement runs. The lamps were calibrated using a radiance doubling technique, as described in Report CBTM S7, using the NPL Ag and Au fixed point blackbody sources. Firstly, though, the following initial measurements were made. 1.3 POSITIONAL EFFECT CHECKS With the lamps at approximately 1100 °C the following positional checks were performed: i) The pyrometer was scanned across the filament, at the height of the notch, in steps of
0.125 mm to each edge of the filament. This was done to assess the horizontal radiance distribution.
ii) The filament was rotated about the vertical axis in 1° steps up to ± 10° from the normal
alignment position. The results of these checks can be found in Section 2.2.
RESTRICTED-COMMERCIAL NPL Report CBTM S31
2
1.4 RESTABILISATION OF LAMPS The lamps were re-stabilised as described in the protocol, by measuring the radiance temperature at approximately 1100 °C, turning them up to 1700 °C for one hour, then re-measuring the radiance temperature at 1100 °C. The results of the measurements can be found in Section 2.3. 1.5 CALIBRATION OF THE LAMPS The lamps were calibrated over the range 962 °C to 1700 °C. At 962 °C, 1000 °C, and 1064 °C they were calibrated by direct comparison with either the Ag or Au fixed point blackbody source. Several melts and freezes were performed for each lamp temperature, and the average result was obtained for each lamp. Above 1064 °C, the calibration was carried out using a radiance doubling technique, using the measurements at the gold point as the reference. However, additional points were included so that measurements were made at all the temperatures defined in the protocol. The calibration was carried out twice. At each temperature, the measurements for each lamp were corrected to a particular current using typical current/temperature relationships for high-stability lamps with, respectively, 1.5 mm and 1.3 mm wide filaments. As the current corrections were small, this did not introduce any significant error. Once the first calibration had been performed, the results were curve-fitted using a sixth order Chebyshev polynomial to provide more accurate current/temperature relationships for the lamps. At and below 1200 °C for lamp number C681, and 1300 °C for lamp number C564, the results were corrected to a base temperature of 20 °C using the supplied polynomial expressions. The measurements were made at a wavelength of approximately 664.3 nm and corrected to 650 nm using the polynomial equation provided in the protocol. An additional measurement at the gold point was performed using a new filter combination with a transmission wavelength centred at approximately 657 nm. This filter combination had been calibrated in-situ as per NPL Report CBTM S7. It was used to measure the radiance temperature of the lamps at a wavelength closer to 650 nm, thus reducing the correction required. This filter was not used for the full calibration of the lamps as it was thought more important to use the same filter for all the measurements made during this CCT intercomparison. Throughout both calibrations, the maximum rate of increase or decrease of current was 1 A per minute. Overnight, the current to both lamps was turned off. The total burning time of the lamps was 48.6 hours for the first run, including the positional and stability checks, and 32.9 hours for the second run. The results of the measurements are given in Tables 3 to 6. The results of all the measurements made against the Ag and Au fixed-point blackbodies are given in Tables 7 to 10.
RESTRICTED-COMMERCIAL NPL Report CBTM S31
3
1.6 SIZE-OF-SOURCE EFFECT MEASUREMENTS The size-of-source effect (SSE) of the pyrometer was measured in the same way as before (see NPL Reports CBTM S7 and S13). The SSE versus aperture diameter data was fitted using the expression: y = a + b[exp(cx)] (1) where y is the SSE relative to a 25 mm diameter aperture, x is the aperture diameter in mm, and a = 99.9561, b= -0.6694, c = -0.6017. This expression was used to calculate the effective diameters for the silver point, gold point and lamps in the same way as before, i.e. allowing for the thermal profile of the furnace and the effect of the strip-shaped lamp filament. The effective diameters were found to be 10.57 mm and 8.09 mm respectively for the gold point and silver point, and 2.71 mm and 2.40 mm respectively for lamp C681 and lamp C564. This leads to a correction in the lamp radiance of +0.17 °C at 962 °C and +0.22 °C at 1064 °C for lamp C681, and +0.19 °C at 962 °C and +0.25 °C at 1064 °C for lamp C564. The SSE correction at higher temperatures may be found by extrapolation. 1.7 RE-MEASUREMENT OF RAMB Before the lamps were transported to their next destination, Ramb was measured again. The results of these measurements are shown in Section 2.1, Table 1. 2. RESULTS OF THE MEASUREMENTS 2.1 RESULTS OF THE MEASUREMENTS OF RAMB Table 1
All the self-heating effect values are insignificant (< 10-5 Ω). 2.2 RESULTS OF THE POSITIONAL EFFECT CHECKS i) Scanning the pyrometer horizontally across the filament of C681 showed a region of
±0.125 mm around the normal alignment position where the radiance temperature varied by < 0.05 °C, and a region of ±0.25 mm where the radiance temperature varied by < 0.4 °C. The target size of lamp C681 was considered sufficient to fill the field-of-view of the pyrometer. For lamp C564, the radiance temperature varied by < 0.05 °C over a region of ±0.125 mm, but by > 2 °C over a region of ±0.25 mm from the normal
RESTRICTED-COMMERCIAL NPL Report CBTM S31
4
alignment position. This was due to the lamp having a narrower filament: at a distance of 0.25 mm from the normal alignment position, the edge of the filament is very close to the edge of the 0.75 mm field-of-view of the pyrometer.
ii) The results of the rotation of each lamp about the vertical axis are shown in Figures 1
and 2. The angular alignment position of C564 is close to the observed radiance peak and hence is critical. The angular alignment of both lamps was checked before each measurement at the lower temperatures, and regularly at the highest temperature where the brightness of the filament made the check more difficult. This would have minimised the errors due to mis-alignment.
Since the width of the filament of C564 is less than twice the nominal field-of-view of the pyrometer this will result in higher correction due to the SSE (see Section 1.6 above), and also make the measurements more susceptible to alignment errors. Additionally, errors will occur due to the angular alignment position being very close to the observed radiance peak ((ii) above). These factors have been taken into account when assessing the alignment uncertainty in Table 13. 2.3 RESTABILISATION OF THE LAMPS The results of the restabilisation measurements are given in the following Table. Table 2
Lamp Number Current (A) Radiance temperature before
restabilisation (°C)
Radiance temperature after
restabilisation (°C)
Difference (°C)
C681 6.7486 1099.857 1099.859 0.002
C564 5.4446 1100.012 1099.871 0.141 The difference in radiance temperature of lamp C681 is insignificant. That of C564 is significant and greater than the 0.05 °C specified in the protocol. This difference could be due to a genuine drift, or it could partly be a result of the lamp’s greater sensitivity to angular alignment. 2.4 CALIBRATION RESULTS FOR C681 AND C564 Tables 3 to 6 show the results of both calibration runs. The third column gives the base temperature coefficient applied at each temperature, while the fourth gives the measured lamp radiance temperature corrected to a base temperature of 20 °C. Corrections were applied to allow for the pyrometer’s SSE (6th column) and to convert the results to a wavelength of 650 nm (9th column). The 10th column gives the reference current as defined in the instructions sent with the lamps. The 11th column gives the change in lamp current per °C change in radiance temperature, used to correct the results in column 9 to the reference current. The last column in the Tables gives the final corrected radiance temperature for each lamp current. Note that, in the Tables, λ is the reference wavelength of the pyrometer from the filter calibration; it is not the effective wavelength λe.
RESTRICTED-COMMERCIAL NPL Report CBTM S31
5
The room temperature and humidity during the first calibration run were (21.2 ± 1.8) °C and (34.7 ± 4.5) % respectively; during the second calibration run they were (21.1 ± 1.1) °C and (39.1 ± 4.5) % respectively. The maximum and minimum room temperature values were 22.9 °C and 18.4 °C respectively for the first calibration run and 23.7 °C and 18.6 °C respectively for the second calibration run. The maximum and minimum relative humidity values were 43.0 % and 26.8 % respectively for the first run and 49.3 % and 27.2 % respectively for the second calibration run. 2.5 RESULTS OF THE FIXED POINT MEASUREMENTS Tables 7 to 10 give the results of the measurements made with the lamps using the silver and gold point blackbody sources as the reference. In the Tables, the fourth column gives the measured radiance temperature, corrected to a base temperature of 20 °C, at a pyrometer reference wavelength of approximately 664.3 nm. The seventh column gives the radiance temperature corrected for SSE, and the eighth gives the radiance temperature corrected to 650 nm. The last column gives the corrected radiance temperature of the lamp at the reference current. The average of the fixed point values have been included in Tables 3 to 6. 2.6 POLYNOMIAL FITTING OF THE CALIBRATION DATA For both lamps, the final corrected radiance temperature at each lamp current (last column in Tables 3 to 6 were fitted using a 6th order Chebyshev polynomial equation. This was used to derive the current temperature relationship for the lamps. The Chebyshev polynomial coefficients for the first calibration run are given in Table 11. The fits for both sets of data were good, with the largest residuals being equivalent to a temperature uncertainty of 0.08 °C.
3. MEASUREMENT UNCERTAINTIES The measurement uncertainties, evaluated at a level of confidence of approximately 95%, are given in Tables 12 and 13.
4. CONCLUSION The results of the two calibration runs for each lamp agreed to well within the measurement uncertainties. The differences between the first and second runs are plotted in Figures 3 and 4. For the first calibration run, the measurements made at the gold point at 657 nm agreed with those made at 664.3 nm to within 0.1 °C. This gives us confidence that the wavelength correction used is valid for these lamps to within 10%. The measurements for lamp C564 showed more scatter/non-repeatability then those for C681. This is more than likely due to it having a narrower filament making it more sensitive to small alignment errors. The non-repeatability impacts on the calibration of both lamps and has been included in the uncertainty budget. It is, in any case, a relatively small component in the uncertainty.
RESTRICTED-COMMERCIAL NPL Report CBTM S31
10
Table 7 - Measurements of C564 using fixed points, first calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 4.4829 0.99 962.587 664.277 0.19 962.777 964.364 4.480 963.89
Ag 4.4829 0.99 962.629 664.287 0.19 962.819 964.407 4.480 963.93
Ag 4.7240 0.59 1000.456 664.318 0.20 1000.656 1002.350 4.721 1001.90
Au 5.1723 1.00 1064.409 664.335 0.25 1064.659 1066.538 5.169 1066.09
Au 5.1723 1.00 1064.425 664.338 0.25 1064.675 1066.554 5.169 1066.11
Au 5.1723 0.98 1065.427 657.277 0.25 1065.677 1066.632 5.169 1066.19
Table 8 - Measurements of C564 using fixed points, second calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 4.4830 0.99 962.528 664.298 0.19 962.718 964.307 4.480 963.82
Ag 4.4830 0.99 962.595 664.304 0.19 962.785 964.375 4.480 963.88
Ag 4.7240 0.59 1000.424 664.308 0.20 1000.624 1002.316 4.721 1001.86
Au 5.1727 1.00 1064.483 664.339 0.25 1064.733 1066.613 5.169 1066.11
Au 5.1727 1.00 1064.499 664.348 0.25 1064.749 1066.630 5.169 1066.13
Table 9 - Measurements of C681 using fixed points, first calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 5.5104 1.00 961.819 664.277 0.17 961.989 963.574 5.508 963.27
Ag 5.5104 1.00 961.802 664.287 0.17 961.972 963.558 5.508 963.25
Ag 5.8250 0.59 1000.031 664.318 0.18 1000.211 1001.903 5.822 1001.55
Au 6.4024 1.00 1064.254 664.335 0.22 1064.474 1066.352 6.399 1065.99
Au 6.4024 1.00 1064.277 664.338 0.22 1064.497 1066.376 6.399 1066.02
Au 6.4024 0.99 1065.288 657.277 0.22 1065.508 1066.463 6.399 1066.11
RESTRICTED-COMMERCIAL NPL Report CBTM S31
11
Table 10 - Measurements of C681 using fixed points, second calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 5.5103 1.00 961.804 664.298 0.17 961.974 963.561 5.508 963.27
Ag 5.5103 1.00 961.789 664.304 0.17 961.959 963.547 5.508 963.26
Ag 5.8248 0.59 999.976 664.308 0.18 1000.156 1001.847 5.822 1001.52
Au 6.4024 1.00 1064.243 664.339 0.22 1064.463 1066.342 6.399 1065.98
Au 6.4024 1.00 1064.225 664.348 0.22 1064.445 1066.325 6.399 1065.97
Table 11 - Chebyshev coefficients for the fit of the first calibration run
CCT Key Comparison: ITS-90 from 962 °°°°C to 1700 °°°°C,
NPL Measurements, April to May 1999
H C McEvoy
ABSTRACT
This report describes the measurements performed at NPL with lamps numbered C860, C864 and C840 at the end of the second circulation of the CCT key comparison. The work was carried out during April and May 1999 under the United Kingdom’s Department of Trade and Industry Programme for Thermal Metrology, Milestone PT981.2.A.03. For a detailed description of the NPL Primary Pyrometer, and for further details of the measurement techniques, refer to NPL Reports numbered CBTM S7, CBTM S13 and CBTM S31.
National Physical Laboratory Queens Road, Teddington, Middlesex, TW11 0LW
This Report is supplied restricted commercial Extracts from this report may be reproduced provided the source is acknowledged
Approved on behalf of the Managing Director, NPL by Dr D W Robinson, Centre for Basic, Thermal & Length Metrology
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
CONTENTS
Page 1. MEASUREMENTS PERFORMED ON THE LAMPS ........................................... 1
1.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE OF THE LAMP FILAMENT............................................................................. 1
1.2 SETTING UP THE LAMPS.............................................................................. 1 1.3 RESTABILISATION OF THE LAMPS............................................................ 1 1.4 CALIBRATION OF THE LAMPS.................................................................... 1 1.5 SIZE-OF-SOURCE EFFECT MEASUREMENTS........................................... 2 1.6 RE-MEASUREMENT OF RAMB....................................................................... 3
2. RESULTS OF THE MEASUREMENTS.................................................................. 3
2.1 RESULTS OF THE MEASUREMENTS OF RAMB.......................................... 3 2.2 RESTABILISATION OF THE LAMPS............................................................ 3 2.3 CALIBRATION RESULTS FOR THE LAMPS............................................... 4 2.4 RESULTS OF THE FIXED POINT MEASUREMENTS ................................ 4 2.5 POLYNOMIAL FITTING OF THE CALIBRATION DATA .......................... 5
CCT key comparison: ITS-90 from 962 °C to 1700 °C NPL measurements, April to May 1999
by
H C McEvoy
The following describes the measurements performed at NPL with lamps numbered C860, C864 and C840 at the end of the second circulation of the CCT key comparison. The work was carried out during April and May 1999. For a detailed description of the NPL Primary Pyrometer, and for further details of the measurement techniques, refer to NPL Reports numbered CBTM S7, CBTM S13 and CBTM S31. 1. MEASUREMENTS PERFORMED ON THE LAMPS 1.1 MEASUREMENT OF RAMB , THE ROOM TEMPERATURE RESISTANCE OF
THE LAMP FILAMENT Before any measurements were performed, Ramb was measured at 20 mA with the lamps in the case and the lid closed as much as possible. The measurements were performed using an ASL F18 bridge (see NPL report CBTM S13 for details). The results are given in Section 2.1 Table 1. 1.2 SETTING UP THE LAMPS The lamps numbered C860 and C864 were set up in front of the NPL Primary Pyrometer as before (see NPL report CBTM S7, Section 2.3). Before the first and second measurement runs the front window of each lamp was cleaned with a few drops of ethanol, then polished thoroughly with a dry lens tissue. The front windows were cleaned of dust particles regularly throughout both measurement runs. 1.3 RESTABILISATION OF THE LAMPS The lamps were re-stabilised as described in the protocol, by measuring the radiance temperature at approximately 1100 °C, turning them up to 1700 °C for one hour, then re-measuring the radiance temperature at 1100 °C. The results of the measurements can be found in Section 2.2. 1.4 CALIBRATION OF THE LAMPS a) Lamps C860 and C864 The lamps were calibrated over the range 962 °C to 1700 °C. At 962 °C, 1000 °C, and 1064 °C they were calibrated by direct comparison with either the Ag or Au fixed-point blackbody source. Several melts and freezes were performed for each lamp temperature, and the average result was obtained for each lamp. Above 1064 °C, the calibration was carried out using a radiance doubling technique, using the measurements at the gold point as the reference, as described in NPL Report CBTM S7. However, additional points were included so that measurements were made at all the temperatures defined in the protocol. The calibration was carried out twice.
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
2
At each temperature, the currents were set close to the reference values specified in the documentation sent with the lamps on the circulation. The measurements were corrected to the exact specified current using the current/temperature tables derived from the curve fits of the second calibration run performed during July-August 1997. b) Lamp C840 The purpose of lamp number C840 was to provide a back-up in case of breakage or failure of C860 or C864. It was not used in the circulation, except for the checks made on C860 and C864 during June 1998 (see NPL Report CBTM S13). However, in view of the results of these checks it was felt important to re-calibrate it. C840 was set up and restabilised as described in Sections 1.2 and 1.3 above. The results of the restabilisation checks can be found in Section 2.2. C840 was recalibrated during the second calibration of C860 and C864, at all temperatures up to 1200 °C, then at various temperatures up to 1700 °C. At 962°C, 1000 °C and 1064 °C, the calibration was performed by direct comparison with either the NPL Ag or Au fixed-point blackbody source. Above 1064 °C, C840 was calibrated by comparing it with either C860 or C864. In this way, a calibration could be performed without the lamp being directly involved in the bootstrap procedure. At each temperature the current was set close to a suitable value determined from the July-August 1997 calibration. The current/temperature relationship for the lamp was used to correct the radiance temperature to the exact chosen current so that the 1997 and 1999 calibrations could be easily compared. C840 was not calibrated a second time. c) Corrections to the lamp results At and below 1100 °C the results were corrected to a base temperature of 20 °C using the polynomial expressions derived during the July-August 1997 calibration. The measurements were made at a wavelength of approximately 664.3 nm and corrected to 650 nm using the polynomial equation provided in the protocol. Throughout both calibrations, the maximum rate of increase or decrease of current was 1 A per minute. Overnight, the current to both lamps was turned off. The total burning time of C860 and C864 for the first run, including the stability checks, was 40 hours. During the second calibration run, the total burning time was 38.5 hours for C860 and 34.5 hours for C864. The total burning time of lamp C840 including the stability checks was 32 hours. The results of the measurements are given in Tables 3 to 7. The results of all the measurements made against the Ag and Au fixed-point blackbodies are given in Tables 8 to 12. 1.5 SIZE-OF-SOURCE EFFECT MEASUREMENTS The size-of-source effect (SSE) of the pyrometer was measured in the same way as before using the NPL large area heat-pipe blackbody source, and set of apertures (see NPL Reports CBTM S7 and S13). The SSE versus aperture diameter data was fitted using the expression: y = a + b[exp(cx)] (1)
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
3
where y is the SSE relative to a 25 mm diameter aperture, x is the aperture diameter in mm, and a = 99.9633, b= -0.7607, c = -0.7613. This expression was used to calculate the effective diameters for the silver point, gold point and lamps in the same way as before, i.e. allowing for the thermal profile of the furnace and the effect of the strip-shaped lamp filament. The effective diameters were found to be 9.04 mm and 6.79 mm respectively for the gold point and silver point, and 2.59 mm for lamps. This leads to a correction in the lamp radiance of +0.18 °C at 962 °C and +0.23 °C at 1064 °C. The SSE correction at higher temperatures may be found by extrapolation. 1.6 RE-MEASUREMENT OF RAMB After the calibration had been completed, Ramb was measured again. The results of these measurements are shown in Section 2.1, Table 1. 2. RESULTS OF THE MEASUREMENTS 2.1 RESULTS OF THE MEASUREMENTS OF RAMB Table 1
All the self-heating effect values are insignificant (< 10-5 Ω). The differences between the pre- and post-calibration values of Ramb are likely to be due to the effect of the temperature coefficient of the filaments. 2.2 RESTABILISATION OF THE LAMPS The results of the restabilisation measurements are given in the following Table.
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
4
Table 2
Lamp Number Current (A) Radiance temperature before
restabilisation (°C)
Radiance temperature after
restabilisation (°C)
Difference (°C)
C860 6.1116 1082.237 1082.230 0.007
C864 6.1209 1100.038 1100.023 0.015
C840 6.2040 1100.549 1100.319 0.230 The differences in radiance temperature of lamps C860 and C864 are well within the stabilisation limits set by the protocol. That of lamp C840 is significant and greater than the 0.05 °C specified in the protocol. It is not clear why the lamp should have drifted by this amount. 2.3 CALIBRATION RESULTS FOR THE LAMPS Tables 3 to 7 show the results of both calibration runs. The third column gives the base temperature coefficient applied at each temperature below 1100 °C, while the fourth gives the measured lamp radiance temperature corrected to a base temperature of 20 °C. Corrections were applied to allow for the pyrometer’s SSE (fifth and sixth columns) and to convert the results to a wavelength of 650 nm (seventh and eighth columns). The ninth column gives the reference current: for C860 and C864 this was the value defined in the instructions sent with the lamps and for C840 it was the chosen current. The tenth column gives the change in lamp current per °C change in radiance temperature, used to correct the results in column eight to the reference current. The last column in the Tables gives the final corrected radiance temperature for each lamp current. Note that, in the Tables, λ is the reference wavelength of the pyrometer from the filter calibration; it is not the effective wavelength λe. For ease of comparison, Figures 1 to 2 show the differences between the first 1999 calibration and the second (reference) 1997 calibration for all three lamps, and the differences between the first and second 1999 calibrations for C860 and C864. The average room temperature and humidity during the first calibration run were 21.2 (±0.7) °C and 34.8 (±5.9) % respectively; during the second calibration run they were 21.9 (±0.8) °C and 29.9 (±4.7) % respectively. The maximum and minimum room temperature values were 22.8 °C and 20.1 C respectively for the first calibration run and 23.2 °C and 20.5 °C respectively for the second calibration run. The maximum and minimum relative humidity values were 46.6 % and 26.3 % respectively for the first run and 46.8 % and 22.9 % respectively for the second calibration run. 2.4 RESULTS OF THE FIXED POINT MEASUREMENTS Tables 8 to 12 give the results of the measurements made with the lamps using the Ag and Au point blackbody sources as the reference. In the Tables, the fourth column gives the measured radiance temperature, corrected to a base temperature of 20 °C, at the pyrometer reference wavelength shown in the fifth column. The seventh column gives the radiance temperature
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
5
corrected for SSE, and the eighth gives the radiance temperature corrected to 650 nm. The last column gives the corrected radiance temperature of the lamp at the reference current. The averages of the fixed point results have been included in Tables 3 to 7. 2.5 POLYNOMIAL FITTING OF THE CALIBRATION DATA For lamps C860 and C864, the final corrected radiance temperature at each lamp current (last column in Tables 3 to 6 were fitted using a 6th order Chebyshev polynomial equation. The Chebyshev polynomial coefficients for the first calibration run are given in Table 13. The fits for both sets of data were good, with the largest residuals being equivalent to a temperature uncertainty of 0.07 °C. The results for lamp C840 were not fitted since the calibration had not been carried out at all temperatures.
3. MEASUREMENT UNCERTAINTIES The measurement uncertainties, evaluated at a level of confidence of approximately 95%, are given in Tables 14 and 15.
4. CONCLUSION The results show that the calibration results of lamps C860 and C864 are in very good agreement with those obtained in 1997 before the start of the intercomparison. This is despite them having been transported between a number laboratories during that time, and being subjected to many hours of burning. Furthermore, the two 1999 calibrations are in excellent agreement. It can be concluded that, within the measurement uncertainties, the lamps have not drifted significantly since their initial calibration. In the event, the back-up lamp C840 was not required in the circulation. In these measurements its calibration agrees reasonably well with that performed in 1997.
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
11
Table 8 - Measurements of C860 using fixed points, first calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 5.0732 0.98 960.357 664.304 0.18 960.537 962.121 5.072 961.97
Ag 5.0732 0.98 960.357 664.307 0.18 960.537 962.121 5.072 961.97
Ag 5.3798 1.65 998.110 664.310 0.19 998.300 999.986 5.380 1000.01
Au 5.9454 0.98 1062.288 664.307 0.23 1062.518 1064.387 5.944 1064.23
Au 5.9454 0.98 1062.282 664.311 0.23 1062.512 1064.381 5.944 1064.23
Table 9 - Measurements of C860 using fixed points, second calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 5.0732 0.98 960.364 664.305 0.18 960.544 962.128 5.072 961.97
Ag 5.0732 0.98 960.361 664.318 0.18 960.541 962.126 5.072 961.97
Ag 5.3798 1.65 998.078 664.325 0.19 998.268 999.956 5.380 999.98
Au 5.9455 0.98 1062.264 664.317 0.23 1062.494 1064.364 5.944 1064.20
Au 5.9455 0.98 1062.268 664.326 0.23 1062.498 1064.369 5.944 1064.21
Table 10 - Measurements of C864 using fixed points, first calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 4.9338 0.98 960.177 664.304 0.18 960.357 961.940 4.933 961.84
Ag 4.9338 0.98 960.146 664.307 0.18 960.326 961.910 4.933 961.81
Ag 5.2367 1.65 998.065 664.310 0.19 998.255 999.941 5.236 999.86
Au 5.7890 0.97 1062.037 664.307 0.23 1062.267 1064.135 5.788 1064.02
Au 5.7890 0.97 1062.039 664.311 0.23 1062.269 1064.138 5.788 1064.03
RESTRICTED-COMMERCIAL NPL Report CBTLM S2
12
Table 11 - Measurements of C864 using fixed points, second calibration run
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 4.9336 0.98 960.029 664.305 0.18 960.209 961.792 4.933 961.71
Ag 4.9336 0.98 960.037 664.318 0.18 960.217 961.802 4.933 961.72
Ag 5.2367 1.65 997.979 664.325 0.19 998.169 999.857 5.236 999.77
Au 5.7890 0.97 1061.976 664.317 0.23 1062.206 1064.075 5.788 1063.96
Au 5.7890 0.97 1061.995 664.326 0.23 1062.225 1064.095 5.788 1063.98
Table 12 - Measurements of C840 using fixed points
Fixed point used
I (A)
Ratio of photo-
currents
Measured t (°C)
λ (nm) SSE correction
(°C)
t corrected for SSE
(°C)
t corrected to 650nm
(°C)
Reference current
(A)
t90 at reference current
(°C)
Ag 5.0031 0.99 960.764 664.305 0.18 960.944 962.529 5.002 962.39
Ag 5.0031 0.99 960.736 664.318 0.18 960.916 962.502 5.002 962.36
Ag 5.3087 1.66 998.525 664.325 0.19 998.715 1000.404 5.307 1000.20
Au 5.8658 0.98 1062.293 664.317 0.23 1062.523 1064.393 5.865 1064.31
Au 5.8658 0.98 1062.281 664.326 0.23 1062.511 1064.382 5.865 1064.29
Table 13 - Chebyshev coefficients for the fit of the first calibration run
Report on the calibration of the tungsten strip lamps C860 and C864 within the CCT key-comparison “Local realizations of the IST-90 between the Silver point and 1700 °C using vacuum Tungsten-strip lamps as transfer standards” By J. Hartmann
I. Experimental and theoretical procedure
The realisation of the ITS- 90, the description of the equipment and the experimental procedure are described in the following references:
1. J. Fischer, H.J. Jung, R. Friedrich, “A new determination of the freezing temperature of gold relative to that of silver by radiation thermometry“, Temperature 6, 53-57 (1992);
2. J. Fischer, H.J. Jung, "Determination of the thermodynamic temperatures of the freezing points of silver and gold by near-infrared pyrometry", Metrologia 26, 245-252 (1989);
The formal definition and derivation of the spectral radiance temperature with explicit reference to corrections applied together with the transfer of the radiance temperature to strip lamp is described in
3. J. Fischer, J. Hartmann, “Calibration of tungsten strip lamps as transfer standards for temperature” Proceedings of Tempmeko´99;
4. H.-J. Jung, J. Verch; “Ein Rechenverfahren zur Auswertung pyrometrischer Messungen“, Optik 38, 95-109 (1973)
In this report only a short description of the reference thermometer characteristics is given.
The limiting effective wavelength λe is calculated for every lamp at every current separately according Ref. 3 and 4 and is given in the final Tables 5 to 8 in the column λe. The measured beam is limited to a diameter of 20 mm. The reference pyrometer has a focus length of 300 mm yielding a f-number of 20/300=1/15. The target distance is 1220 mm and the target field is circular, with a diameter of 0.5 mm.
The size-of source-effect (SSE) with respect to a gold fixed-point blackbody with an aperture of 3 mm diameter (effective source diameter 30 mm) was measured for the two wavelengths (650 nm and 950 nm) for two different strip widths. The results are given in Table 1.
Wavelength / nm SSE for the 1.5 mm strip SSE for the 3 mm strip
650 6.89x10-4 5.26x10-4
950 7.13x10-4 4.24x10-4
Table 1: Size-of-source effect measured with respect to a gold fixed point with an aperture of 3 mm diameter (effective source diameter 30 mm) for two strip widths.
report_f.doc 27/06/2003 JH
The transfer lamps C860 and C864 The transfer lamps arrived at PTB on January 27th and left PTB on March 30th 1999. The conditions of the measurements together with the total burning times for both lamps are given in Table 2.
Lamp C860 Lamp C864 Orientation as prescribed in the protocol as prescribed in the protocol Base temperature TB 20 °C ± 0.1 °C 20 °C ± 0.1 °C Total burning time 65 h 64 h Ambient temperature (min) 22 °C 21 °C Ambient temperature (max) 23 °C 23 °C Ambient temperature (aver.) 22 °C 22 °C Relative humidity (min) 33% 35% Relative humidity (max) 38% 40% Relative humidity (average) 35% 37% Lamp resistor Rambient(begin) 0.04020985 Ω 0.04183677 Ω Temperature Tambient(begin) 23.6 °C 23.4 °C Lamp resistor Rambient(end) 0.04016385 Ω 0.04190437 Ω Temperature Tambient(end) 22.75 °C 22.55 °C
Table 2: Measurements conditions for the two lamps C860 and C864.
The measured horizontal and angular distributions of the strip radiance temperatures are presented in Fig. 1-2.
0.9975
0.9985
0.9995
1.0005
1.0015
1.0025
-1.00 -0.50 0.00 0.50 1.00distance / mm
norm
. Sig
nal
5.072A7.298A
0.9975
0.9985
0.9995
1.0005
1.0015
1.0025
-1.00 -0.50 0.00 0.50 1.00distance / mm
norm
. Sig
nal
4.933A5.788A9.314A11.767A
Figure 1: Horizontal distribution of the signal along the strip obtained at the lamp C860 (left) and C864 (right). The signals have been normalized to the signal in the middle of the strip. The nodge is located on the right side, i.e. in direction of positive x.
Figure 2: Angular distribution of the signal when rotating the lamps (C860 on the left, C864 on the right) on an axis perpendicular to the optical axis and to the floor. The signals have been normalized with respect to the signal at zero angle.
report_f.doc 27/06/2003 JH
II. Uncertainties-Identification of uncertainty components The calibration scheme performed includes three steps (see Ref. 3): At first, two first order working standards (WS) were calibrated with reference to the gold fixed point blackbody. These two first order WS (C514 and C520) were operated at only one radiance temperature (C514 at 1800 K, C520 at 1337 K).
In a second step, a second order WS (P95) is calibrated with reference to the two first order WS at different radiance temperatures. In the last step, the lamps C860 and C864 are calibrated with reference to the second order WS at nearly the same radiance temperatures.
In the following the contributions to the overall uncertainty are given separately for every calibration step. The uncertainties ui are given at an coverage factor k=1.
a) Calibration of the first order WS with reference to the gold fixed point blackbody
1. Realization of the reference temperature of the gold fixed point. This uncertainty is caused by the impurity of the gold metal inlet (5N, i.e. 0.99999), the emissivity of the cavity (0.99996±0.00001) and the temperature difference ∆T across the bottom of the cavity (<1 mK). The realization of the reference temperature Tr=1337.33 K is within ±0.01 K resulting in a standard uncertainty for the radiance temperature of
2
2 3K01.0
=
rTTu
2. Long term stability of the interference filters used (includes the mean effective wavelength, the spectral transmission of the interference filter, the spectral responsivity of the detector): ±0.05 nm resulting in a standard uncertainty for the radiance temperature of
3nm05.013 λ
−=
rTTTu
3. Uncertainty in radiance comparison including a lamp (spatial and angular distribution of the spectral readiance, cleaning of the window, alignment, ratio of feed back resistors, non-linearity, SSE) ∆L/L=1.5x10-3. This results in a standard uncertainty of radiance temperature of (with c2 being Planck’s second radiation constant)
2
23
4 3105.1
cTu λ−⋅=
4. Uncertainty due to the measurement of the lamp current. With a relative uncertainty u=2.4x10-5 for the voltage measurement and u=1x10-5 for the standard resistor we obtain a resulting standard uncertainty in radiance temperature (with dT/di being the slope of the lamp characteristic T=T(i))
( ) ( )25255 104.210 −− ⋅+⋅
=
didTiu
5. Short term stability of a vacuum tungsten strip lamp of 0.1 K resulting in a standard uncertainty of radiance temperature
3K1.0
9 =u
report_f.doc 27/06/2003 JH
6. Absorption of water vapor (at 950 nm). This may cause a shift in wavelength of 0.065 nm resulting in a standard uncertainty of radiance temperature:
3nm950
nm065.0110
−=
rTTTu
b) Calibration of lamp P95 with reference to Lamps C514 and C520
1. When comparing two sources with different radiance temperatures an uncertainty due to poor blocking of the interference filter caused by parasitic transmission at long wavelengths arises. Using “edge filters” as RG780, RG715 and RG9 a rough estimate of the standard uncertainty in radiance temperature can be made, which is presented in Table 3.
T / K u1 / K T / K u1 / K
900 0,16 1500 0,01
1100 0,03 1700 0,02
1300 0,00 1900 0,02 Table 3: Standard uncertainty in radiance temperature due to blocking error with reference to a blackbody at 1337 K.
2. Realization of the reference temperature u2=u2(first order WS)(T/Tr)2
3. Long term stability of the interference filter u3
4. Uncertainty in radiance comparison including a lamp u4
5. Measurement of the lamp current u5
6. Influence of the temperature of the base. The standard deviation for maximum changes of the base temperature Tb of ±0.1 K is (with dTS/dTB being the change in radiance temperature when changing the base temperature by 1 K)
3K1.0
6B
S
dTdT
u =
7. Resolution of the IR-pyrometer in terms of the photocurrents equals ±2x10-15 A. The photocurrent at 650 nm and 1337 K is 5,4x10-10 A. At 950 nm and 1285 K the photocurrent is 2,6x10-9A. The standard deviation for resolution in radiance temperature is then
for 650 nm 3
nm650
nm650exp
33.1337nm650exp
104.5102 2
2
2
2
10
15
7
⋅
⋅
⋅
⋅⋅
⋅= −
− cT
Tc
Kc
u
report_f.doc 27/06/2003 JH
and for 950 nm 3
nm950
nm950exp
1285nm950exp
106.2102 2
2
2
2
10
15
7
⋅
⋅
⋅
⋅⋅
⋅= −
− cT
Tc
Kc
u
8. Short term stability of the vacuum tungsten strip lamps u9 9. Absorption of water vapour at 950 nm u10
c) Calibration of lamps C860 and C864 with reference to Lamp P95 1. Realization of the reference temperature
u2=u2(second order WS)(T/Tr)2 2. Long term stability of the interference filter u3 3. Uncertainty in radiance comparison including a lamp: in contrast to the first two
calibration steps an uncertainty in radiance of ∆L/L=2.0x10-3 is considered for u4 as the lamps C860 and C864 have not been investigated as thoroughly as the lamps C514, C520 and P95
4. Measurement of the lamp current u5 5. Influence of the temperature of the base u6 6. Resolution of the IR-pyrometer in terms of the photocurrents u7 7. Short term stability of the vacuum tungsten strip lamps u9 8. Absorption of water vapour at 950 nm u10 Collecting all the uncertainties mentioned above the final overall uncertainty at the coverage factor k=1 presented in Table 4 is obtained. Table 4 a) Table 4 b)T (650 nm)/K Uncertainty (k=1, 650 nm) / K T (950 nm)/K Uncertainty (k=1, 950 nm) / K
Table 4: Overall uncertainty as a function of radiance temperature for 650 nm wavelength ( a) and for 950 nm wavelength ( b)
III Calibration results The calibration results of the lamp C860 are shown in Tables 5 a) to d) for the 650 nm wavelength and in Tables 6 a) to d) for the 950 nm measurement. The calibration results of the lamp C864 are shown in Tables 7 a) to d) for the 650 nm wavelength and in Tables 8 a) to d) for the 950 nm wavelength measurement. The tables are slightly modified compared to the tables prescribed in the protocol as we left out some corrections. First no corrections due to non-linearity and water absorption have been
report_f.doc 27/06/2003 JH
made. These two effects have been considered with in the uncertainty budget. Second no correction due to the base temperature were made as our temperature stabilization is sufficient accurate and stable. A possible effect due to a slight variation has been considered in the uncertainty budget. Third the controlling of the strip current is accurate and stable, so no correction to this has been applied (see columns 2-4). However, as we perform a Spline interpolation to the measured radiance temperatures for identifying measurement errors ( see Ref. 3) we applied an additional correction shown in the column named “Spline correction “in Tables 5 to 8.
1
VNIIM. TO THE PROTOCOL ON THE INTERCOMPARISON OF LOCAL REALIZATIONS OF THE ITS-90 BETWEEN THE SILVER POINT AND 1700°C, USING VACUUM TUNGSTEN-STRIP LAMPS AS TRANSFER STANDARDS. REPORTING 1. Experimental and theoretical procedures. 1.1. Realization of the ITS-90. ITS-90 is realized in accordance to its definition by measurement of a ratio of radiances for a standard source as a model of blackbody at the temperatures of the fixed points of a pure metal: Ag, Au and Cu. The models are executed as cylindrical graphite cavity with geometry ensuring an emissivity about 0.9994. A size of the emanating hole is about 1.7 mm. To generate a uniform spatial temperature field a horizontal tubular electric furnace provided by four heaters and PID regulating controllers is used. Construction of the furnace and the crucible form ensure a homogeneity of a temperature field for melting and freezing of metal about 0.07 °C at a time from 20 about 40 minutes. The transfer of the radiance values in fixed points to the secondary measurement standards is carried out with the spectrocomparator - a monochromatic pyrometer that ensures indication of equality of monochromatic radiances of two emitters. The operating principle of the spectrocomparator is based on modulation of two optical radiations and on determination of their spectral luminance equality at a specified wavelength. The device has two symmetric optical channels. Images of radiators under comparison by objectives in the planes of diaphragms, where radiation from periphery radiating surface is preliminary delayed. Selected radiations are directed by means of objectives and prism to mirror of modulator from which they are in turn projected onto the input slit of double monochromator at a modulating frequency of 1070 Hz. Position of images of the sources being compared can be observed with microscope. Substitution of one radiant flux another is carried out by the modulator mirror in such a way that the total illumination of radiation detector remains constant when luminances of the source images are equal, and when they aren't equal - the total flux contains a variable component having the modulation frequency. When the equality of luminances is achieved, the photocurrent variable component is a minimum at the modulation frequency, and it is determined by null-indicator, connected to synchronous detector and selective amplifier. A photomultiplier with a multialcaline cathode or silicon diode were used as a detector of radiation. A vacuum tungsten ribbon lamp with a water-cooled socket is used as a lamp-container. The temperature of the socket is checked by a thermoresistor and its variations had to be less than 0,3 °C. The standard resistors and standard cell used as reference means for absolute measurements of currents and voltages are also thermosetted and checked. The blackbody models in the form of graphite cavities (45 mm length, 6 mm diameter with an aperture 2 mm diameter), being submerged into a crucible filled with metal, were used as standard radiators. The crucibles were made of graphite, the ash content of which was less
2
than 0,03 %. The content of impurities in the metals was less than 0,002 %. To obtain the current values of the lamp at some other temperature a double luminance method was applied. This method is commonly known as the method of combining luminous fluxes. As the secondary measurement standards are used tungsten strip lamps, which simultaneously are interpolation devices realizing a temperature scale in a continuous range of temperatures higher of 962 °C. A scale extrapolates from fixed points by a method of doubling of radiances with the special installation permitting to double visible radiances of a lamp strip. In performing such calibration an additional lamp which is placed instead of the standard radiator, was used. The arrangement with two semitransparent mirror was also used to obtain luminances, multiple of 2±1, by changing the light paths through the mirrors. From these values a corresponding temperature value is calculated using Planck's law. A correction includes to the SSE, the base temperature, the effective wavelength, the stray light. 2. Presentation of results 2.1. Local conditions. 2.1.1. Reference thermometer. - Effective wavelength was λe = 656.3 nm; - Half-width of spectral response function was about 4.5 nm. - Aperture ratio was about f/12. - Target distance was about 330 mm. - Target field dimensions was 0.7 x 0.7 mm. Size-of-source effect covering a range in radii (r0,rmax), with rmax about 12 mm, 2.1.2. Transfer lamps. - Orientation of lamps was normal. - Nominal base temperature was 21.8 ± 0.3 °C. - Total burning time was: 19.5 h for C564; 19 h for C681. 2.1.3. Ambient conditions. - Tamb was in limits: 20.9 ... 21.5 °C - RH was in limits: 55 ... 65 %
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 1
CCT - Key comparison :Comparison of the Local Realizations of theITS-90 between Silver point and 1700 °C
Initial measurements on C564, C681 and C680 (Set I)
NMi/VSL - contributionMarch 1999
R. BosmaE.W.M. van der Ham
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 2
This report describes the initial measurements from VSL for the CCT project ‘Comparison of theLocal Realizations of the ITS-90 between Silver point and 1700 °C’. The project uses two sets oftungsten strip lamps. One of these sets is provided by VSL, the second by NPL.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 3
Description of the scale realisation and lamp setup.
The optical pyrometer.
The pyrometer is a modified version of the pyrometer described in [1-3]. Figure 1 shows a schematicview of the pyrometer.
1
2
3 45 6
7 8 910
Figure 1 Schematic view of optical system of pyrometer1. The optical axis of pyrometer2. Two achromatic lenses with focal length of 500 mm (closely spaced together)3. An aperture stop, with diameter of 1 mm4. A lens with focal length of 75 mm5. The gray filter positions6. The spectral filter position; under 0,3 ° with optical axis7. A lens with focal length of 176 mm8. The field stop, with diameter of 1,76 mm9. The photo diode10. An aperture stop, with diameter of 8,33 mm, which defines aperture ratio to f/9
The objective (2) of this pyrometer consists of two equal achromatic lenses with a focal length of500 mm and a diameter of 80 mm. The lenses have an anti-reflection coating for use between 650nm and 950 nm. The 1 mm aperture (3) is placed in the focus of the lens (2), which results in a 1:1image of the observed source. A second lens (4) creates an parallel beam for the neutral densityfilters (5), the interference filters (6) and the aperture stop (10). To prevent interreflections theinterference filters are placed at an angle of 0,3 ° with the optical axis of the pyrometer. Theaperture stop in the parallel beam defines the aperture ratio to f/9. A third lens (7) produces animage on the field stop. The field stop has a diameter of 1,76 mm which results, after correction forthe the magnification of lenses 4 and 7, in a effective target spot of 0,75 mm. The remainingradiation is collected with a silicon photo diode (9), type Hamamatsu S1337 - 1010BQ. Both the 1mm aperture and the field stop can be observed via a mirror with an objective. The current of thephotodiode is amplified by a Keithley 428 current-to-voltage-converter from which the output ismeasured with a Keithley 181 Nanovoltmeter. The detector is cooled to 19,5 °C with a stability of50 mK.
The pyrometer has three interference filters. The filters have an effective wavelength of 661 nm,672 nm and 959 nm. The full width at half maximum transmission (FWHM) for the filters is 10 nm ,53 nm and 33 nm respectively. The 661 nm filter is used in combination with two additional filtersto improve the out-of-band blocking. All filters are positioned on a computer controlled filterwheel.
In this report the pyrometer has only been used with two filter/amplification combinations, i.e., the661 nm filter with an amplification of 108 VA and the 959 nm filter with 106 VA. Both settings arereferred to by the filter 661nm and 959 nm respectively.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 4
The fixed point furnace
Figure 2 shows the design of the insert of the Ag fixed point furnace. The total insert isincorporated in a quartz tube, with is closed by flanges at both ends. The front flange is equippedwith a quartz window, which is removed during the actual measurements. The fixed point cavity ispositioned in the center of the insert and can be flushed with Argon gas from the rear flange. Withthe window closed the insert can be maintained under Argon pressure. This reduces the decay ofthe four graphite rings in front of the cavity. After a few days of operation these rings arereplaced. The temperature of the cavity is controlled and measured with a S-type thermocouplejust below the fixed point cavity. With this setup an average plateau duration of one hour isobtained.
Figure 2 A schematic view of the insert of the silver fixed point furnace, that is used by NMi/VSL
The Ag fixed point contains approximately 0,8 kg of silver with a purity of 99,9999 %. The graphitecavity has a length of 176 mm, a diameter of 9,5 mm and a bottom with a 120 ° cone. The apertureof the cavity is 3 mm. The calculated effective emissivity is 0,999994 under a iso-thermal cavityconditions. The calculations were conducted with the software program Steep 3 that was design byVega International (1998).
Scale realization
The VSL pyrometer is used as a comparator of radiances. The radiances of the silver point arecompared with the strip lamp operating at an unknown temperature. This temperature isdetermined using an integral method. For this purpose the signal of the pyrometer is representedby:
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 5
Q(T) ' m?0 %
12
? ?
?0 &12
? ?
g(?,T) @ t (?) @L(?,T) @S(?) @d? (1)
were Q(T) the signal of the pyrometer at temperature T80 the effective wavelength of the pyrometer)8 the spectral range of the pyrometerg the emissivity of the sourcet the transmission of the sourceL the radiance according to Planck’s LawS the sensitivity of the pyrometer
The temperature of the lamp is calculated from:
Q(TX) ' q(TX) @Q(TAg) 'i(TX)i(TAg)
@Q(TAg) (2)
were Tx the unknown temperatureTag the silver point temperatureq(TX) the relative response at Tx
i(TX) the photo current of the pyrometer at Tx
The relative response can also by calculated from the quotient of the photo currents. The photocurrents have to be corrected for linearity en size-of-source-effect. Using equation 1 and 2 the truetemperature of the source is calculated after correcting the response is for the emissivity of thetungsten strip lamp. The spectral radiance temperature at a reference wavelength 8r is calculatedusing:
Tr,?r'
c2
?r
@ 1
ln 1g @ t
@ ec2
Tt @?r & 1 % 1(3)
were Tr the spectral radiance temperature at wavelength 8r c2 the second radiation constant, 0,014388 m.KTt the true temperature of the source
A calculation of the uncertainties that is performed directly starting with the integral in Equation 1results in complex calculations. Therefore it is chosen to obtain the uncertainty from themonochromatic point of view as given in the supplementary information for the internationaltemperature scale of 1990 [4]. For each parameter the contribution to the overall uncertainty canthan be calculated with the partial differential quite simply.
The tungsten strip lamp setup
The tungsten strip lamp is mounted on a translation stage which allows movement with submilimeter accuracy in all three planes, in angular rotation and in a tilt angle adjustment. Afterpositioning the lamp, its vertical position was checked using a plumb line. The plumb line was
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 6
positioned between the pyrometer and the lamp. For the adjustment the lamp strip and the plumpline were viewed from the rear of the lamp. The other positions were adjusted looking through anobjective from which the 1 mm aperture (position 3 in figure 1) is visible. For the angularadjustment an additional aperture was mounted in front of the pyrometer objective, increasing thefocal depth considerably. The filament notch and marker on the rear of the lamp were visible. Alladjustment were performed with the lamp at room temperature and using a light tubeilluminating the lamp filament from the rear.
The lamp current is provided by a Heinzinger TNS 15-450 power supply, which is controlled using aRS232/DAC interface. The current is measured using a Holec 300 SEP zero-flux-current-transformer(ZFCT) and the Keithley 181 Nanovoltmeter. The ZFCT transforms a 0 to 20 A dc-current to a 0 to10 V dc-voltage. Adjustments of the current are controlled by computer with a maximum currentrate of 1 A/min.
The base temperature of the lamp is measured using a Pt-100 and an ASL-F300 resistance bridge.The temperature is controlled with a Tamson TC-3 circulating bath. The bath has an externalcooling; an Eurotherm 818 controller with the feedback resistor is mounted in the bath to stabilizethe bath 0,03 °C accuracy. Note that the temperature of the base is not directly controlled by theEurotherm controller; with each current change the setpoint of the controller is changed manually.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 7
Description of the measurements
Selection of lamps
High-stability vacuum tungsten strip lamps, manufactured by GEC, were used for the comparison.These lamps are of type 10/V with a nominal width of the strip of 1,5 mm. For the comparisonthree lamps were selected based on a stabilization test as described in the protocol to thecomparison [5]. During a period of 100 hours the lamps were operated at a current equivalent to atemperature of 1700 °C. During this time both the temperature and current of the lamp weremonitored. The temperature drift during this period of time should not exceed 0,3 K, otherwisethe lamp should be rejected.
Effects of lamp positioning
To check the quality of the lamp and possible interreflections between the lamp and the pyrometerthe influence of position in the three planes and the angular rotation were measured. Thesemeasurements, preformed at approximately 1500 °C, were used in the uncertainty budget for thelamps temperatures.
Spectral sensitivity of pyrometer
The spectral sensitivity of the pyrometer is measured using a single monochromator, Jobin Yvon,type HR1000. The normal stray light suppression of the monochromator is approximately 1·10-4. Forhigh accuracy temperature measurements this is not sufficient. Therefore the stray light is reducedby using an extra band filter at the entrance of the monochromator, resulting in a straysuppression of 1·10-6 around 650 nm. Towards higher wavelengths this suppression decreased to1·10-5. The latter greatly effects measurements on the upper edge of the 959 nm filter andtherefore the accuracy of the spectral response measurement.
Size-of-source effect from pyrometer
The size-of-source effect measurements were performed in 1996 [3]. The setup consists of a 140 mmdisk with a 3 mm black spot in the center to suppress the main signal. The disk was illuminated by a250 W photo lamp. Apertures, ranging from 3 mm to 100 mm, were placed in front of the disk. Theprofile of the illuminated disk was also measured in order to correct for non-uniformity.
In order to determine the size-of-source correction on the pyrometer additional measurementswere conducted concerning the temperature profile of the fixed point furnace used and the sizze-of-source effect when measuring on a high-stability strip lamp.
Linearity of pyrometer
The linearity of the pyrometer was determined using a superposition method [6]. The setup usedfor these measurements consists of two high-stability strip lamps and a beamsplitter with twoshutters. The setup was checked upon interreflections between the lamps; effects were seen smallerthan 1·10-6. Implementing the methode by Coslovi [6] it is not strictly necessary to operate bothlamps at exactly the same temperature.
Transfer of fixed point onto pyrometer
The use of a silicon diode in the pyrometer improved the time stability of the response. Thisaffected our way of realization of the ITS-90. The shortterm stability of the pyrometer is sufficientto transfer the fixed point to the pyrometer instead of transferring it directly onto the strip lamp.Therefore it is not necessary to operate the lamps during the fixed point transfer.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 8
Scale realization on lamp
The realization of the scale was performed at currents specified in the protocol [5]. The lamps werestabilized for 1 hours at 1700 °C the day before the actual scale realization. For actual realizationall currents of appendix D of the protocol were measured successively. For temperatures below1100 °C the stabilization time was 30 minutes and above 1100 °C approximately 15 minutes. Duringthe stabilization time the response of the pyrometer was monitored. Based on these measurementsthe stabilization time was decreased or increased. At each setting the response and detectortemperature of the pyrometer together with the current and base temperature of the lamp weremeasured at least four times. All data were stored in a database for later use.
Measurement of ambient resistance of lamp
After the initial measurements the ambient temperature of the lamp was measured using aHewlett Packard 34420A nanovolt/micro-ohm meter. The meter was connected with four wires tothe base of the lamp; two on the current terminals and two on the connections for the watertubes. A 100 µS standard resistance was used to check the calibration of the instrument. A Pt-100element was used to measure the temperature of the base during these measurements.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 9
Results
Results of lamp selection
Three lamps were tested on stability. Table 1 presents the results.
Lamp identification Drift over 100 hours /K
C564 0,13
C681 0,12
C680 0,15
Table 1 : Stability test of lamps
The lamps C564 and C681 were used for the circulation scheme. Lamp C680 was retained as sparelamp in case of damage on one of the other lamps.
Results of lamp positioning
Figures 3 to 7 present the results of the position measurements on lamp C564. The “normalposition” refers to the position as described in the section “The tungsten strip lamp setup’.
-100
-80
-60
-40
-20
0
20
Res
pons
e /%
-1 -0.5 0 0.5 1 Position /mm
Figure 3 Horizontal position of C564;(0,0) is normal position
-2
-1.5
-1
-0.5
0
0.5
1 R
espo
nse
/%
-6 -4 -2 0 2 4 6 Position /mm
Figure 4 Vertical postion of C564; (0,0) isnormal position
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 10
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.005 R
espo
nse
/%
-3 -2 -1 0 1 2 3 Position /mm
Figure 5 Focus position of C564; (0,0) isnormal position
-0.300
-0.200
-0.100
0.000
0.100
Res
pons
e /%
-10 -5 0 5 10 Position /°
Figure 6 Angular rotation (vertical axis)of C564; (0,0) is normal position
-0.012 -0.01
-0.008 -0.006 -0.004 -0.002
0 0.002 0.004
Res
pons
e /%
-1.5 -1 -0.5 0 0.5 1 1.5 2 Position /°
Figure 7 Angular rotation (horizontalaxis) of C564; (0,0) is normal position
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 11
Figures 8 to 11 present the results of the position measurements of lamp C681.
-100
-80
-60
-40
-20
0
20 R
espo
nse
/%
-0.6 -0.4 -0.2 0 0.2 0.4 Position /mm
Figure 8 Horizontal position of C681;(0,0) is normal position
-6
-4
-2
0
2
4
Res
pons
e /%
-10 -5 0 5 10 Position /mm
Figure 9 Vertical position of C681; (0,0) isnormal position
-0.2
-0.1
0
0.1
0.2
Res
pons
e /%
-15 -10 -5 0 5 10 15 20 25 Position /°
Figure 10 Angular rotation (vertical axis)of C681; (0,0) is normal position
-0.14
-0.12
-0.10
-0.08
-0.06
-0.04
-0.02
0.00
Res
pons
e /%
-3 -2 -1 0 1 2 3 Position /°
Figure 11 Angular rotation (horizontalaxis) of C681; (0,0) is normal position
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 12
Figures 12 to 15 present the results of the position measurements on lamp C680
Figure 12 Horizontal position of C680;(0,0) is normal position
-7 -6 -5 -4 -3 -2 -1 0 1
Res
pons
e /%
-10 -5 0 5 10 Position /mm
Figure 13 Vertical position of C680; (0,0)is normal position
-0.03
-0.02
-0.01
0.00
0.01
Res
pons
e /%
-3 -2 -1 0 1 2 Position /mm
Figure 14 Focus position of C680; (0,0) isnormal position
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
Res
pons
e /%
-15 -10 -5 0 5 10 15 20 Position /°
Figure 15 Angular position of C680; (0,0)is normal position
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 13
Table 2 presents the uncertainties obtained from the position measurements for the three lamps.
Position Adjustmentaccuracy
Uncertainties in percentage of response /%
C564 C681 C680
Horizontal 0,1 mm 0,01 0,036 0,01
Vertical 0,1 mm 0,01 0,12 0,01
Focus 0,5 mm 0,005 - 0,0075
Angular (verticalaxis)
1 ° 0,05 0,054 0,07
Angular(horizontal axis)
1 ° 0,02 0,056 0,07(1)
Combined uncertainty 0,06 0,15 0,10
Table 2: Position uncertainties of lamp; (1) same as vertical axis
Results of spectral sensitivity of pyrometer
Figure 16 and 17 present the spectral measurements of the pyrometer. The blocking with the661 nm filter is measure to be beter than 1·10-6. Due to problems with the stray light of the present monochromator around 950 nm the blocking of the 959 nm filter could not be measured beterthan 5·10-4. The uncertainty in the wavelength for both filters was 0,05 nm resp. 0,2 nm.
0.0
0.2
0.4
0.6
0.8
1.0
Rel
ativ
e se
nsiti
vity
/1
640 650 660 670 680 Wavelength /nm
Figure 16 Sensitivity of pyrometer with661 nm filter
0.0
0.2
0.4
0.6
0.8
1.0
Rel
ativ
e se
nsiti
vity
/1
920 930 940 950 960 970 980 990 Wavelength /nm
Figure 17 Sensitivity of pyrometer with959 nm filter
Results of Size-of-Source effect
Figure 16 presents the result of the Size-of-source effect measurements on the flat plate and thestrip lamp. The flat plate had a 3 mm black spot to suppress the main signal. The black spot withthe strip lamp was 1 mm. Additional measurement were preformed with a flat plate and a 1 mm
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 14
black spot. These measurements are not presented in figure 18.
0.0
0.2
0.4
0.6
0.8
1.0
s(r,r
0) /1
E-3
0 20 40 60 80 100 120 140 Diameter /mm
661 nm filter 959 nm filter661 with strip lamp
Figure 18 Size-of-source effect ofpyrometer with flat plate and strip lamp
For the Size-of-Source effect (SSE) correction the profile of the fixed point furnace is alsomeasured. Equation 4 shows the formula used for the correction.
SSE&correction ' s 0,5(striplamp) & s 0,5(1,5) % s 1,5(AG) (4)
were F0,5(strip lamp) SSE measured on strip lampF0,5(1,5) SSE for 3 mm source measure with 1 mm black spotF1,5(AG) SSE for fixed point source measure with 3 mm black spot
The first two terms are measured with the 661 nm filter. The results are also used for the 959 nmfilter. Table 3 presents the measured and calculated terms. The uncertainty is estimate to 5 %.
Wavelength 661 nm 959 nm
F0,5(strip lamp) 0,16·10-3
F0,5(1,5) 0,21·10-3
F1,5(AG) 0,69·10-3 0,66·10-3
SSE-correction -0,74·10-3 -0,71·10-3
Table 3: Size-of-Source effect correction of pyrometer
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 15
Results of linearity
Table 4 present the linearity measurements for the 661 nm filter.
The average of the uncorrected non-linearity is (-26,9 ± 11,6) ppm. The average of the correctednon-linearity is (-112,7 ± 22,2) ppm. It seems that the correction with the Coslovi-method [6] doesnot work, as the corrected non-linearity increases. The difference between applying the Coslovi-correction and assuming a linear system is 0,08 K at 1700 °C. It is therefore chosen to assume alinear system and to take the non-linearity in account as uncertainty.
The measurements for the 959 nm filter yields the same result as with the 661 nm filter. Table 5presents the linearity corrections for both filters.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 16
i ai
661 nm filter 959 nm filter
012
0,001,000,00
0,001,000,00
Uncertainty 3·10-5 2·10-5
Table 5: Linearity of pyrometer for 661 and 959 nm filter.
Results of transfer of fixed point onto pyrometer
Table 6 presents the results of the fixed point transfer onto the pyrometer.
Filter Fixed pointresponse
[mV]
Uncertainty
[mV]
Date Drift afterprevious transfer
[mK/month]
661 nm 7,89095 0,002 21-05-1997 31
959 nm 34,17147 0,003 27-05-1997 -
661 nm 7,830296 0,002 10-03-1998 58
959 nm 34,13907 0,007 10-03-1998 10
Table 6: Fixed point transfer to pyrometer
The first two fixed point responses were used for the measurements on lamp C564 and C681. Thescale realization was performed within one month. The monthly drift is used as an additionaluncertainty. The data measured on 10-03-1998 was used for the measurements on lamp C680. Themeasurements were completed within two weeks after the fixed point transfer.
Results of scale realization on lamp
In the tables 7 to 19 the following measured or calculated values are presented according to theprotocol:í the measured lamp currentí the measured base temperature of the lampí the ratio of the measured photo current at the lamp and the fixed pointí the ratio of the photo current corrected for Size-of-Source effect and linearityí the calculated true temperatureí the calculated radiance temperature of the lampí the calculated effective wavelength of the pyrometerí the correction due to the deviation of the base temperature from 20 °Cí the spectral radiance temperature given at reference wavelength (650 nm or 950 nm)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 17
Table 7: Run one on lamp C564 with reference wavelength 650 nm. Measured on 29-05-1997 withlaboratory conditions: t = (23,2 ± 0,5) °C and rh = (43 ± 10) %.(3001997.11)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 18
Table 8: Run one on lamp C564 with reference wavelength 950 nm. Measured on 04-06-1997 withlaboratory conditions: t = (23,1 ± 0,5) °C and rh = (47 ± 10) %.(3001997.13)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 19
Table 9: Run two on lamp C564 with reference wavelength 950 nm. Measured on 009-06-1997with laboratory conditions: t = (23,2 ± 0,5) °C and rh = (44 ± 10) %.(3001997.18)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 20
Table 10: Run one on lamp C681 with reference wavelength 650 nm. Measured on 18-06-1997 withlaboratory conditions: t = (23,3 ± 0,5) °C and rh = (42 ± 10) %.(3001997.19)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 21
Table 11: Run two on lamp C681 with reference wavelength 650 nm. Measured on 20-06-1997with laboratory conditions: t = (23,1 ± 0,5) °C and rh = (43 ± 10) %.(3001997.21)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 22
Table 12: Run three on lamp C681 with reference wavelength 650 nm. Measured on 24-06-1997with laboratory conditions: t = (23,1 ± 0,5) °C and rh = (42 ± 10) %.(3001997.23)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 23
Table 13: Run one on lamp C681 with reference wavelength 950 nm. Measured on 19-06-1997 withlaboratory conditions: t = (23,2 ± 0,5) °C and rh = (43 ± 10) %.(3001997.20)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 24
Table 14: Run two on lamp C681 with reference wavelength 950 nm. Measured on 25-06-1997with laboratory conditions: t = (23,1 ± 0,5) °C and rh = (44 ± 10) %.(3001997.24)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 25
Table 15: Run one on lamp C680 with reference wavelength 650 nm. Measured on 11-03-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001998.04)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 26
Table 16: Run two on lamp C680 with reference wavelength 650 nm. Measured on 12-03-1998with laboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001998.05)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 27
Table 17: Run three on lamp C680 with reference wavelength 650 nm. Measured on 13-03-1998with laboratory conditions: t = (23,0 ± 0,5) °C and rh = (45 ± 10) %.(3001998.06)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 28
Table 18: Run one on lamp C680 with reference wavelength 950 nm. Measured on 23-03-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001998.07)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 29
Table 19: Run two on lamp C680 with reference wavelength 950 nm. Measured on 24-03-1998with laboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001998.08)
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 30
Table 20 and 21 present the final results of the calibration on lamp C564. The results werecalculated with a polynomial fit; t = 3ai·Ln(I)i with i = 0..5. For the reference wavelength 650 nmonly one run was preformed. For wavelength 950 nm two runs were measured. Final analysisshowed that the first run was incorrect. The final temperature is therefore only based on thesecond run.
Table 21: Final results C564 with reference wavelength 950 nm. For the average only run 2 istaken.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 31
Table 22 and 23 present the final results of the calibration on lamp C681. The results werecalculated with a polynomial fit; t = 3ai·Ln(I)i with i = 0..5. The final temperature is the average of3 and 2 runs, respectively.
Table 23: Final results C681 with reference wavelength 950 nm.
CCT-Key comparison. Initial measurements on C564, C681 and C680 by NMi/VSL 32
Table 24 and 25 present the final results of the calibration on lamp C680. The results werecalculated with a polynomial fit; t = 3ai·Ln(I)i with i = 0..5. The final temperature is the average of3 and 2 runs, respectively.
[2] Bezemer, J., Metrologia, Vol. 10, 1974, pp. 47 -54.
[3] The characterization of radiation thermometers subject to the size-of-source effectP.Bloembergen, Y.Duan, R.Bosma, Z.YuanProceedings of Tempmeko ‘96
[4] The supplementary information for the international temperature scale of 1990Bureau International des poids et mesures1990, Pavillon de Breteuil, F-92310 SEVRES
[5] Protocol to the comparison of local realizations of the ITS-90 between the silver point and1700 °C using vacuum tungsten-strip lamps as transfer standards.Not published.
[6] Fast determination of the nonlinearity of photodetectorsL.Coslovi and F.RighiniApplied Optics / Vol. 19, No. 18 / 15 September 1980
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 1
CCT - Key comparison :Comparison of the Local Realizations of theITS-90 between Silver point and 1700 °C
2nd measurements on C564 and C681 (Set I)
NMi/VSL - contributionMarch 1999
R. BosmaE.W.M. van der Ham
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 2
This report describes the 2e measurements from NMi/VSL on the VSL-lamp set for the CCT project‘Comparison of the Local Realizations of the ITS-90 between Silver point and 1700 °C’. After theprevious measurements at NMi/VSL they were in chronologically order measured by CSIRO(Australia), KRISS (Korea), NIM (China), NMC (Singapore) and NLRM (Japan). The measurements atthese laboratories were performed in period from July 1997 to June 1998.
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 3
Description of the scale realisation and lamp setup.
The measurements were performed with the same setup as used for the initial measurements [1].For the spectral sensitivity, the size-of-source-effect and the linearity the same calibration valueswere used as with the initial measurements. Only the transfer of the fixed point values wasmeasured again.
Description of the measurements
Stability check on the lamps
According to the Protocol [2] a initial check was made to see if there was mechanical stress presentin the strip. The lamp current and temperature were measured at the silver point current. After thelamp was operated at the 1700 °C- current for one hour, the current was again adjusted to thesilver point. After about 30 minutes the current and temperature were measured. The difference intemperature, corrected for the current difference, should be smaller than 0,25 K.
Effects of lamp positioning
The effects of lamp positioning was not measured again. For the uncertainty budget the results ofthe former measurements were used.
Scale realization on lamp
The scale realization was performed at 650 nm, not at 950 nm.
Measurement of ambient resistance of lamp
The ambient resistance of the lamps was measured before and after all measurements. The samesetup was used as described in the initial measurements [1].
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 4
Results
Results of stability check on lamp
The lamps were tested on stability. Table 1 presents the results.
Lamp identification Drift after stabilisation /K
C564 - first run 0,35
C564 - second run -0,02
C681 - first run 0,01
Table 1 : Stability test of lamps
The drift of lamp C564 was after the first run larger than the 0,25 K as stated in the protocol [2]. Itwas decided to repeat the measurement. In the second run the drift was far below the requestedvalue. The large in the 1st run was probably due to mechanical stress in the lamp.
Results of transfer of fixed point onto pyrometer
Table 2 presents the results of the fixed point transfer onto the pyrometer.
Filter Fixed pointresponse
[mV]
Uncertainty
[mV]
Date Drift afterprevious transfer
[mK]
661 nm 7,800693 0,002 23-06-1998 266
661 nm 7,798007 0,002 01-07-1998 24
Table 2: Fixed point transfer to pyrometer
The first fixed point measurements were related to lamp C564. The scale realization was performedwithin two weeks. The second fixed point measurements were related to lamp C681. Themeasurements were completed within one week after the fixed point transfer. The drift betweenboth fixed point realization was used as an additional uncertainty.
Results of scale realization on lamp
In the tables 3 to 9 the following measured or calculated values are presented according to theprotocol:í the measured lamp currentí the measured base temperature of the lampí the ratio of the measured photo current at the lamp and the fixed pointí the ratio of the photo current corrected for size-of-source effect and linearityí the calculated true temperatureí the calculated radiance temperature of the lampí the calculated effective wavelength of the pyrometerí the correction due to the deviation of the base temperature from 20 °Cí the spectral radiance temperature given at reference wavelength
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 5
Table 3: Run one on lamp C564 with reference wavelength 650 nm. Measured on 24-06-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (47 ± 10) %.(3001998.12)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 6
Table 4: Run two on lamp C564 with reference wavelength 650 nm. Measured on 25-06-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (47 ± 10) %.(3001998.13)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 7
Table 5: Run three on lamp C564 with reference wavelength 650 nm. Measured on 29-06-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (46 ± 10) %.(3001998.14)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 8
Table 6: Run one on lamp C681 with reference wavelength 650 nm. Measured on 02-07-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (45 ± 10) %.(3001998.16)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 9
Table 7: Run two on lamp C681 with reference wavelength 650 nm. Measured on 03-07-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001998.17)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 10
Table 8: Run three on lamp C681 with reference wavelength 650 nm. Measured on 06-07-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (46 ± 10) %.(3001998.18)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 11
Table 9: Run four on lamp C681 with reference wavelength 650 nm. Measured on 07-07-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001998.19)
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 12
Table 10 presents the final results of the calibration on lamp C564. The results were calculated witha polynomial fit; t = 3ai·Ln(I)i with i = 0..5. Because of the large drift between run one and two anadditional run was measured. The final results were calculated with run two and three.
Table 10: Final results C564 with reference wavelength 650 nm.
Table 11 presents the final results of the calibration on lamp C681. The results were calculated witha polynomial fit; t = 3ai·Ln(I)i with i = 0..5. The final temperature is the average over runs two tofour. Run one is rejected because of the large drift (-0,5 K at 1700 °C).
Table 12: Drift of lamps C564 and C681 since initial measurements
-0.10
-0.05
0.00
0.05
0.10
0.15
Drif
t /°C
800 1000 1200 1400 1600 1800 Temperature /°C
C564 C681
Figure 1 Drift of lamps C564 and C681 since initialmeasurements
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 14
Table 13 presents the uncertainty for the scale realization at 650 nm.
Source of uncertainty Type Uncertainty (2F) /°C
tAg tAu 1300 °C 1500 °C 1700 °C
Fixed point
Realization of fixed point B 0,017 0,020 0,027 0,035 0,043
Emissivity of fixed point B 0,001 0,001 0,001 0,001 0,002
Pyrometer
Response A 0,016 0,013 0,017 0,022 0,027
Linearity B 0,002 0,002 0,003 0,004 0,005
SSE B 0,003 0,003 0,005 0,006 0,007
Wavelength B 0,000 0,008 0,033 0,059 0,089
Drift B 0,030 0,035 0,049 0,062 0,077
Lamp
Positioning B 0,105 0,123 0,171 0,217 0,268
Current A 0,109 0,106 0,117 0,135 0,154
Emissivity B 0,006 0,007 0,010 0,012 0,015
Transmission of window B 0,001 0,001 0,002 0,002 0,003
Quality of polynomial fit A 0,052
Total (2F) 0,17 0,18 0,22 0,28 0,34
Total (1F) 0,09 0,09 0,11 0,14 0,17
Table 13: Uncertainty in scale realization at 650 nm
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 15
Results of Measurement of ambient resistance of lamp
Table 14 presents the results of the ambient resistance measurements.
Lamp Resistance[mS]
Temperature[°C]
C564 - before 40,2936 ± 0,00440,3376 ± 0,004
23,36 ± 0,0223,67 ± 0,02
C564 - after 40,3826 ± 0,004 23,88 ± 0,02
C681 - before 34,3526 ± 0,00434,4336 ± 0,004
22,97 ± 0,0223,59 ± 0,02
C681 - after 34,4046 ± 0,00434,4546 ± 0,004
23,65 ± 0,0224,02 ± 0,02
Table 14: Ambient resistance measurements of lamp C564 and C681.
CCT-Key comparison. 2nd measurements on C564 and C681 NMi/VSL 16
References
[1] CCT - Key comparison : Comparison of the Local Realizations of the ITS-90 between Silverpoint and 1700 °CInitial measurements on C564, C681 and C680NMi/VSL - contributionMarch 1999Not published.
[2] Protocol to the comparison of local realizations of the ITS-90 between the silver point and1700 °C using vacuum tungsten-strip lamps as transfer standards.Not published.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 1
CCT - Key comparison :Comparison of the Local Realizations of theITS-90 between Silver point and 1700 °C
Measurements on C860 and C864 (Set II)
NMi/VSL - contributionMarch 1999
R. BosmaE.W.M. van der Ham
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 2
This report describes the measurements from NMi/VSL on the NPL-lamp set for the CCT project‘Comparison of the Local Realizations of the ITS-90 between Silver point and 1700 °C’. Before thelamps were measured at NMi/VSL they were in chronologically order measured by NPL (UK), NIST(USA), CENAM (Mexico), NPL (UK), INM (France) and IMGC (Italy). The measurements at theselaboratories were performed in period from July 1997 to November 1998.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 3
Description of the scale realisation and lamp setup.
The measurements were performed with the same setup as used for the initial measurements [1].For the spectral sensitivity, the size-of-source-effect and the linearity the same calibration valueswere used as with the initial measurements. Only the transfer of the fixed point values wasmeasured again.
Description of the measurements
Stability check on the lamps
According to the Protocol [2] a initial check was made to see if there was mechanical stress presentin the strip. The lamp current and temperature were measured at the silver point current. After thelamp was operated at the 1700 °C- current for one hour, the current was again adjusted to thesilver point. After about 30 minutes the current and temperature were measured. The difference intemperature, corrected for the current difference, should be smaller than 0,25 K.
Effects of lamp positioning
To check the quality of the lamp and possible interreflections between the lamp and the pyrometerthe influence of position in the three planes and the angular rotation were measured. Thesemeasurements were used in the uncertainty budget for the lamps temperatures. The measurementswere preformed at approximately 1500 °C.
Scale realization on lamp
The scale realization was performed at 650 nm and 950 nm.
Measurement of ambient resistance of lamp
The ambient resistance of the lamps was measured before and after all measurements. The samesetup was used as described in the initial measurements [1].
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 4
Results
Results of stability check on lamp
The lamps were tested on stability. Table 1 presents the results.
Lamp identification Drift after stabilisation /K
C860 - first run -0,03
C864 - first run 0,26
Table 1 : Stability test of lamps
The drift of both lamps was within the value stated in the protocol [2].
Results of lamp positioning
Figure 1 to 4 present the results of the position measurements on lamp C860.
-70
-60 -50
-40 -30
-20 -10
0
10
Res
pons
e /%
-1.2 -0.8 -0.4 0.0 0.4 0.8 Position /mm
Figure 1 Horizontal position of C860;(0,0) is normal position
0
1
2
3
4
5
Res
pons
e /%
-5 0 5 10 15 Position /mm
Figure 2 Vertical position of C860; (0,0) isnormal position
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 5
-0.025
-0.020
-0.015
-0.010
-0.005
0.000 R
espo
nse
/%
-3 -2 -1 0 1 2 3 Position /mm
Figure 3 Focus position of C860; (0,0) isnormal position
-0.05
0.00
0.05
0.10
0.15
0.20
Res
pons
e / %
-10 -5 0 5 10 Position /°
Figure 4 Angular rotation (vertical axis)of C860; (0,0) is normal position
Figure 5 to 8 present the results of the position measurements on lamp C860.
-90
-70
-50
-30
-10
10
Res
pons
e /%
-1 -0.5 0 0.5 1 1.5 Position /mm
Rel. response
Figure 5 Horizontal position of C864;(0,0) is normal position
-6
-5
-4
-3
-2
-1
0
1 R
espo
nse
/%
-5 0 5 10 15 Position /mm
Figure 6 Vertical position of C864; (0,0) isnormal position
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 6
-8
-6
-4
-2
0
2 R
espo
nse
/%
-8 -6 -4 -2 0 2 4 Position /mm
Figure 7 Focus position of C864; (0,0) isnormal position
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
Res
pons
e /%
-10 -5 0 5 10 Position /°
Figure 8 Angular rotation (vertical axis)of C864; (0,0) is normal position
Table 2 presents the uncertainties obtained from the position measurements for the lamps.
Position Adjustment accuracy Uncertainties in percentage of response /%
C860 C864
Horizontal 0,1 mm 0,02 0,03
Vertical 0,1 mm 0,02 0,03
Focus 0,5 mm 0,0 0,0
Angular (vertical axis) 1 ° 0,05 0,03
Combined uncertainty 0,06 0,05
Table 2: Position uncertainties of lamp
Results of transfer of fixed point onto pyrometer
Table 3 presents the results of the fixed point transfer onto the pyrometer.
Filter Fixed pointresponse
[mV]
Uncertainty
[mV]
Date Drift afterprevious transfer
[mK/month]
661 nm 7,734898 0,002 03-12-1998 123
959 nm 34,13964 0,007 04-12-1998 -0,2
Table 3: Fixed point transfer to pyrometer
The measurements with the 661 nm setup were completed within two weeks after the fixed pointtransfer. The drift in the fixed point realizations was used as an additional uncertainty.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 7
Results of scale realization on lamp
In the tables 4 to 11 the following measured or calculated values are presented according to theprotocol:í the measured lamp currentí the measured base temperature of the lampí the ratio of the measured photo current at the lamp and the fixed pointí the ratio of the photo current corrected for size-of-source effect and linearityí the calculated true temperatureí the calculated radiance temperature of the lampí the calculated effective wavelength of the pyrometerí the correction due to the deviation of the base temperature from 20 °Cí the spectral radiance temperature given at reference wavelength
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 8
Table 4: Run one on lamp C860 with reference wavelength 650 nm. Measured on 08-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (42 ± 10) %.(3001998.31)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 9
Table 5: Run two on lamp C860 with reference wavelength 650 nm. Measured on 09-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001998.13)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 10
Table 6: Run one on lamp C860 with reference wavelength 950 nm. Measured on 10-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001998.33)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 11
Table 7: Run two on lamp C860 with reference wavelength 950 nm. Measured on 11-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001998.34)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 12
Table 8: Run one on lamp C864 with reference wavelength 650 nm. Measured on 16-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001998.35)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 13
Table 9: Run two on lamp C864 with reference wavelength 650 nm. Measured on 21-12-1998 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (42 ± 10) %.(3001998.36)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 14
Table 10: Run one on lamp C864 with reference wavelength 950 nm. Measured on 05-01-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001999.01)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 15
Table 11: Run two on lamp C864 with reference wavelength 950 nm. Measured on 06-01-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (42 ± 10) %.(3001999.02)
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 16
Table 12 and 13 present the final results of the calibration on lamp C860. The results werecalculated with a polynomial fit; t = 3ai·Ln(I)i with i = 0..5.
Table 13: Final results C860 with reference wavelength 950 nm.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 17
Table 14 and 15 present the final results of the calibration on lamp C864. The results werecalculated with a polynomial fit; t = 3ai·Ln(I)i with i = 0..5.
Table 15: Final results C864 with reference wavelength 950 nm.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 18
Table 16 and 17 present the uncertainty for the scale realization at 650 nm and 950 nm,
Source of uncertainty Type Uncertainty (2F) /°C
tAg tAu 1300 °C 1500 °C 1700 °C
Fixed point
Realization of fixed point B 0,02 0,02 0,03 0,04 0,04
Emissivity of fixed point B 0,001 0,001 0,001 0,001 0,002
Pyrometer
Response A 0,02 0,02 0,02 0,02 0,03
Linearity B 0,003 0,003 0,003 0,005 0,006
SSE B 0,003 0,003 0,005 0,006 0,007
Wavelength B 0,00 0,01 0,03 0,06 0,09
Drift B 0,06 0,07 0,10 0,12 0,15
Lamp
Positioning B 0,05 0,05 0,07 0,09 0,13
Current A 0,13 0,13 0,15 0,18 0,20
Emissivity B 0,006 0,007 0,010 0,012 0,015
Transmission of window B 0,001 0,001 0,002 0,002 0,003
Quality of polynomial fit A 0,06
Total (2F) 0,17 0,17 0,21 0,25 0,31
Total (1F) 0,08 0,08 0,10 0,13 0,15
Table 16: Uncertainty in scale realization at 650 nm
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 19
Source of uncertainty Type Uncertainty (2F) /°C
920 °C 1014 °C 1230 °C 1409 °C 1586 °C
Fixed point
Realization of fixed point B 0,02 0,02 0,03 0,04 0,05
Emissivity of fixed point B 0,001 0,001 0,001 0,002 0,002
Pyrometer
Response A 0,02 0,02 0,03 0,04 0,05
Linearity B 0,002 0,002 0,003 0,004 0,005
SSE B 0,003 0,004 0,005 0,007 0,008
Wavelength B 0,01 0,01 0,07 0,13 0,20
Drift B 0,0 0,0 0,0 0,0 0,0
Lamp
Positioning B 0,06 0,07 0,09 0,11 0,14
Current A 0,13 0,13 0,15 0,17 0,20
Emissivity B 0,004 0,005 0,006 0,008 0,009
Transmission of window B 0,001 0,001 0,001 0,002 0,002
Quality of polynomial fit A 0,02
Total (2F) 0,15 0,15 0,19 0,25 0,32
Total (1F) 0,07 0,08 0,10 0,12 0,16
Table 17: Uncertainty in scale realization at 950 nm
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 20
Results of Measurement of ambient resistance of lamp
Table 18 presents the results of the ambient resistance measurements.
Lamp Resistance[mS]
Temperature[°C]
C860 - before 40,0686 ± 0,004 22,81 ± 0,02
C860 - after 40,1036 ± 0,00440,0836 ± 0,004
23,06 ± 0,0222,96 ± 0,02
C864 - before 41,8546 ± 0,004 22,87 ± 0,02
C864 - after 41,7596 ± 0,00441,7796 ± 0,004
22,59 ± 0,0222,63 ± 0,02
Table 18: Ambient resistance measurements of lamp C680 and C864.
CCT-Key comparison. Measurements on C860 and C864 NMi/VSL 21
References
[1] CCT - Key comparison : Comparison of the Local Realizations of the ITS-90 between Silverpoint and 1700 °CInitial measurements on C564, C681 and C680NMi/VSL - contributionMarch 1999Not published.
[2] Protocol to the comparison of local realizations of the ITS-90 between the silver point and1700 °C using vacuum tungsten-strip lamps as transfer standards.Not published.
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 1
CCT - Key comparison :Comparison of the Local Realizations of theITS-90 between Silver point and 1700 °C
3th measurements on C564 and C681 (Set I)
NMi/VSL - contributionJuly 1999
R. BosmaE.W.M. van der Ham
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 2
This report describes the 3th measurements from NMi/VSL on the VSL-lamp set for the CCT project‘Comparison of the Local Realizations of the ITS-90 between Silver point and 1700 °C’. After theprevious measurements at NMi/VSL they were in chronologically order measured by NPL (GreatBrittain) and VNIIM(Russia). The measurements at these laboratories were performed in periodfrom August 1998 to May 1999.
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 3
Description of the scale realisation and lamp setup.
The measurements were performed with the same setup as used for the initial measurements [1].For the spectral sensitivity, the size-of-source-effect and the linearity the same calibration valueswere used as with the initial measurements. Only the transfer of the fixed point values wasmeasured again.
Description of the measurements
Stability check on the lamps
According to the Protocol [2] a initial check was made to see if there was mechanical stress presentin the strip. The lamp current and temperature were measured at the silver point current. After thelamp was operated at the 1700 °C- current for one hour, the current was again adjusted to thesilver point. After about 30 minutes the current and temperature were measured. The difference intemperature, corrected for the current difference, should be smaller than 0,25 K.
Effects of lamp positioning
The effects of lamp positioning was not measured again. For the uncertainty budget the results ofthe former measurements were used.
Scale realization on lamp
The scale realization was performed at 650 nm, not at 950 nm.
Measurement of ambient resistance of lamp
The ambient resistance of the lamps was measured before and after all measurements. The samesetup was used as described in the initial measurements [1].
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 4
Results
Results of stability check on lamp
The lamps were tested on stability. Table 1 presents the results.
Lamp identification Drift after stabilisation /K
C564 - first run 0,38
C564 - second run 0,09
C681 - first run 0,04
Table 1 : Stability test of lamps
The drift of lamp C564 was after the first run larger than the 0,25 K as stated in the protocol [2]. Itwas decided to repeat the measurement. In the second run the drift was far below the requestedvalue. The large in the 1st run was probably due to mechanical stress in the lamp.
Results of transfer of fixed point onto pyrometer
Table 2 presents the results of the fixed point transfer onto the pyrometer.
Filter Fixed pointresponse
[mV]
Uncertainty
[mV]
Date Drift afterprevious transfer
[mK]
661 nm 7.583074 0,002 06-07-1999
661 nm 7.571304 0,002 20-07-1999 109
Table 2: Fixed point transfer to pyrometer
The first fixed point measurements were related to lamp C564 and C681. The scale realization onboth lamps was performed within two weeks. The second fixed point measurements was measuredafter the realizations on the lamps. The drift between both fixed point realization was used as anadditional uncertainty.
Results of scale realization on lamp
In the tables 3 to 8 the following measured or calculated values are presented according to theprotocol:í the measured lamp currentí the measured base temperature of the lampí the ratio of the measured photo current at the lamp and the fixed pointí the ratio of the photo current corrected for size-of-source effect and linearityí the calculated true temperatureí the calculated radiance temperature of the lampí the calculated effective wavelength of the pyrometerí the correction due to the deviation of the base temperature from 20 °Cí the spectral radiance temperature given at reference wavelength
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 5
Table 3: Run one on lamp C564 with reference wavelength 650 nm. Measured on 08-07-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (45 ± 10) %.(3001999.04)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 6
Table 4: Run two on lamp C564 with reference wavelength 650 nm. Measured on 09-07-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (44 ± 10) %.(3001999.06)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 7
Table 5: Run three on lamp C564 with reference wavelength 650 nm. Measured on 12-06-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (45 ± 10) %.(3001999.07)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 8
Table 6: Run one on lamp C681 with reference wavelength 650 nm. Measured on 14-07-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001999.08)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 9
Table 7: Run two on lamp C681 with reference wavelength 650 nm. Measured on 15-07-1999 withlaboratory conditions: t = (23,0 ± 0,5) °C and rh = (43 ± 10) %.(3001999.09)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 10
Table 8: Run three on lamp C681 with reference wavelength 650 nm. Measured on 16-07-1999 withlaboratory conditions: t = (23,1 ± 0,5) °C and rh = (43 ± 10) %.(3001999.10)
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 11
Table 9 presents the final results of the calibration on lamp C564. The results were calculated witha polynomial fit; t = 3ai·Ln(I)i with i = 0..5. Because of the large drift between run one and two anadditional run was measured. The final results were calculated with run two and three.
Table 9: Final results C564 with reference wavelength 650 nm.
Table 10 presents the final results of the calibration on lamp C681. The results were calculated witha polynomial fit; t = 3ai·Ln(I)i with i = 0..5. Because of the large drift between run one and two anadditional run was measured. The final results were calculated with run two and three.
Table 12: Drift of lamps C681 since initial measurements
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 13
-0.3
-0.2
-0.1
0
0.1
0.2
Dev
iatio
n fro
m 1
st ru
n /°C
800 1000 1200 1400 1600 1800 Temperature /°C
C564 - 2nd run C564 - 3th run C681 - 2nd run C681 - 3th run
Figure 1 Drift of lamps C564 and C681 since initial measurements
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 14
Table 13 presents the uncertainty budget for the scale realization at 650 nm.
Source of uncertainty Type Uncertainty (2F) /°C
tAg tAu 1300 °C 1500 °C 1700 °C
Fixed point
Realization of fixed point B 0,017 0,020 0,027 0,035 0,043
Emissivity of fixed point B 0,001 0,001 0,001 0,001 0,002
Pyrometer
Response A+B 0,016 0,013 0,017 0,022 0,027
Linearity B 0,002 0,002 0,003 0,004 0,005
SSE B 0,003 0,003 0,005 0,006 0,007
Wavelength B 0,000 0,008 0,033 0,059 0,089
Drift B 0,100 0,117 0,163 0,207 0,257
Lamp
Positioning B 0,105 0,123 0,171 0,217 0,268
Current A+B 0,109 0,106 0,117 0,135 0,154
Emissivity B 0,006 0,007 0,010 0,012 0,015
Transmission of window B 0,001 0,001 0,002 0,002 0,003
Quality of polynomial fit A 0,052
Total (2F) 0,19 0,21 0,27 0,34 0,42
Total (1F) 0,10 0,10 0,14 0,17 0,21
Table 13: Uncertainty budget for scale realization at 650 nm
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 15
Results of Measurement of ambient resistance of lamp
Table 14 presents the results of the ambient resistance measurements.
Lamp Resistance[mS]
Temperature[°C]
C564 - before
C564 - after 40,3026 ± 0,00440,3026 ± 0,004
23,59 ± 0,0223,57 ± 0,02
C681 - before
C681 - after 34,4566 ± 0,00434,4486 ± 0,004
23,83 ± 0,0223,76 ± 0,02
Table 14: Ambient resistance measurements of lamp C564 and C681.
CCT-Key comparison, 3th measurements on C564 and C681 NMi/VSL 16
References
[1] CCT - Key comparison : Comparison of the Local Realizations of the ITS-90 between Silverpoint and 1700 °CInitial measurements on C564, C681 and C680NMi/VSL - contributionMarch 1999Not published.
[2] Protocol to the comparison of local realizations of the ITS-90 between the silver point and1700 °C using vacuum tungsten-strip lamps as transfer standards.Not published.