Radiation thermometry Uncertainties on measurements on Tungsten striplamps

Radiation thermometry Uncertainties on measurements on Tungsten strip-lamps 1 September 7 th 2001, Evd. Ham@NMi. nl

Subject of interest Strip Base 2 September 7 th 2001, Evd. Ham@NMi. nl

Uncertainty budget Phrase subject into context Sources (alphabetic order) • Base temperature • Current • Drift • Emissivity • Positioning • Quality of polynomial fit • Scale realization • Transmission of window 1. Derive its magnitude 2. Find order of importance September 7 th 2001, Evd. Ham@NMi. nl 3

Magnitude - Base temperature has influence on temperature of strip and needs to be corrected for when deviating from Trefbase (Uncertainty 10%) Tstrip (ºC) 700 800 900 1000 1100 Tstrip / Tbase (m. K/ºC) 470 250 110 47 17 • Minimize correction, water-cooled base: Tbase=(20± 0. 1)°C over entire range of operation • Magnitude 5 m. K at 962°C • Type B evaluation 4 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Current Instruments involved • Current supply (stability I/I 5 x 10 -5) • Zero-flux meter (equivalent I/I 1 x 10 -5) • Voltmeter (equivalent I/I 1 x 10 -4, Keithley 181) • Magnitude 0. 44 m. A @ 4 A 54 m. K at 962°C • Type A evaluation 5 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Drift Stability test: operate lamp for 100 hrs @ 1700°C Requirement: T/ t < 3 m. K / hr over 100 hrs • History provides information on correction • Drift behaviour can be used to give an estimate of drift during calibration • Traceability route and context of measurement determines whether this source of uncertainty is relevant • Magnitude negligible • Type B evaluation 6 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Emissivity The emissivity ( ) of Tungsten is involved when converting Trad from pyrometer to 650 nm. • Values ( ) of de Vos are used • Only uncertainty in relative change of ( ) is of importance; estimate 0. 5% • Magnitude 3 m. K at 962°C • Type B evaluation 7 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Positioning-1 • Displacement 2 x horizontal, 1 x vertical • Rotation around vertical/horizontal From profile measurements • Determine sensitivity around zero (%/mm, %/°) • Estimate uncertainty in determination of position Vert. Rot. Hor. • Magnitude 53 m. K at 962°C • Type A evaluation Rot. September 7 th 2001, Evd. Ham@NMi. nl Hor. 8

Magnitude - Positioning-2 Sensitivity Estimate in zero Uncertainty Left-right 0. 1 %/mm 0. 1 mm 0. 01% Up-down 0. 1 %/mm 0. 1 mm 0. 01% 0. 01 %/mm 0. 5 mm 0. 005% Rotate vertical 0. 05 %/° 1° 0. 05% Rotate horizontal 0. 02 %/° 1° 0. 02% Action Back-forward Total 0. 06% 53 m. K 9 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Quality of fit Polynomial fit: t 90= ai ln(S), with a=0. . 5 Least squares residuals result in 26 m. K • Magnitude 26 m. K at 962°C • Type A evaluation 10 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Scale realization Standard pyrometer @ 650 nm • Magnitude 51 m. K at 962°C • Type B evaluation Source Type Uncertainty (m. K) 962°C 1300°C 1700°C Realization B 8 14 22 Emissivity B 1 1 1 Response A+B 8 9 14 Linearity B 1 2 3 SSE B 2 3 4 Wavelength B 0 17 45 Drift B 50 82 128 51 86 138 Total 11 September 7 th 2001, Evd. Ham@NMi. nl

Magnitude - Trans. window The transmission ( ) of the window is involved when converting Trad from pyrometer to 650 nm. • Values ( ) of Pyrex material are used • Only uncertainty in relative change of ( ) is of importance; estimate 0. 1% • Magnitude 1 m. K at 962°C • Type B evaluation 12 September 7 th 2001, Evd. Ham@NMi. nl

Order of importance Type of evaluation Important ? Magnitude Current A YES 54 m. K Positioning A YES 53 m. K Scale realization B YES 51 m. K Quality fit A YES/NO 26 m. K Base B NO 5 m. K Emissivity B NO 3 m. K Transmission window B NO 1 m. K Drift B - - Source 13 September 7 th 2001, Evd. Ham@NMi. nl

EA-4/02 Expression of the uncertainty of Measurement in Calibration Quantity Current Estimate Standard uncertainty 4 A Positioning Scale realization 962°C Sensitivity coefficient Contribution to the standard uncertainty 0. 44 m. A 8. 2 m. A/K 54 m. K 0. 06% 1. 13 %/K 53 m. K 51 m. K 1 51 m. K Quality fit 26 m. K Combined 95 m. K K-factor 2 Expanded 190 m. K 14 September 7 th 2001, Evd. Ham@NMi. nl

Conclusion • Uncertainty in measurements associated with Tungsten strip lamps is at NMi-VSL dominated by 4 contributions (1. Current) 2. Positioning 3. Scale realization 4. Quality of the fit • For NMi-VSL the overall uncertainty is evaluated to values for k=2 and 95% confidence interval: T 90(X) 962°C 1300 °C 1700 °C U(T 90(X)) 190 m. K 280 m. K 420 m. K 15 September 7 th 2001, Evd. Ham@NMi. nl