DEUTSCHER KALIBRIERDIENST

Transcription

1 DEUTSCHER KALIBRIERDIENST Abschlussbericht DKD-Ringvergleich konventioneller Wägewert 2003

2 Herausgegeben von der Akkreditierungsstelle des Deutschen Kalibrierdienstes (DKD) bei der Physikalisch- Technischen Bundesanstalt in Zusammenarbeit mit der OE 1.11 der PTB. Copyright 2002 by DKD Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Grenzen des Urheberrechtsgesetzes ist ohne Zustimmung unzulässig und strafbar. Das gilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbeitung in elektronischen Systemen. Deutscher Kalibrierdienst (DKD) Im DKD sind Kalibrierlaboratorien von Industrieunternehmen, Forschungsinstituten, technischen Behörden, Überwachungs- und Prüfinstitutionen zusammengeschlossen. Sie werden von der Akkreditierungsstelle des DKD bei der PTB akkreditiert und überwacht. Sie führen Kalibrierungen von Messgeräten und Maßverkörperungen für die bei der Akkreditierung festgelegten Messgrößen und Messbereiche durch. Die von ihnen ausgestellten DKD-Kalibrierscheine sind ein Nachweis für die Rückführung auf nationale Normale, wie sie von der Normenfamilie DIN EN ISO 9000 und der DIN EN ISO/IEC gefordert wird. Kalibrierungen durch DKD-Laboratorien geben dem Anwender Sicherheit für die Verlässlichkeit von Messergebnissen, erhöhen das Vertrauen der Kunden und die Wettbewerbsfähigkeit auf dem nationalen und internationalen Markt und dienen als messtechnische Grundlage für die Mess- und Prüfmittelüberwachung im Rahmen von Qualitätssicherungsmaßnahmen. Im DKD werden Kalibriermöglichkeiten für elektrische Messgrößen, für Länge, Winkel und weitere geometrische Größen, für Rauheit, Koordinaten- und Formmesstechnik, für Zeit und Frequenz, für Kraft, Drehmoment, Beschleunigung, Druck, Durchfluss, Temperatur, Feuchte, medizinische Messgrößen, akustische Messgrößen, optische Messgrößen, ionisierende Strahlung und weitere Messgrößen angeboten. Veröffentlichungen: siehe Internet Anschrift: Deutscher Kalibrierdienst bei der Physikalisch-Technischen Bundesanstalt Bundesallee 100, D Braunschweig Postfach 33 45, D Braunschweig Telefon Sekretariat: (05 31) Fax: (05 31) dkd@ptb.de Internet:

3 Deutscher Kalibrierdienst Physikalisch-Technische Bundesanstalt FL: Masse Abschlussbericht zum DKD-Ringvergleich DKD Intercomparison Measurement Conventional Mass DKD / PTB FL 1.11 Oktober 2003

4 Contents Page 1 Introduction 5 2 General information Organisation Device to be circulated Transportation Participants and time schedule 5 3 Results Reference s Results of participating laboratories Conclusions 8 A1 Measurement instructions 9 A2 Additional information 11 A3 A5: Results of participants B Corrective actions 19 4

5 1. Introduction 1.1 This DKD Interlaboratory Comparison for Conventional Mass was prepared between July and September 2000 and carried out between October 2000 and December The measurements and evaluations, including the statistical analysis of the calibration results and the determination of the reference s with their uncertainties, have been made in accordance with the regulations of the European co-operation for Accreditation of laboratories that are valid for EA Interlaboratory Comparisons (in particular, EAL-P7 EAL Interlaboratory Comparisons, edit. 1, March 1996 (EA- 2/03)). Accordingly, the parameter E n defined in Appendix H of this document was used to verify the compliance of any calibration laboratory with its best (accredited) of measurement. 1.3 For this Interlaboratory Comparison, a modified E n was used to consider an additional component for the mass drift of the weights. See for further information. 2. General information 2.1 Organisation The DKD Interlaboratory Comparison was organised by the Deutscher Kalibrierdienst, DKD in co-operation with the mass laboratory of the Physikalisch - Technische Bundesanstalt (PTB) as the Reference. 2.2 Device to be circulated The device to be circulated consists of three weights (1 kg, 100 g, 100 mg) described in the Measurement Instructions (see Appendix A1) and in the Additional Information (see Appendix A2) Shapes, materials, magnetic susceptibilities, surface roughnesses as well as densities of the weights fulfill the requirements of accuracy class E2 according to OIML R 111. The 1 kg weight and the 100 g weight have been replaced by new ones after DKD3 and DKD5 calibrations due to pollution and magnetization problems. 2.3 Transportation The weights were contained in special boxes, which were protected from shocks by carton boxes filled with foam chunks. The comparison pack was transported by international transportation agencies or by hand-carrying. The participating laboratories were responsible for transportation to the respective next laboratory according to the circulation schedule. 2.4 Participants and time schedule The organising accreditation body the Deutscher Kalibrierdienst, DKD set up the time schedule as specified in document EAL-P7, taking the special wishes of participating laboratories into account. The circulation schedule was, in most cases, adhered to. But at certain points, necessity arose to make slight changes to it. 5

6 2.4.2 It was decided to include three laboratories from Mexico not accredited by the DKD in this Interlaboratory Comparison The following table gives the sequence of the measurements. Table 1: Sequence of measurements Code Period of measurement October 2000 DKD1 October 2000 DKD2 November 2000 DKD3 November 2000 December 2000 DKD4 January 2001 January 2001 DKD5 March 2001 June 2001 MEX1 October 2001 MEX2 November 2001 MEX3 November 2001 December 2001 Reference laboratory 3. Results 3.1 Reference s For this Intercomparison, the uncertainties of the reference s were significantly better than the of class E2 weights The conventional mass of the artifacts was determined by comparison with the reference laboratory s working standards using weighing schemes with substitution weighings The expanded of the reference s are equivalent to the combined standard multiplied with the coverage factor k = 2. It was determined according to EAL-R2 (EA-4/02). The assigned expanded corresponds to a coverage probability of approximately 95 %. The expanded was combined from the components of of measurement of the used reference standards, of the weighings and of the air buoyancy correction. An estimation of long-term changes was not included The following table gives the measured minimum (first) and maximum (second) conventional mass changes between two recalibrations by the reference laboratory and the resulting drift uncertainties calculated according to equation (3), see below. 6

7 Table 2: Mass changes between two recalibrations Nominal Reference (k = 2) 1 kg 0,15 mg 100 g 0,015 mg 100 mg 0,0015 mg m 0,01 mg 0,03 mg 0,006 mg 0,152 mg 0,0005 mg 0,0013 mg Drift (k = 2) 0,00 mg 0,02 mg 0,003 mg 0,088 mg 0,0003 mg 0,0008 mg It can be seen that there were some large changes in the conventional mass of the weights with time. The maximum mass change exceeded more than ten times (see 100 g) the reference, but it was not considerably larger than the best measurement capability of the corresponding laboratory. In one case (DKD3) soiling of the 1 kg weight was reported directly after receipt at the corresponding laboratory. Nevertheless, the measurements carried out with this soiled weight are within the reference. Though, the 1 kg weight was cleaned after the conventional mass determination at the participating laboratory and a large mass decrease was registered at the reference laboratory (cf. Appendix A3). Therefore, the weight was exchanged. Since there was no reliable reference of the 1 kg weight after the measurements at DKD3 the reference s for the measurements of DKD1, DKD2 and DKD3 were set to the same reference before the measurements were started at DKD1 while the drift was set equal to zero (cf. Appendix A3) The apparent instability has to be taken into account in the calculation of the reference s at the laboratory calibration dates and of the E n ratio as an additional component Therefore, the reference s at the time of measurements carried out by participants were calculated in conformity with the following rules: i) If two consecutive determinations of reference s are in agreement within the limits of reference, their mean is used for all participants. ii) If the two reference s, m 1 and m 2, calibrated at time t 1 and t 2, differ significantly, a time-based linear drift is assumed and for a participant i measuring at time t i the reference was interpolated according to the following equation: m cr,i = m cr,1 + ( m m ) cr,2 cr,1 ti t1 t t 2 1 (1) 7

8 3.2 Results of participating laboratories The results of the measurements are displayed in tabular form and in graphical representation, see Appendix A3 - A5. The E n ratio is shown in tabular form only and is expressed by the following equation: E n = m m 2 cl cl U + U 2 cr cr + U 2 cd where U cl and U cr are the uncertainties associated with the laboratory s conventional mass and the reference, respectively. Additionally, a drift U cd for the mass instability of the artifacts is taken into account: U cd = k 1 12 ( m m ) 2 cr,1 cr, If the drift U cd is greater than the laboratory s U cl, the E n ratio was written in bold and italics The following measurement results gave rise to E n ratio greater than one and should be investigated. The findings and any necessary corrective actions will be included in the final report (cf. Appendix B): Table 3: E n ratios larger than one Nominal Participants 1 kg MEX1, MEX3 100 g DKD1*, MEX1, MEX3 100 mg MEX3 * cf. Appendix B Further Observations Some of the participants calculated and stated smaller measurement uncertainties in the required tables of results than their respective best measurement capability. In the evaluation for this report, the greater of the two data was used in any case After the 100 g weight has been received at the reference laboratory subsequent to the conventional mass determination at DKD5 a change in the magnetization was registered and the weight was exchanged. 3.3 Conclusions Of a total of 24 measurement results, all were used for calculating the respective E n s, six of which are greater than one. This equals a proportion of 25 %, with deviations especially located at the mexican laboratories MEX1 and MEX3. Therefore, another interlaboratory comparison for the measurement quantity conventional weight should be prepared for these participants. Of the laboratories accredited by DKD only one E n was larger than one (cf. Appendix B) Participants measurement uncertainties are spread over a very wide range. With a view to comparable measurement conditions and equivalent handling of comparison devices, the time between two recalibrations should be reduced as much as possible in the future in order to minimize drift uncertainties. (2) (3) 8

9 Appendix A1: Measurement Instructions DKD INTEABORATORY COMPARISON (conventional mass) 1. General Information 1.1 Accreditation Body, responsible for the organisation of this interlaboratory comparison: Physikalisch-Technische Bundesanstalt (PTB) DKD Tel: Postfach Fax: D Braunschweig Germany Contact: Mr. E. Fay, Reference Physikalisch-Technische Bundesanstalt (PTB) 1.11: Mass Tel: Postfach Fax: D Braunschweig Germany Contact: Dr. M. Gläser, 2. Device The device to be circulated comprises the following items, three boxes each marked with and packed in a mailing box of approx. dimensions: 26 cm x 26 cm x 15 cm. Type: 1) 2 cylindrical mass standards with knob (1 kg, 100 g) 2) 1 polygonal wire weight (100 mg) Material: stainless steel Manufacturer: Häfner (1 kg, 100 g), Mettler (100 mg) Nominal s: 1 kg, 100 g, 100 mg 3. Transportation The device has to be carried by car, train or air plane as it appears the safest for the device. The device should be unpacked by an expert of mass calibration immediately after receipt by the calibration laboratory and checked for damage, in particular, a visual inspection of the surfaces should be made and the results be noted on the receipt form. 9

10 4. Handling and storing The mass standards should be manipulated with appropriate pincers. After arrival, they should be removed from the transportation box and be stored under bell jars. The standards shall be stored at least three days in the laboratory, preferably in the balance room, before starting the mass determinations. The surface of the traveling standards should not be cleaned, except for dust removal by using a soft brush. In case of major pollution, the laboratory has to contact the Reference. 5. Measurements The measurements have to be performed according to the normal procedure as agreed with the Accreditation Body. In particular, the conventional mass of each of the three standards has to be determined. In general, the determination of conventional mass does not require the correction for air buoyancy. However, under special conditions, this correction could be necessary; for example, if the density of the reference standard or if the air density differs too much from certain s. For information, see OIML R 111, Appendix B. For this case, the densities and volumes of the traveling standards will be communicated with the final instructions. 6. Circulation Scheme The circulation scheme will be arranged between participants and the Reference. 7. Uncertainty The calibration has to be performed with the best measurement capability (smallest ) agreed with the Accreditation Body. The is to be calculated according to EAL-R2 (EA-04/02). 8. Report Each participant shall send the completed receipt form to the Reference immediately upon receipt of the device and a formal DKD certificate within two weeks after the calibration, also to the Reference. 10

11 Appendix A2: Additional Information DKD INTEABORATORY COMPARISON (conventional mass) Reference Physikalisch-Technische Bundesanstalt 1.11: Mass Postfach Telephone: D Braunschweig Telefax: Germany Contact person: Dr. M. Gläser, Description of the weights ADDITIONAL INFORMATION Nominal Marking Shape Material 1 kg none cylindrical weight non-magnetic stainless steel 1 kg 25 cylindrical weight non-magnetic stainless steel 100 g 3 cylindrical weight non-magnetic stainless steel 100 g point cylindrical weight non-magnetic stainless steel 100 mg triangle polygonal wire non-magnetic stainless steel The weights are contained in a wooden box. The control mark is affixed on the box. The 1 kg weight marked 25 and the 100 g weight marked point are the replacing weights (cf , 3.1.4, 3.2.5). Volume and density s Nominal Marking Volume at 20 C Volume (k = 2) Density at 20 C Density (k = 2) 1 kg None 127,249 cm³ 0,04 cm³ 7858,8 kg/m³ 2,5 kg/m³ 1 kg ,23 cm³ 0,04 cm³ 8049,4 kg/m³ 2,5 kg/m³ 100 g 3 12,7487 cm³ 0,0010 cm³ 7844,0 kg/m³ 0,6 kg/m³ 100 g Point 12,5574 cm³ 0,0010 cm³ 7963,4 kg/m³ 0,6 kg/m³ 100 mg Triangle 0,0126 cm³ 0,0001 cm³ 7960 kg/m³ 48 kg/m³ The volumes and densities of weights from 100 g to 1 kg were determined at the PTB Density. The assumed density of the 100 mg weight was given with 7960 kg/m³. 11

12 The expanded of measurement of volume and of density results from the combined standard by multiplying with the coverage factor k=2. It was determined according to EAL-R2. The of the measurand is normally within the assigned interval of s with a confidence level of approximately 95 %. The standard of volume for the 100 mg weight is based on assumed density of 48 kg/m 3 of stainless steel. 12

13 Appendix A3 Table: Results of 1 kg conventional mass Date of calibration Nominal m n Reference (m cr -m n ) / mg Reference U cr / mg Drift U cd / mg (m cl -m n ) / mg U cl / mg kg 0,08 0,15 0,08 0,15 best meas. capability U bmc / mg Difference m cl -m cr / mg DKD kg 0,08 0,15 0,00 0,40 0,50 0,50 0,32 0,613 DKD kg 0,08 0,15 0,00 0,40 0,50 0,50 0,32 0,613 DKD kg 0,08 0,15 0,00 0,10 0,15 0,50 0,02 0,038 * kg -0,14 0,15-0,14 0,15 ** kg 0,16 0,15 0,16 0,15 DKD kg 0,15 0,15 0,02 0,10 0,50 0,50-0,05-0, kg 0,13 0,15 0,13 0,15 DKD kg 0,14 0,15 0,01-0,12 1,00 1,50-0,26-0, kg 0,15 0,15 0,15 0,15 MEX kg 0,15 0,15 0,01-0,43 0,17 0,50-0,58-1,101 MEX kg 0,15 0,15 0,01-0,21 0,51 1,70-0,36-0,208 MEX kg 0,15 0,15 0,01 0,81 0,50 0,50 0,67 1, kg 0,14 0,15 0,14 0,15 E n ratio E n ratio: Generally, for the calculation of the E n ratio the larger between the calculated measurement and the best measurement capability was used. *: Reference standard has been cleaned **: New reference standard 13

14 Graph: 1 kg conventional mass Deviation from the nominal 1,50 *The reference standard has been cleaned **New reference standard Deviation from the nominal / mg 1,00 0,50 0,00-0,50-1,00-1,50 DKD1 DKD2 DKD3 * ** DKD4 DKD5 MEX1 MEX2 MEX

15 Appendix A4 Table: Results of 100 g conventional mass Date of calibration Nominal m n Reference (m cr -m n ) / mg Reference U cr / mg Drift U cd / mg (m cl -m n ) / mg U cl / mg g -0,060 0,015-0,060 0,015 best meas. capability U bmc / mg Difference m cl -m cr / mg DKD g -0,057 0,015 0,012 0,010 0,050 0,050 0,067 1,262 DKD g -0,054 0,015 0,012-0,050 0,050 0,050 0,004 0,066 DKD g -0,049 0,015 0,012-0,073 0,016 0,050-0,024-0, g -0,040 0,015-0,040 0,015 DKD g -0,088 0,015 0,043-0,070 0,050 0,050 0,018 0, g -0,114 0,015-0,114 0,015 DKD g -0,069 0,015 0,088-0,069 0,051 0,150 0,000 0, g 0,038 0,015 0,038 0,015 ** g -0,016 0,015-0,016 0,015 MEX g -0,013 0,015 0,003 0,070 0,050 0,050 0,083 1,585 MEX g -0,013 0,015 0,003 0,025 0,051 0,170 0,038 0,222 MEX g -0,013 0,015 0,003 0,104 0,050 0,050 0,117 2, g -0,010 0,015-0,010 0,015 E n ratio E n ratio: Generally, for the calculation of the E n ratio the larger between the calculated measurement and the best measurement capability was used. E n ratios (italic and bold): Drift is larger than the measurement of the laboratory. **: New reference standard 15

16 Graph: 100 g conventional mass Deviation from the nominal 0,20 **New reference standard 0,15 Deviation from the nominal / mg 0,10 0,05 0,00-0,05-0,10-0,15 DKD1 DKD2 DKD3 DKD4 DKD ** MEX1 MEX2 MEX3 16

17 Appendix A5 Table: Results of 100 mg conventional mass Date of calibration Nominal m n Reference (m cr -m n ) / mg Reference U cr / mg Drift U cd / mg (m cl -m n ) / mg U cl / mg mg -0,0003 0,0015-0,0003 0,0015 best meas. capability U bmc / mg Difference m cl -m cr / mg DKD mg -0,0006 0,0015 0,0003 0,0000 0,0050 0,0050 0,0006 0,105 DKD mg -0,0006 0,0015 0,0003 0,0000 0,0050 0,0050 0,0006 0,105 DKD mg -0,0006 0,0015 0,0003-0,0007 0,0010 0,0050-0,0002-0, mg -0,0008 0,0015-0,0008 0,0015 DKD mg -0,0013 0,0015 0,0006-0,0010 0,0050 0,0050 0,0003 0, mg -0,0018 0,0015-0,0018 0,0015 DKD mg -0,0021 0,0015 0,0003-0,0020 0,0053 0,0160 0,0000 0, mg -0,0023 0,0015-0,0023 0,0015 MEX mg -0,0030 0,0015 0,0008 0,0003 0,0050 0,0050 0,0033 0,616 MEX mg -0,0030 0,0015 0,0008 0,0055 0,0096 0,0170 0,0085 0,495 MEX mg -0,0030 0,0015 0,0008-0,0360 0,0050 0,0050-0,0331-6, mg -0,0036 0,0015-0,0036 0,0015 E n ratio E n ratio: Generally, for the calculation of the E n ratio the larger between the calculated measurement and the best measurement capability was used. 17

18 Graph: 100 mg conventional mass Deviation from the nominal 0,02 0,01 0,00-0,01-0,02-0,03-0,04 DKD1 DKD2 DKD3 DKD4 DKD5 MEX1 MEX2 MEX3 Deviation from the nominal / mg -0,

19 Appendix B: Corrective Actions In one case (DKD1, 100 g) an E n ratio larger than one was achieved for a laboratory accredited by DKD (cf. Appendix A4). During the intercomparison some corrective actions were made by the corresponding participant, so that at the end of the intercomparison the measurements for the 100 g weight could be repeated successfully (E n ratio smaller than one). The following table gives the appropriate s: Table: Results of 100 g conventional mass Date of calibration Nominal Reference (m cr -m n ) / mg Reference U cr / mg Drift U cd / mg (m cl -m n ) / mg U cl / mg best meas. capability U bmc / mg Difference m n m cl -m cr / mg g -0,010 0,015-0,010 0,015 DKD g -0,003 0,015 0,008 0,020 0,050 0,050 0,023 0, g 0,003 0,015 0,003 0,015 E n ratio E n ratio: Generally, for the calculation of the E n ratio the larger between the calculated measurement and the best measurement capability was used. 19