As LAB C5 Metrology & Calibration

specific criteria for accreditation
Metrology and Calibration Laboratories 5
Published by: International Accreditation New Zealand 626 Great South Road, Greenlane, Auckland 1005 Private Bag 28 908, Remuera, Auckland 1136, New Zealand Telephone 64 9 525 6655 Facsimile 64 9 525 2266 Email: info@ianz.govt.nz Internet: http://www.ianz.govt.nz AS LAB C 5, April 2005. ISBN: 0908611 39 0 Copyright International Accreditation New Zealand 2005
Metrology and Calibration Laboratories
Contents
Page
1 2 3 4
Introduction ............................................................................................................................. 3 Definitions ...............................................................................................................................3 Classes of Test ....................................................................................................................... 3 Laboratory Accommodation and Safety .................................................................................. 3 4.1 Accommodation ............................................................................................................. 3 4.2 Safety............................................................................................................................. 4 Traceability of Measurement ................................................................................................... 4 Equipment Management and Calibration ................................................................................ 4 6.1 Calibration...................................................................................................................... 4 6.2 Branch Laboratories and Site Calibration....................................................................... 5 Computer-Controlled Calibration Equipment........................................................................... 5 Laboratory Staff....................................................................................................................... 5 Calibration Methods ................................................................................................................ 6 9.1 Balance calibration......................................................................................................... 6 9.2 Calibration of thermometers........................................................................................... 6 9.3 Calibration of Pressure Gauges and Transducers ......................................................... 6 9.4 Revenue Meter Calibration ............................................................................................ 6
5 6
7 8 9
10 Uncertainty of Measurement ................................................................................................... 6 11 Certificates of Calibration ........................................................................................................ 7 12 Compliance with Specification................................................................................................. 7 13 Measurement Audits ............................................................................................................... 7 14 References.............................................................................................................................. 7 APPENDIX 1 Classes of Test......................................................................................................... 8 Notes on classes of tests ...................................................................................................... 12 APPENDIX 2 Equipment Calibration Intervals.............................................................................. 13 APPENDIX 3 Example: Scope of Accreditation............................................................................ 18 APPENDIX 4 Approved Signatories and Other Staff.................................................................... 19 APPENDIX 5 Examples: Evaluation of Uncertainty of Measurement ........................................... 21
Introduction
Classes of Test
International Accreditation New Zealand (IANZ)s Specific Criteria Schedules amplify or specialise the IANZ general criteria for particular areas of technology. This schedule must be read together with current issues of the IANZ general criteria for accreditation ISO 17025, General requirements for the competence of testing and calibration laboratories, and Procedures and Conditions of Accreditation, the latter document describing the organisation and operation of the IANZ Laboratory Accreditation Programme. Specific Criteria Schedule No 5 defines technical requirements for accreditation of Metrology and Calibration laboratories in addition to those of ISO 17025. This schedule provides information on classes of test (Appendix 1), staff, accommodation, equipment and other aspects of good laboratory management practice which are considered a minimum standard for Metrology and Calibration laboratories. Detailed criteria are also described for uncertainty of measurement, including some worked examples, as well as recommended calibration periods for equipment. The field covers all types of calibration including: Engineering metrology Electrical Heat and temperature Optical and radiometry Mass, volume and density Pressure and flow Reference material and artefact calibration A list of Specific Criteria Schedules published to date is available from International Accreditation New Zealand on request.
Accreditation by International Accreditation New Zealand does not constitute a blanket approval of all of a laboratorys activities. Therefore, a means of identifying those activities where competence has been demonstrated and for which accreditation is granted is necessary. Accreditation is normally granted only for work which is performed regularly and for which the laboratory is appropriately equipped and has demonstrated its capability. The field of Metrology and Calibration covers those measurements listed in Appendix 1. Note that where reference is made to tests these should be interpreted as measurements for the purposes of accreditation in this field. An example scope of accreditation is shown in Appendix 3.
Laboratory Accommodation and Safety
4.1 Accommodation Accommodation requirements for Metrology and Calibration laboratories vary depending upon the nature of the items to be calibrated and the uncertainty with which measurements are to be made. A formal laboratory area will be required for precise measurements but many calibrations may be satisfactorily performed in production areas or in the field. Formal laboratory areas must have good lighting (400-500 Lux) is recommended for laboratory areas), adequate bench space, freedom from dust and fumes, freedom from vibration and acoustic noise and have appropriate control of temperature and humidity. The extent to which these environmental factors apply will vary according to the uncertainty to which calibrations are performed. When precise measurements are to be made in laboratories the following factors will assume greater importance: (a) Isolation from sources of mechanical vibration and shock likely to have a detrimental effect on sensitive instruments, e.g. lifts, plant rooms, busy roads, etc. (b) Smooth, antistatic finishes for walls, ceilings and floors and, where necessary, air filtration to facilitate dust control. (c) Insulation of walls and shading from direct sunlight. (d) Temperature control of the laboratory at a selected temperature (20C or 23C 2C for example). (e) Humidity control in the region of 35% to 70% RH.
3
2
2.1
Definitions
Best Measurement Capability (BMC) The smallest uncertainty that a laboratory can achieve when calibrating a device that is effectively ideal i.e. where there is no contribution from the device itself to the uncertainty. Usually BMC appears in the laboratorys scope of accreditation associated with a measurement parameter or range of parameters. International Laboratory Accreditation Cooperation (ILAC). Asia Pacific Laboratory Accreditation Cooperation (APLAC).
2.2 2.3.
(f) (g)
(h) (i)
Freedom from fumes likely to have an adverse effect on equipment (e.g. corrosion of switch contacts). Isolation from electromagnetic interference. This is less likely to be necessary for D.C. and low frequency A.C. measurements but assumes importance at R.F. frequencies. Screening may be necessary for some precise electrical calibrations. Stabilisation or filtering of incoming mains power supply where purity of waveform and constancy of voltage is important. Specific housekeeping requirements may apply to certification of reference materials.
(b)
(c)
From an IANZ accredited calibration laboratory which is accredited for the particular measurement or which is accredited by a national accreditation body (such as NATA, UKAS etc) with which IANZ has a mutual recognition agreement. The calibration certificates issued by accredited laboratories must be endorsed in accordance with the requirements of the accreditation bodies concerned. This constitutes proof of traceability to national standards.
Note that ISO 17025 requires that the environmental conditions that can effect the results are identified, documented and the means of controlling them is described. 4.2 Safety Metrology and Calibration laboratories are expected to comply with the Electrical Safety Regulations and any other health and safety requirements which apply. Australian Standard AS 2243 is recommended as a guide to safe practices in laboratories. Part 1 of this Standard provides general guidelines and Parts 6 and 7 provide specific guidelines relating to mechanical and electrical laboratories.
*Critical measurements/calibrations are those which will significantly affect the accuracy or proper performance of tests. IRL NPL NML Industrial Research Limited National Physical Laboratory National Measurement Laboratory of CSIRO.
Equipment Management and Calibration
Management and calibration requirements for equipment are given in clauses 5.5 and 5.6 of ISO 17025. There are some new requirements of ISO 17025 for intermediate checks on equipment and for updating of new calibration constants in software. Calibration labels on your own equipment need to show the expiry date or the date of calibration. These need to be covered in the quality system. Guidelines on re-calibration intervals are given in Appendix 2. 6.1 Calibration Calibration involves controlled comparison of the device under calibration (DUC) against a known instrument over the range of values of use of the DUC. The differences between the known instrument and the DUC are tabulated for a range of pre-selected calibration points. The uncertainty and these differences must be reported for the comparison process. IANZ will not usually accept the reporting of compliance with specification unless the results and uncertainty are also reported. Note that calibration does not involve adjustment that is a separate process which may also be carried out as part of the service offered by the calibration agency. Many calibrations (and electrical in particular) are complicated by the possibility of hardware and/or software adjustment during the comparison.
Traceability of Measurement
Traceability of measurement is ensured when there is an unbroken chain of comparisons of equipment of known uncertainty which link one measurement result to the next and eventually to a national standard of measurement (and therefore to the SI system). Each link in the chain compares equipment with reference equipment of the same or (usually) smaller uncertainty and may involve reference artefacts or materials. The concept of traceability also includes the competence of all the people involved, the fitness of each measurement environment, the suitability of the methods used and all other aspects of the quality and technical systems involved at each step in the chain of measurements. Traceability must be established for all critical* measurement and calibration equipment either: (a) Directly to the national metrology institute (IRL - Measurement Standards Laboratory) or another such national body (eg NPL - UK, NML - Australia) that is a part of the international mutual recognition agreement for NMIs.
Where instruments submitted to a calibration laboratory are likely to be adjusted, appropriate as received measurements must be made and reported. The full calibration can then be carried out after the adjustment. If this procedure is not followed then historical stability data is lost and also the submitting laboratorys ability to take appropriate corrective action on out-of-calibration equipment. Where adjustment of hardware or software takes place during servicing this may invalidate the current calibration certificate, e.g. any of the following will invalidate a balance calibration certificate: (a) Re-programming of scale factors stored in microprocessor memory (b) Any adjustments, such as to linearity or corner-load error (c) Replacement, re-machining or alteration of internal weights or parts associated with the load cell. Manufacturers calibration certificates are acceptable for traceability purposes providing they are endorsed with the logo of a laboratory accreditation body. Calibration certificates endorsed with the logo of an ISO 9000 certifying body alone will not be accepted for critical measurements where traceability is required. 6.2 Branch Laboratories and Site Calibration Some metrological measurements for which formal laboratory accommodation is not essential, e.g. machine tool checks, balance calibration, field pressure or temperature calibrations, field calibration of metering installations etc may be performed in situ and accreditation may be granted for these tests even though the laboratory has no test room set aside for this work. In these circumstances IANZ considers that the staff, reference standards, measuring equipment, storage facilities, transport, and office facilities make a laboratory. Where an accredited calibration laboratory offers site calibration from more than one branch, with all branches having the same scope and procedures, the accredited main branch(s) will be fully assessed as usual. The assessment will cover staff from all sites and their records and equipment details. In addition, each subsidiary branch operation shall, where possible, be assessed at least once between routine reassessments. Every branch must be included in the accredited laboratorys internal audit schedule and the accredited laboratorys quality management system must encompass all branch and site operations. The accredited laboratorys scope of accreditation shall list all branches from which the site calibration services are offered. Site calibration capability must be clearly identified in the scope of accreditation
along with the best measurement capability for each accredited measurement range.
Computer-Controlled Calibration Equipment
Appropriate quality assurance is needed of all inhouse developed software (see ISO 17025, 5.4.7.2). Automatic test equipment must be calibrated in a similar manner to other calibration equipment. The following comments apply to the use of computers for direct data capture and control of the calibration operation. Where control is by proprietary software such as that supplied with some calibrators, validation will only be required of the individual calibration routines for instruments and not for the programme supplied by the manufacturer. For in-house developed software, standard packages of raw data can be developed for feeding through the system to check routines on development or modification of the system. Care should be taken to ensure that such packages cover the expected range of values and include combinations of peculiar circumstances to highlight faults in basic logic of the programme or its subroutines. Alternative systems using spreadsheets or other software may also be used. Reference artefacts may be held to check the operation of the whole system at appropriate intervals. The results of this testing should be recorded and incorporated in the maintenance history. Software maintenance should include a back-up regime and a system recovery plan. Electronic data must be treated in an equivalent way to hard copy to ensure it is not lost or changed without an audit trail.
Laboratory Staff
ISO 17025 gives the general requirements for laboratory staff and management. Laboratory staff seeking signatory approval within Metrology and Calibration laboratories are expected to hold an appropriate tertiary qualification, e.g. a degree, NZ Certificate or other post secondary qualification together with suitable experience. Signatory requirements are specified in Appendix 4 of this document. At least one staff member must be able to determine least uncertainties as required by Section 10 of this Schedule.
Calibration Methods
Where calibration methods are based on manufacturers methods, these must be specialised for the laboratorys own equipment. The procedures must exercise all relevant parts of the hardware and software of the instrument, particularly for calibration purposes. Where standard methods or references are available for calibration purposes (e.g. BS EN 837, AS 1349 for pressure gauges, Traceable Temperatures, ASTM E77 for temperature, AS 2162.1 for volumetric glassware, EN ISO 7500-1 for materials testing machines etc) these should be used. Some detailed technical requirements are given below. These are not exhaustive and cover only a small range of the devices that can be accredited in the metrology and calibration field. 9.1 Balance calibration The reference method for calibration of balances and masses is Prowse: The Calibration of Balances. However, this document is now somewhat dated and needs to be supplemented with the Balances and Weighing workshop course notes from Industrial Research Limited. This course explains how to make specific measurements such as repeatability, off centre loading and linearity (departure from nominal value) amongst other things. It also describes how uncertainty in the balance is determined and gives example layouts for calibration certificates. 9.2 Calibration of thermometers Reference 4 gives much detail about the working of thermometers and how to make traceable measurements including during calibration. It describes calibration procedures and how to determine uncertainty for glass thermometers, resistance thermometers, thermocouples, amongst others. It also gives example certificates. 9.3 Calibration of Pressure Gauges and Transducers Industrial Research Limited course on Traceable Pressure Measurements and the published course notes have valuable information on making pressure measurements, on determining uncertainty using the normal cycling measurements on bourdon tube gauges and also on how to report. Information on the accuracy specifications for different classes of gauge as well as the accuracy of the reference device required are given in standards like BS EN 837-1 and AS 1349. 9.4 Revenue Meter Calibration Calibration of electricity revenue meters is covered in the Metering and Reconciliation Industry Agreement
6
(MARIA) Codes of Practice. COP 2 defines calibration intervals and the necessary accuracy of reference meters that are needed to calibrate specified accuracy class meters. Worked examples of uncertainty for meters, current transformers and voltage transformers may be found on the MSL website at www.irl.cri.nz/teams/msl/ or the MARIA website at: www.m-co.co.nz/c3brules/GuidetoCOPs.doc. Accuracy specifications for electricity meters along with limits for variation in supply voltage, temperature, electromagnetic fields and other influence quantities are given in meter standards such as IEC 61036 and IEC 60687. Calibration procedures for gas meters are given in NZS 5259. Useful guidance on other measurements can be found in the European co-operation for Accreditation (EA) Guidelines documents. A list of these can be obtained from the IANZ librarian. Note that although there are few standard methods in calibration IANZ considers that assessment and appropriate measurement audits can be considered to validate in house developed methods.
10
Uncertainty of Measurement
Metrology and Calibration laboratories must document a policy on calculation of the uncertainty of measurement of any DUC. Laboratories are strongly advised to follow the policy outlined in the guidance document Guide to the Expression of Uncertainty in Measurement (reference 3) and to incorporate its guidance and the following IANZ policy into their own policy statement. The IANZ measurement uncertainty policy is: All known components of uncertainty arising from type A and type B contributions must be considered, evaluated and prepared in an uncertainty budget including: (a) Calibration uncertainty from the certificate of the higher level calibrating laboratory. (b) Any known drift related effects not allowed for in the original calibration certificate or covered by the specified calibration interval. (c) Any influence quantity effects caused by method, environment, DUC, or operator effects and not able to be randomised (i.e. allowed for in a type A assessment). Where information is available from inter-laboratory trials or measurement audits this should be used. (d) Type A contributions obtained by repeating the measurements, preferably 10 times or
more but certainly not less than 5 over a range of normal operation of the equipment. These contributions will be designated and evaluated as standard uncertainty from type A or type B components. The combined standard uncertainty is determined using the root sum square (RSS) method of reference 3. A coverage factor of 2 is then used to give a 95% level of confidence about the result of the measurement (the best estimate of the value of the measurand). An accredited laboratory must also calculate its least uncertainty (or best measurement capability-BMC). The least uncertainty is determined by considering a near ideal DUC (e.g. a device similar to the laboratorys reference device). The contributions (less those from the DUC) are then evaluated, as above for this device, to give the least uncertainty for the laboratory over appropriate ranges of its capability. Appendix 5 provides sample calculations of uncertainty and least uncertainty. References 2 and 3 also provide worked examples that are informative and helpful. There are also some helpful uncertainty calculators available on the Web.
Alternatively, the laboratory may state compliance up to the limit but draw attention to the uncertainty. Under no circumstances can compliance be implied above an upper limit or below a lower limit (see references 7 and 8 for further guidance). Whatever the policy, it needs to be documented in the quality manual.
13
Measurement Audits
Where appropriate and possible, at each assessment or surveillance visit the laboratory may be asked to calibrate at least one artifact or device (e.g. mass, thermometer, burette, DVM, etc) covered by its terms of accreditation. The laboratorys performance will be assessed against known values for the artifact based on measurements by the Measurement Standards Laboratory of New Zealand or by a consensus of other accredited laboratories. The laboratorys own assigned least uncertainties of measurement (BMC) will be considered when assessing its performance in these audits and may be increased if its measurements differ by too much from the reference values. The laboratory will also be expected to participate in any inter-laboratory trials organised by IANZ or by other recognised accreditation agencies or National Metrology Institutes in the Asia-Pacific region where these are available and appropriate.
11
Certificates of Calibration
Clauses 5.10.2 and 5.10.4 of ISO 17025 and Appendix 1 in Procedures and Conditions of Accreditation give IANZ requirements for certificate formats. In addition to these requirements, calibration certificates shall contain: (a) the title Calibration Certificate or Certificate of Calibration. (b) The laboratorys accreditation number. (c) The results of the measurements (d) The uncertainty (or uncertainties) to be associated with the results of the calibration and the applicable level of confidence (95%) along with the actual k value used. (e) The statement that the certificate shall not be reproduced except in full.
14
1. 2. 3. 4. 5. 6. 7. 8.
References
The Expression of Uncertainty and Confidence in Measurement, NAMAS M3003, December 1997. Assessment of Uncertainties of Measurement; R R Cook, NATA, 1999. Guide to the Expression of Uncertainty in Measurement, ISO/IEC/OIML/BIPM (1995) Traceable Temperatures; JV Nicholas and D R White, John Wiley, 1994. International Vocabulary of Basic and General Terms in Metrology, ISO/IEC/OIML/BIPM (1993) Planning of Metrology and Testing Laboratories, OIML. Guidelines on Assessment and Reporting of Compliance with Specification, ILAC-G8. Method of Stating Test Results and Compliance with Specification, APLAC TC004.
12
Compliance with Specification
Laboratories may report compliance with a metrological specification or a manufacturers specification, along with the calibration results, in order to verify a DUC. The preferred way to do this is to report compliance with the specification to the edge of the window of uncertainty (e.g. upper limit minus uncertainty) and not make a statement within the window below the limit.
APPENDIX 1 Classes of Test

5.01 Engineers Limit Gauges (a) Plain plug, ring and gap gauges. Taper plug and ring gauges. (b) Parallel screw plug and ring gauges. Adjustable thread calliper gauges for parallel threads. (c) Taper screw plug and ring gauges. Adjustable thread calliper gauges (3 roll type) for taper threads. (d) Profile gauges (e) Position and receiver gauges involving both linear and angular measurements. (f) Spline and serration gauges. (g) Other gauges involving measurements similar to those under (a) and including depth gauges, height gauges and gauges involving plane coordinated position of holes and spigots. 5.02 Jigs, Fixtures, Cutting Tools and Components 5.03 Engineers Measuring Tools and Instruments (a) Examination to nominated national or international standards (New Zealand, British, Australian, ASTM, ISO etc) of measurement tools and instruments. Test sieves Engineers straightedges Toolmakers straightedges Surface plates Toolmakers flats Engineers parallels (steel) Vernier callipers Vernier height gauges Feeler gauges External micrometers Internal micrometers (including stick micrometers) Engineers comparators for external measurement (magnification up to 2,000) (b) Other tools and instruments. 5.04 Machine Tools (a) Geometric testing including Flatness of beds and tables Straightness of guideways Alignments (parallelism, squareness, etc) Accuracy of lead screws Accuracy of gear drives Accuracy of built-in measuring systems (b) Practical tests including Performance tests Deflection tests
8
5.05 Geometric Form (a) Surface texture (b) Roundness (c) Straightness (d) Flatness (e) Eccentricity (f) Squareness (g) Angle 5.06 Gears, Splines and Serrations (a) Machine cut gears - helical and straight spur (b) Bevel gears (machine cut) (c) Gears for traction (d) Worm gearing (e) Fine pitch gears (f) Gears for turbines and similar drives (g) Straight-sided splines and serrations 5.11 Working Standards of Length and Angle (a) Gauge blocks and accessories (b) Length bars and accessories (c) Cylindrical standards, internal and external (d) Spherical standards (e) Thread measurement accessories (f) Precision linear scales (g) Laser measuring systems (h) Precision graticules including stage micrometers and haemacytometer counting chambers (i) Surveying tapes and petroleum dip tapes (j) Screw pitch reference standards (k) Angle gauges and precision polygons (l) Precision circular scales (m) Reference standards for surface finish 5.12 Precision Measuring Instruments (a) Length measuring machines (b) Screw diameter measuring machines (c) Screw pitch measuring machines (d) Precision projection apparatus (e) Taper measuring machines (f) Drunkenness measuring machines (g) Auto-collimators and other optical instruments used in engineering metrology (h) Dividing heads and tables (i) Gear in hob measuring equipment (j) Extensometers for determination of modulus of elasticity (k) Engineers comparators with magnification greater than 2,000 (l) Sine bars and sine tables (m) Cryptometers
5.15 Reference Materials (a) Reference gases and mixtures (b) Other specified reference materials 5.21 Masses (a) Examination of laboratory standards of mass (b) Examination of industrial standards of mass (c) Determination of the mass of solid objects 5.22 Precision Laboratory Balances Examination of the performance of precision laboratory balances having a nominal measurement uncertainty not exceeding 1 part in 100,000 of maximum capacity. 5.23 Industrial Balances Examination of the performance of industrial balances having a nominal measurement uncertainty exceeding 1 part in 100,000 of maximum capacity. 5.24 Industrial Weighing Appliances 5.25 Pattern Approval Tests 5.28 Flow Measuring Devices (a) Anemometers (b) Mechanical type meters (c) Variable aperture meters (d) Dry meters (e) Wet meters (f) Electromagnetic meters (g) Turbine meters (h) Vortex meters (i) Differential pressure meters (j) Open channel water meters (k) Weir type structures (l) Other devices 5.31 Volumetric Equipment (a) Examination of laboratory volumetric glassware including examination for compliance with the Class A or Class B requirements of the relevant national or international standards (b) Examination of other types of volumetric apparatus (including gas meters) (c) Calibration of tanks, cylinders and other industrial volumetric apparatus including calibration by strapping and calibration by internal measurements 5.32 Density (a) Density of solids (b) Density of liquids (c) Density of gases
5.33 Hydrometers (a) Density hydrometers (b) Specific gravity hydrometers (c) Brix hydrometers (d) Proof spirit hydrometers 5.34 Densitometers (a) Liquid densitometers (b) Gas densitometers 5.35 Hygrometry (a) Humidity measuring devices (b) Environmental chambers 5.41 Barometers (a) Aneroid barometers (including digital barometers) (b) Barographs (c) Mercury barometers, Fortin & Kew types 5.42 Differential Pressure Measuring Devices (including Manometers) (a) Diaphragm types (b) Liquid column types, inclined and vertical (c) Other types 5.43 Pressure Balances Gauge Testers and Pressure
5.44 Pressure and Vacuum Measurement (a) Pressure gauges (b) Vacuum gauges (c) Pressure transducers (d) Pressure recorders 5.51 Force Measuring Devices (a) Proving devices (b) Elastic force measuring devices (c) Force dynamometers 5.52 Strain and Displacement Measurement (a) Extensometers (b) Strain gauges (c) Dial gauges (d) Other devices 5.53 Testing Machines (a) Tension and universal machines in tension (b) Compression and universal machines in compression (c) Vickers hardness machines (d) Rockwell hardness machines (e) Brinell hardness machines (f) Izod impact machines (g) Charpy impact machines (h) Deadweight rubber hardness testers (i) Micro-hardness rubber testers (j) Rubber hardness meters (durometers) (k) Plastic hardness testers
9
(l) (m) (n) (o)
Torsion machines Tension-torque machines Road friction testers Other testing machines
5.67 Refractometry 5.68 Radiant Flux 5.69 Illuminance and Luminance (a) Luminance meters (b) Illuminance meters 5.70 Other Optical and Radiometric Calibration (a) Luminous transmittance or reflectance (b) Spectrophotometry (c) Chromaticity (d) Spectral radiance or irradiance 5.71 Acoustic Equipment (a) Microphones (b) Sound level meters including sound survey meters and impulse sound level meters (c) Sound spectrum analysers and filters (d) Acoustic calibrators, including pistonphones and reference sound sources (e) Sound level recording systems (f) Instrumentation and professional tape recorders (g) Acoustic interferometers (impedance tubes) (h) Standard sound sources (i) Noise dose meters (j) Audiometers (k) Other specified equipment 5.75 Vibration Measuring and Calibrating Equipment (a) Vibration transducers including accelerometers, velocity and displacement pickups (b) Vibration measuring systems (c) Analysers and filters (d) Vibration recording systems (e) Vibration calibrators (f) Other vibration measuring equipment 5.81 Conductors and Resistance Alloys (a) Resistivity (b) Resistance of length of conductor (c) Temperature coefficient (d) Physical properties 5.82 Resistors, Resistance Boxes and Potential Dividers (a) Precision resistors, resistance boxes and conductance boxes (b) Volt ratio boxes and potential dividers (c) DC shunts (d) AC shunts
5.54 Mechanical Testing Equipment (a) Portable Brinell measuring microscopes (b) Indenters for hardness machines (c) Hardness blocks for metals testing (d) Other equipment 5.55 Speed Measuring Devices (a) Tachometers (b) Speedometers (c) Velocity transducers 5.61 Temperature Measuring Equipment (a) Rare metal thermocouples (b) Base metal themocouples (c) Platinum (and other metallic) resistance thermometers (d) Germanium thermometers (e) Thermistors and other semi-conductor thermometers (f) Liquid-in-glass thermometers (g) Clinical thermometers (h) Optical pyrometers (i) Strip lamps (j) Radiation pyrometers (k) Vapour pressure thermometers (l) Filled metal systems (m) Bimetallic systems (n) Digital quartz frequency units (o) Indicators, recorders and controllers (p) Other direct reading temperature measuring systems 5.63 Temperature controlled enclosures (a) Ovens and furnaces (including autoclaves) (b) Baths (c) Incubators (d) Refrigerators and freezers (e) Conditioning rooms and cabinets (f) Other enclosures 5.64 Thermal Radiation 5.65 Irradiance including Solar Radiation (a) Pyrheliometers (b) Pyranometers (c) Ultraviolet pyranometers (d) Pyrradiometers (e) Albedometers (f) Pyr-albedometers 5.66 Polarimetric Instruments (a) Polarimeters (b) Saccharimeters (c) Quartz control plates
10
5.84 Capacitors (a) Precision capacitors (b) Capacitance attenuators (c) Capacitance potential dividers 5.85 Inductors and Transformers (a) Inductors, self and mutual (b) Ratio transformers (c) Power transformers and reactors (d) Current transformers: protection and measurement (e) Voltage transformers: protection and measurement (f) Audio transformers (g) RF transformers (h) Pulse transformers (i) Auto transformers (j) Phase shifting transformers (k) Reference ballasts (l) Neutral earthing transformers 5.86 Voltage Standards and Current Standards (a) Standard cells (b) Electronic emf reference devices 5.87 Transfer Instruments (AC/DC) 5.88 Calibrators for Instrumentation (a) DC voltage (b) AC voltage (c) DC current (d) AC current (e) Resistance 5.89 Indicating Instruments and Recording Instruments (a) DC voltmeters (b) AC voltmeters (c) DC ammeters (d) AC ammeters (e) Wattmeters (f) Varmeters (g) Phase angle indicators (h) Power factor meters (i) Ohmmeters (j) LCR meters (k) Galvanometers and null detectors (l) Energy meters (m) Graphic recording instruments (n) Digital storage recorders (o) Instrumentation tape recorders (p) Electric field strength meters (q) Other specified devices 5.90 Bridges, Potentiometers and Test Sets (a) DC bridges (b) DC potentiometers (c) AC bridges (d) AC potentiometers (e) Ratiometers (f) Current transformer testing sets
(g) (h)
Voltage transformer testing sets Partial discharge test equipment and Time
5.91 Frequency Measurement Measurement (a) Frequency meters (b) Wavemeters (c) Counters (d) Time interval meters (e) Clocks and watches (f) Stroboscopes (g) Frequency standards 5.92 Waveform Measurement (a) Frequency characteristics (b) Input characteristics (c) Timing characteristics (d) Distortion (e) Other characteristics 5.93 Signal Sources (a) Frequency characteristics (b) Output characteristics (c) Modulation characteristics (d) Sweep characteristics (e) Other characteristics
5.95 Communications Equipment (a) Line transmission measuring equipment (b) Radio transmission measuring equipment (c) Field intensity measuring equipment (d) Electrical noise and interference measuring equipment (e) Impedance and reflection measuring equipment (f) Spectrum analysis measuring equipment (g) Data transmission equipment (h) Power measuring equipment (i) Attenuators and amplifiers (j) Waveguide and coaxial components (k) Communication systems (l) Data acquisition systems (m) Processor controlled systems (n) Other equipment 5.96 Electronic Equipment (a) High voltage impulse and disturbance tests (b) Transducer indicators and calibrators (c) Doppler radar equipment (d) Miscellaneous equipment and tests 5.97 High Voltage Testing (a) Direct voltage (b) Alternating voltage (c) Impulse voltage (d) Impulse current (e) Partial discharge (f) Dielectric tests (g) Other specified tests
11
Notes on classes of tests (i) Accreditation for adjustable thread calliper gauges (class 5.01) is granted only to laboratories which can perform complete examinations. It is not granted for setting only. Facilities are required for checking geometry of the anvils including thread form, eccentricity and taper of roller anvils and relative position of fixed anvils. It is usually necessary for laboratories to have their own facilities for the checking of setting plugs. In some circumstances regular checking of setting plugs by an IANZ accredited laboratory or by the Measurement Standards Laboratory may be accepted. (ii) Class 5.02 covers measurement of jigs, fixtures, cutting tools and components which are similar to the gauge measurement in class 5.01. It includes also the examination of laboratory moulds and cutting dies such as cement cube moulds, vicat moulds, rubber specimen cutters such as tension test dies, compression set test cutters and tear test dies. Engineers comparators with magnifications greater than 2000 are included in class 5.12. A wide range of laboratory instruments come within the scope of class 5.03 (b). Examples include: Vicat apparatus (cement testing) Pensky-Martens apparatus (petroleum testing) Penetration needles (bitumen and grease testing) Thickness gauges for rubber, plastics, textiles and other materials Airflow nozzles and petrol jets. Precision measuring instruments are included in class 5.12. Acceptance test charts for metal working and wood working machines have been published: jointly by the Institutions of Mechanical Engineers and Production Engineers by the Machinery Publishing Co Limited Testing Machine Tools, by Dr G Schelsinger and Dr F Koenigsberger ISO 230 - Machine Tool Test Code together with specific ISO recommendations.
(vi)
Examination of gear cutting tools such as hobs and rack type cutters are included in class 5.02. Examination of the gauges specified in BS 2059 is included in class 5.01 (g). Special purpose laboratory items are included in class 5.31 (b), examples include: Dean and Stark apparatus Diluent trap for testing crankcase oil Degassing chamber for gas content of insulating oils. Similar tests are included within the field of Electrical Testing, e.g. 5.81, 5.82, 5.83, 5.97.
(vii)
(viii)
(iii) (iv)
(v)
12
APPENDIX 2 Equipment Calibration Intervals

The following table sets out the normal periods between successive calibrations for a number of reference standards and measuring instruments. It must be stressed that each period is generally considered to be the maximum appropriate in each case providing that the other criteria specified below are met: (a) the equipment is of good quality and of proven stability, and (b) the laboratory has both the equipment capability and staff expertise to perform adequate internal checks, and (c) if any suspicion or indication of overloading or mishandling arises the equipment will be checked immediately and thereafter at frequent intervals until it can be shown that stability has not been impaired. Where the above criteria cannot be met appropriately shorter intervals may be necessary. IANZ is however prepared to consider submissions for extension of calibration intervals based on factors such as history of stability, frequency of use, accuracy required, ability of staff to perform regular checks and successful participation in measurement audits. It is the responsibility of the laboratory to provide evidence that its calibration system would ensure that confidence in the equipment could be maintained. Items marked * in the table are those which can be calibrated by the staff of a laboratory if it is suitably equipped and the staff is competent to perform such re-calibrations. Inter-comparisons are usually performed by laboratory staff. Where calibrations have been performed by the staff of a laboratory adequate records of these measurements must be maintained. The first column shows the maximum recommended period between the initial calibration and the first recalibration. The second shows the maximum period between subsequent re-calibrations provided that the two earlier calibrations indicate that the item is stable. These re-calibration intervals apply only to equipment of good quality and stability that is used, handled and stored with care. Excessive usage of equipment would lead to a reduction in these periods.
Equipment
Maximum period (years) between successive calibrations Initial Calibration Subsequent Calibration
Accelerometers Acoustic calibrators Anemometers Angle gauges - reference - working Attenuators
one one one three two three (Frequency response). Resistance and return loss check annually where appropriate. Inter-compare every six months
One One One five four three
Balances Bandpass filter sets Barometers - aneroid Bridges Calibration baths and furnaces - (Fortin and Kew types)*
one two five three three three (See NATA Technical Note no. 8) (Full calibration). Range check annually. Complete temperature survey initially.
three two five* three three five
13
Metrology and Calibration Laboratories Equipment Maximum period (years) between successive calibrations Initial Calibration Subsequent Calibration
Capacitors Dead weight testers Density bottles Digital multimeters* Digital calibrators with self checking Dividing heads Environmental chambers Filters (optical for calibrating spectrophotometers for and absorbance) Frequency analysers Frequency Counters* Frequency standards Force testing machines
three five two one two three three five five one
Inter-compare annually.
three five five two two five
Time and spatial (temperature variations, recovery time, rate of ventilation).
five ten five two
See Time below. Two to five years depending on type. (Where required by a standard method this period will be less) three two two three Six months (compare thermometers at room temperature with wick dry). Five years (complete calibration). two one Three months (with Assman Psychrometer) three three five three (Full calibration). Annual intercomparison of transformers to detect major problems. Inter-compare annually Inter-compare every six months or more frequently as required three three five five five four four five
Gauge blocks - reference - working Height setting micrometers and riser blocks Hydrometers Hygrometers* (i) assman and sling type psychrometers (ii) recorders accurate to 1% (iii) other recorders including hair types Inductors Instruments, electrical* (analogue) Instrument and ratio transformers Instrument transformer test sets
Length bars - reference - working Levels - precision

14
three two three
five four five
Linear scales - precision Manometers - reference - working* Masses (integral, stainless steel or nickel chrome alloys) - reference - working* Masses (all other types) - working* Micrometers, dial gauges, callipers etc
three five three
five five three
three three one See IANZ Technical Guide No one
five five three
Microphones
one
three monthly check of frequency response and sensitivity. Calibrate annually or when one dB change is detected whichever is sooner.
one
Optical flats Optical parallels Orifice plates and nozzles Pistonphones Potentiometers Precision polygons Pressure and vacuum gauges -reference -working* RF noise sources RF power measuring equipment
three three Initial dimensional calibration. six monthly visual inspection. one five three one three months two one Power references. three years for thermistor sensors. Annual check of VSWR. Inter-compare every six months
five five
five five one six months two three
Reference glass filters Resistors Rollers and balls Roundness standards Roughness standards Screw check plugs for ring gauges Screw pitch reference standards
five one three three three three three After initial drift rate has been established. Inter-compare annually.
ten three five five five five five

15
Setting cylinders Setting rings Signal generators Sound level meters Sound power sources Spectrophometers Squares - trysquares - block squares Standard cells, electronic
three three one two five six monthly. Wavelength and absorbance calibration. See IANZ Technical Guide No. four. two three one Intercompare at least three monthly to establish drift rate of a group. One cell in a group needs to be calibrated annually. Then intercompare within group as required. three three Check every three months
five five two two five six monthly.
five five one
Surface plates - cast iron - granite Thermocouples (i) rare metal (ii) base metal
five five
ten0 hours use or three years whichever is the sooner. Calibration intervals to suit the particular application. five (full calibration). Check ice point immediately after initial calibration then at least every six months. five (full calibration). Check ice point immediately after initial calibration then at least every six months. Intercompare with reference thermometer(s) at points in the working range every six months. (see IANZ Technical Guide No three). one (full calibration) five five
Thermometers (i) reference liquid-in-glass (ii) working liquid-in-glass*
or alternatively
(iii) electronic (sensors that are thermocouples, thermistors or other integrated circuit devices)* (iv) resistance
one five
five (full calibration) or when the ice point drift is more than five times the uncertainty of calibration. Check at ice point before use or at least every six months. Working hand-held resistance thermometers can be checked using the alternative procedure for glass thermometers Thread measurement cylinders
16
three
five
Thread measurement vee pieces Time*, time interval, and frequency standards*
two one But calibration interval dependent on equipment frequency type and accuracy required. This may be as frequently as daily if the highest possible performance is required (via TV line six). Audit the data collection system every two years. five Maximum with annual self-check for a standalone passive instrument. Two years active devices. three five one Inter-compare every three months. Annual resistance checks.
five
Transfer standards, ac-dc
five
Volt ratio boxes Volumetric glassware* Watthour meters (electromechanical) Wattmeters and watthour meters (electronic)
three five two two
two With regular intercomparisons - intervals to be based on history of performance.
17
APPENDIX 3 Example: Scope of Accreditation

CONFORMANCE STANDARD: NZS/ISO/IEC 17025: General requirements for the competence of testing and calibration laboratories. Nominal calibration temperature 20 2C (in laboratory) All measurement uncertainties are based on a confidence interval of 95% Calibrations are performed at the premises of the accredited laboratory or in the field. Classes of Test: 5.85 INDUCTORS AND TRANSFORMERS (d) Current transformers Calibration of CTs of classes 0.2 and 0.5 to the accuracy requirements of IEC 60044 1 in accordance with in house methods and MARIA COP 2. Least Uncertainty Ratio Error Phase Error 0.033% 1.2 minutes (e) Voltage transformers
Calibration of VTs of classes 0.2 and 0.5 to the accuracy requirements of IEC 60044 2 in accordance with in house methods and MARIA COP 2. Least Uncertainty Ratio Error Phase Error 0.023% 1.2 minutes 0.022% 1.2 minutes
33kV 11kV
5.89 INDICATING INSTRUMENTS (l) Energy meters Calibration of multiple phase energy meters to the accuracy requirements of classes 0.2, 0.5, 1.0 and 2.0 as defined in IEC 61036, 60687 in accordance with in house methods and MARIA COP 2. Active Meters Power factor 1.0 0.5 lag 0.8 lag 0.866 lag & lead Power factor 1.0 0.5 lead & lag Least Uncertainty 0.043% 0.054% 0.048% 0.047% Least Uncertainty 0.066% 0.072
Reactive Meters
18
APPENDIX 4 Approved Signatories and Other Staff

Supervisory staff in accredited organisations must be competent and experienced in the technical areas covered by their accreditation. They must be able to oversee the operations and cope with any problems that might arise in their work or that of their colleagues or subordinates. Such staff members, nominated by their organisations, may be granted signatory approval by the Testing Laboratory Registration Council on the recommendation of IANZ. Approved Signatories may authorise technical procedures and the release of IANZ endorsed work. The qualifications and experience required of Approved Signatories and other technical staff members, cannot be rigidly specified but must be appropriate to the work in which they are engaged. Approved Signatories would normally hold tertiary qualifications or equivalent professional recognition in the relevant discipline. Organisations engaged in a restricted range of repetitive work may have that work controlled by a Signatory with appropriate practical experience and specific training in that work but without formal qualifications. Approved Signatories Approved Signatories are the knowledgeable staff members who, where relevant: (a) develop and implement new operational procedures; (b) design quality control procedures, set action criteria and take corrective actions; (c) identify and resolve problems; (d) authorise the release of all IANZ endorsed reports; (e) take responsibility for the validity of the results. Every accredited organisation must have at least one Approved Signatory covering each item of its scope of accreditation. Accreditation is automatically suspended for any scope item(s) where there is no Signatory for the item(s) due to Signatory/ies leaving the organisation. All IANZ endorsed calibration certificates or reports must be authorised by an Approved Signatory holding approval in that discipline, who will take full responsibility for the validity of the work. Authorisation can be by signing or by preprinted signature or electronic signature with appropriate software safeguards covering release of the report information. Signatory approval is a recognition of personal competence. However, it relates to the accreditation of the employing organisation and is therefore not automatically transferable to another organisation. It lapses when a Signatory leaves the accredited organisation or changes their role significantly within the accredited organisation. The following are considered when IANZ assesses the suitability of staff members as Approved Signatories: (a) Relevant qualifications and/or experience. If the signatories do not have relevant tertiary qualifications they must have sufficient relevant experience enabling them to comply with the requirements listed below. (b) Position in the staff structure. Approved Signatories must be technical personnel closely involved in the day to day operations of the accredited organisation. (c) Familiarity with procedures and awareness of any limitations of these procedures. Approved Signatories must have appropriate personal experience in the work procedures for which they hold approval. They must be aware of any limitations of these procedures, and must understand the scientific basis of the procedures. (d) Ability to evaluate results critically and a position in the staff structure which makes them responsible for the adequacy of results. (e) Knowledge of the quality assurance procedures in operation and ability to take appropriate and effective corrective action, when required. (f) Knowledge of and a commitment to the IANZ requirements for Signatories and for accreditation. This will include being conversant with the principles of effective quality management embodied in ISO 17025 and relevant Specific Criteria. (g) Sufficient experience with the accredited organisation to address the above points. It is difficult to specify an exact time a proposed Signatory should have spent in the organisation, as it is dependent on their previous knowledge and experience and their current role in the accredited organisation. It is unlikely that the time would be less than six months, but exceptional circumstances may apply. Signatory approval is normally granted only to a staff member in charge, a section leader, a departmental manager or a senior staff member who authorises the release of and reports/certificates who can also satisfy the above requirements.
19
Staff members may be granted signatory approval for all of the work included in their organisations scope of accreditation or for only specific work or classes of work relating to their area of personal expertise. Signatory approval is available to a person engaged by an accredited organisation as a consultant, with respect to work done within the scope of accreditation of that organisation, provided that there is a written agreement between the parties setting out the extent of the authority and responsibility of the consultant in relation to the services provided. The consultants position in the organisation must be such that they can perform their role as a technical decision maker, as effectively as if they were an employee. Staff members of the accredited organisation who are not engaged full time are also eligible for signatory
approval, provided that the circumstances in which they are called upon to exercise their signatory function and their access to, and knowledge of, the technical operations are such that they are able to take full responsibility for the results they authorise. The position and function of an Approved Signatory are quite distinct from that of an Authorised Representative. An organisation will normally have only one Authorised Representative who is appointed by the organisation and is only the contact point for IANZ and need not have any particular professional or technical expertise. The organisation may, however, have several Signatories approved by IANZ and with their own individual areas of expertise. An Authorised Representative who is not also an Approved Signatory may not authorise the release of IANZ endorsed reports.
20
APPENDIX 5 Examples: Evaluation of Uncertainty of Measurement

General The uncertainty of a measurement defines its quality and is an aid to calculating risk in commercial decisions. It enables the determination of compliance with manufacturing tolerances, minimum production weights, and, therefore, the likelihood of distributing product out of specification. In the calibration environment an uncertainty analysis also demonstrates an understanding of important effects. The most helpful uncertainty statement is realistic, not conservative or optimistic. Uncertainty calculations improve with experience. In addition, there is usually more than one way of evaluating a contribution. Therefore, record your calculations and assumptions so that you and other staff can follow what youve done, and perhaps later add improvements. When you are calculating measurement uncertainty, keep in mind: (a) Where known, errors must be corrected and not incorporated into the uncertainty. (b) The total uncertainty will usually be dominated by just one or two large contributions. Once they have been identified, make the best use of your time by focusing your efforts on them. (c) Obtaining a model is an essential step in an uncertainty analysis. While this may appear difficult, in most cases the model is simple with the various terms simply added together (e.g. corrections to thermometer or pressure gauge readings), so the sensitivity coefficients are +1 or -1. (d) However well known a calibration process is, there may still be unrecognised contributions that become evident in a measurement audit, inter-laboratory trial, or in repeat calibrations. Wherever possible, the source of these should be identified, the contribution evaluated, and the uncertainty analysis updated. (e) The meanings of technical terms such a k value, effective degrees of freedom, standard uncertainty, correlation, coverage factor are as defined in references 1, 2 and 3. (f) To prevent roundoff errors, carry at least two or three significant digits in values of the various terms while you are doing the calculation. The final uncertainty should be rounded to one or at most two significant digits. (g) For many calibrations done in IANZ accredited calibration laboratories, type B determinations will dominate the uncertainty calculation. These usually have high degrees of freedom associated with their standard uncertainties so we can use k=2 in most cases.
Example 1 Calibration of a glass thermometer against a reference platinum resistance thermometer in a stirred water bath.
1 Identify the sources of uncertainty in the thermometer calibration. The largest sources of uncertainty in liquid-in-glass thermometers are irregularities in the scale markings and bore diameter, small random effects as the mercury sticks in the bore, and the slow expansion and contraction of the glass with temperature giving rise to hysteresis (the reading depending on previous exposure to different temperatures). The mercury stiction is minimised by tapping the thermometer lightly (e.g. with a pencil) before reading. The hysteresis is minimised by only measuring rising or falling temperatures. The simplest calibration assesses the errors over a narrow range of temperatures; say six readings distributed over a few scale divisions above and below the temperature of interest. We apply a single correction, calculated from the average error, to account for most of the errors, and use the standard
deviation to account for the remaining unpredictable variations. We will assume the reference thermometer is an electronic platinum resistance thermometer and the thermometer under calibration is a mercury-in-glass thermometer divided to 0.05C. Type A Uncertainties 1. The standard deviation calculated from the six comparison measurements with the reference thermometer samples most of the important effects: the irregularities in the bore and scale markings, mercury stiction, and fluctuations in the temperature of the calibration bath. Because the six comparison temperatures are chosen randomly and do not fall exactly on scale markings of the thermometer, the standard deviation also includes effects due to the finite resolution of the thermometer and the readability effects arising from the operator having to interpolate between scale markings. Type B Uncertainties 2. The reference platinum resistance thermometer (PRT) is calibrated by the Measurement Standards Laboratory of New
21
Zealand at Industrial Research. The calibration certificate gives the uncertainty, the level of confidence, and coverage factor. We assume the electronic reference thermometer is direct reading in temperature. 3. Spatial non-uniformity of calibration bath temperature. This must be assessed before hand (say once every few years) to ensure it does not have a significant effect on the calibration, and to define the controlled volume of the bath. As described here there is no need to add a term for the resolution of the instrument. However, if the six comparison readings are chosen to fall near the scale markings then a contribution due to the resolution and the uncertainty in reading the scale of the thermometer should be included. Uncertainty due to hysteresis of the thermometer. The hysteresis is minimised by calibrating (and using) the thermometer only to measure rising temperatures (for T > 20C), or falling temperatures for (for T < 20C).
confidence of 95%, and the coverage factor is 2.13 corresponding to 15 degrees of freedom. The standard uncertainty is found by dividing the expanded uncertainty by the coverage factor. If the effective degrees of freedom are quoted with 95% confidence, we can find the k factor from the Students t tables. If neither the coverage factor nor the degrees of freedom are given, we assume a coverage factor of k = 2 and divide by 2. In this case
Uref = 0.008 2.13 = 0.0038 C
4.
3.3
5.
Variation with position in the calibration bath Measurements of time and position variation in the bath were made previously at several radial positions about a central position, at several immersion depths, and at several temperatures (e.g. the minimum, maximum and a mid temperature). A number of measurements (e.g. 10 or so) are taken at each point. These measurements should be repeated every few years (no more than 5), or if the bath is modified. The time variation of the bath is assessed as the standard deviation of the 10 or so measurements made at each point. In a well-stirred bath, the standard deviations will be similar at different positions at a given temperature. The worst-case variation normally occurs when the difference between the bath and the room temperature is greatest. The standard deviation is not used directly in the uncertainty analysis because the time variation will already contribute to the standard deviation in 3.1. However, this term must be included in the calculation of the best measurement capability as it limits the possible uncertainty in a calibration. The spatial variations in the bath temperature can be assessed using the average of the measurements at each point. They will affect a calibration but not contribute to the standard deviation in 3.1. Therefore, we must add a term to account for the possible temperature gradients in the bath. Providing the thermometers involved in the calibration are always kept within the assessed diameter and immersion depths we can use the maximum variation we have measured as the contribution to the uncertainty of calibration by considering it to be a rectangular distribution. In this case we find that so long as the thermometers are in the centre of the bath and within 100 mm of each other, the maximum measured observed spatial difference is 0.01C. The standard uncertainty for this contribution is determined by treating this difference as the range of a rectangular distribution (where, if 2a=0.01, the standard uncertainty is a/3):
2 Corrections applied to readings An ice point check of the reference platinum resistance thermometer is regularly conducted, and a correction determined for any shift in the ice point reading so the uncertainty due to drift is negligible. Corrections to the reference thermometer, reported in its calibration certificate, are applied. as
In principle, large spatial errors in the calibration bath could be corrected, but it is better to ensure that they are negligible by ensuring that the bath has low nonuniformity and the sensors of the two thermometers are placed close together. 3 3.1 Determination of standard uncertainties Repeatability of the thermometer under calibration. The standard deviation of the six comparison measurements includes most of the important effects. The standard deviation was found to be Uduc = 0.009 C Since six calculated, associated distribution be normal. readings are taken and a mean is the number of degrees of freedom with the standard deviation is five. The of the thermometer errors is assumed to
3.2 Uncertainty in the reference thermometer The reference thermometer certificate tells us that it has an uncertainty of 0.008C, given at a level of
22
Ubath =
0.005 3
= 0.0029 C.
3.4
Resolution of the thermometer under calibration The readability or resolution of liquid-in-glass thermometers varies according to the quality of the scale markings on a thermometer. The scale is analog scale, so the qualities of the markings affect how the operator interpolates between the markings. The standard uncertainty due to reading errors ranges from about 0.1 of a scale division for a high quality thermometer to 0.25 of a scale division for a very ordinary thermometer. If the comparison points in the calibration are chosen randomly then the effects are included in the standard deviation calculated in 3.1 and do not need to be included separately. 3.5 Hysteresis of the thermometer under calibration The slow expansion and contraction of the glass introduces a small hysteresis effect (typically no more than 0.1% of the temperature change) into the readings of the liquid-in-glass thermometer. The change can be detected by doing an ice point check at the beginning and the end of the calibration process. The difference is taken as the range of a rectangular distribution, and hence the standard uncertainty is estimated, (2a=0.02 C), as:
Uhys = 0.01 3 = 0.0058C
glass thermometer calibrations, the value is between 2.1 and 2.5. Because the uncertainties are based on a few measurements dominated by random effects, the uncertainty itself is only known to about 30%, so using a value of 2 and rounding up will be OK. In this case, we can expect U = 0.024 C to be close. An accurate value of the coverage factor can be calculated by calculating the effective number of degrees of freedom in the total:
eff =
i =1
4 Uc N Ui4 i
Using the values from the uncertainty table (below) we get

eff =
( 0.0117)4 ( 0.009) / 5 + ( 0.0038) / 15 + ( 0.0029)4 / + ( 0.0058)4 /
4 4
(0.0117)4 (0.009) / 5 + (0.0038)4 / 15

4
= 14.1
Note that the terms with infinite degrees of freedom drop out of the calculation (in the spreadsheet calculation we have used 10 to the power 10 to approximate infinite degrees of freedom). This gives an effective number of degrees of freedom of 14 and so we use k = 2.14. The expanded uncertainty is therefore:
U = 2.14 X 0.0117 = 0.025C.

4 Combination of Uncertainties The uncertainties are combined using the root sum square method (in quadrature). The combined uncertainty is thus:
2 2 2 2 2 Uc = Uduc + Uref + Ubath + Uhys
2 Uc = (0.009)2 + (0.0038)2 + (0.0029)2 + (0.0058)2 oC
which gives a slightly different answer from before. Often, as in this case, such differences will be small when compared with the uncertainty in the estimate itself. However that is not always the case. 6 A model and the uncertainty budget The process described above is a way of assessing uncertainty based on experience and commonsense. The ISO Guide to the expression of uncertainty in measurement (reference 3) has a much more structured way to do this and prepare the uncertainty budget based on a model of the measurement. The model is the mathematical relationship between the inputs to the calibration and the output(s). The factors we have included as inputs, bath variation, repeatability, reference uncertainty and hysteresis are all directly related to the output (the correction at a temperature). For many calibrations, the mathematical model is a simple sum of readings and corrections. This greatly simplifies the propagation of uncertainty equations as the sensitivity coefficients
23
Uc = 0.0117o C
5 Expanded uncertainty To calculate the expanded uncertainty, U, we need the coverage factor k for a level of confidence of 95%.
U = kuc
If all the contributions have a large number of degrees of freedom, this value is 2. More typically in
are all equal to +1 or 1. Our model for the correction to the thermometer under calibration is:
= duc - ref - bath - hys
IANZ expects the uncertainty budgets in calibration laboratories to be prepared as shown in the example below. A number of spreadsheets including those developed by IRL are already in this format. 7 Best Measurement Capability The best measurement capability is calculated as the total uncertainty combining all sources attributable to the reference instruments and the calibration procedure. This includes the spatial and temporal variations in the bath temperature, and the uncertainties in the reference thermometer. If we use the standard uncertainties we have calculated for the reference thermometer and the spatial variation in the bath and combine these with the time variation (which we will assume is the same size as the spatial for this example) we get:
2 Ubmc = (0.0038)2 + (0.0038)2 + (0.0029)2
where
= The difference between the readings of the reference thermometer and the thermometer under calibration = The reading of the thermometer under calibration = The corrected reading of reference thermometer = Temperature error due to bath non uniformity = Temperature error due to hysteresis in thermometer under calibration
duc ref
bath hys
The correction is based on the average observed error: Correction = Average (T) = Average (Tduc - Tref) Note that the errors due to hysteresis and bath nonuniformity are assumed to average to zero. The equation for the uncertainty has a form corresponding to the model except that all of the terms are summed as squares:
2 2 2 2 2 Uc = Uduc + Uref + Ubath + Uhys
Ubmc = 0.0061 C
which combine to give a best measurement capability of 0.012C (95% level of confidence with k=2). The spreadsheet for this example is shown below.
Note also that we can convert models that involve division, multiplication and powers to direct relationship between input and output (or linear) form by using relative uncertainties (or % values).
The easiest way to develop the structured uncertainty budget is to use a spreadsheet. We need the following: (a) A description of the source of the contribution. (b) The type of uncertainty estimate, Type A or Type B. (c) The estimated size of the contribution (the standard uncertainty). (d) The type of distribution assumed e.g. normal, rectangular, triangular etc. (e) The number of degrees of freedom for Type A, and where necessary the effective degrees of freedom for Type B, for each contribution. (f) The sensitivity coefficient. (g) Where necessary, the effective degrees of freedom in the output so we can calculate the coverage factor k. (h) The combined uncertainty. (i) The expanded uncertainty.
24
Symbol U duc U spatial U ref U hyst
Source Repeatabili Spatial var. Reference Hysteresis
Type A B B B
Uncertainty in Correction of Glass Thermometer ci Units Value Dist Divisor C C C C 0.009 Norm 0.005 Rect 0.008 Norm 0.01 Rect 1 1.73 2.13 1.73 1 1 1 1
ui
vi
0.009 5.00E+00 0.003 1.00E+10 0.004 1.50E+01 0.006 1.00E+10
Combined Uncertainty: Expanded Uncertainty: k =?
0.012 2.140 1.41E+01 0.025 C
Example 2. Calibration of a bourdon tube pressure gauge against a dead weight tester.
1 Sources of uncertainty in the gauge calibration The uncertainty we are looking for is the total uncertainty in the correction to the gauge at a point. We will assume the reference device is a dead weight tester (dwt) and the gauge being calibrated is a test gauge (accuracy class 0.25) covering the range 0 to 40 MPa with a minor scale interval of 200 kPa. 3
laboratories correct for local gravity and generally maintain their calibration areas within 2C. Temperature effects on the gauge under calibration. This contribution is considered to be negligible compared to the resolution of the gauge. Resolution limitations in reading the scale of the gauge under calibration. Normally we take the larger of the repeatability or resolution uncertainty. We will include both in this example as any dual counting of contributions will be small. Uncertainty due to hysteresis of the gauge under calibration. This contribution is significant but is incorporated in the repeatability estimation. No extra contribution is necessary. Uncertainty due to a difference in the height of the gauge from the dwt reference level.
Type A Uncertainties 1 Repeatability is usually determined by making more than one run up and down the scale of a gauge. For multiple runs the preferred method of calculating the uncertainty is to use Appendix C of the MSL Pressure Calibration Course Notes. An alternative approach is to make repeat measurements (at least 5 repeats) at several key pressures and calculate the standard deviation at each pressure. Lets assume we have a standard deviation from 5 measurements of 45 kPa. Type B Uncertainties 2 The reference device (a dwt) is calibrated by Industrial Research at the Measurement Standards Laboratory of New Zealand. These calibrations have the uncertainty reported on the certificate as a percentage of the pressure generated and also as a lower limit in pressure units. To keep this example simple we will use the calibration certificate uncertainty and assume it covers the generated pressure variations due to temperature and gravity. This is a reasonable assumption as NZ accredited
2 2.1
Determination of standard uncertainties Repeatability of the gauge under calibration. The standard uncertainty arising from random variation in the correction to 5 readings at a particular pressure is calculated from the standard deviation of those readings as noted above. This value is the worst case variation from measurements at 5 pressures. We shall assume 24 degrees of freedom and a normal distribution (if we had only one set of 5 repeats the number of degrees of freedom would be 4).
Uduc = 0.045 MPa Uduc = 45 kPa

25
2.2 Uncertainty in the dead weight tester The dwt has an expanded uncertainty of 0.025% on the calibration certificate with a lower limit of 2.5 kPa. The k factor is given as 2 so we will assume 30 degrees of freedom. Then:
Uref = 0.025 x 40,000 2 x 100 Uref = 5 kPa kPa
3 Combination of Uncertainties The uncertainties are combined using the root sum square method. The combined uncertainty is thus:
2 2 2 2 2 Uc = Uduc + Ures + Uref + Uhead
2 Uc = ( 45 )2 + ( 14.4 )2 + ( 5 )2 + ( 0.04 )2
Uc = 47.5kPa
2.3 Uncertainty due to gauge resolution The scale division size is 200 kPa. The resolution is assumed to be 1/4 of a division or 50 kPa. The standard uncertainty is found by treating it as a rectangular distribution.
Ures = 50 2 3 kPa
4 Expanded uncertainty To calculate the expanded uncertainty, U, we need the coverage factor k for a level of confidence of approximately 95%. If we have 30 effective degrees of freedom, k = 2.09. The expanded uncertainty is thus:
U = 2.09 x 47.5 kPa = 99.3 kPa.

Remember this is the uncertainty in the correction to the gauge being calibrated at a point. The uncertainty will need to be evaluated at other points as well. The MSL method is much more efficient at determining the uncertainty for that reason and should be used by accredited laboratories calibrating gauges for others, particularly if compliance with a gauge class tolerance is required.
Ures = 14.4 kPa
We will assume a large number of degrees of freedom. 2.4 Uncertainty due to height difference between gauge and dwt. The pressure generated by a head difference is: P = gh Every mm of head generates 0.00686 kPa of pressure error for a liquid of density 700 kg/m3. This is considered to be rectangular with a = 10 mm and standard uncertainty = a/3. So:
Uhead = 10 x 0.00686 3 kPa
Where the pressure was required in the calculations, 40 MPa has been used.
5 A model and the uncertainty budget
C = Pdwt + Phead + Rgauge

Where C = the gauge correction at P Rgauge = indicated pressure of the gauge under calibration Pdwt = pressure generated by the dwt Phead = pressure due to head difference between the dwt and the gauge
Uhead =
0.04 kPa
Again we assume a large number of degrees of freedom for this rectangular distribution. As this contribution is tiny we could neglect it. It is included for completeness.
We could have calculated the uncertainty in the generated pressure using the equation relating pressure to force generated by the masses on the piston and cylinder assembly which would have provided a more complicated model. The model we have chosen is much simpler to use.
The spreadsheet for this example is shown below.
26
Symbol U duc U res U ref U head
Source Repeatabili Resolution Reference Head differ
Uncertainty in Correction of Bourdon Tube Pressure Gauge ci Type Units Value Dist Divisor A B B B kPa kPa kPa mm 45 Norm 25 Rect 10 Norm 10 Rect 1 1.73 2 1.73 1 1 1 0.00686
ui 45.0 14.4 5.0 5.8
vi 24.0E+0 10.0E+9 30.0E+0 10.0E+9
47.5 2.1 29.8E+0 99.3 kPa
6 Best Measurement Capability The best measurement capability is calculated as the total uncertainty combining all sources attributable to the reference instruments and the calibration procedure. This includes only uref and uhead which combine to give a best measurement capability of 10 kPa at 40 MPa (95% level of confidence).
make measurements at the top of the lowest range i.e. 50g.
Example 3. Calibration of a 200g analytical balance.

1 Sources of uncertainty in the balance calibration The uncertainty we are looking for is the total uncertainty in the balance indication at a particular weight. Traditionally the CSIRO document The Calibration of Balances by D B Prowse has been used in New Zealand for balance calibration. This calculates properties called the Limit of Performance or Best Accuracy of the balance which includes uncorrected errors and type A uncertainties in the same final figure. As balances are rarely corrected for linearity errors, there is some justification for continuing this policy while acknowledging that the uncertainty is a different number. We will follow the current MSL policy which is covered in their training courses. We will assume stainless steel integral masses of OIML class E2 are used as the reference weights. These are calibrated at MSL and the certificate includes the actual mass values and the uncertainty at a level of confidence of 95% and k factor of 2. The balance has a maximum indication of 201g and reads to 0.1mg on the lowest range.
Type B Uncertainties 2 Uncertainty in the masses. These are calibrated by Industrial Research at the Measurement Standards Laboratory of New Zealand. The actual value of each mass is reported along with the uncertainty.
3 Resolution limitations in reading the scale or with a digital indication. One could argue that this is included in the repeatability as for 1 above. We will include it in this example. Uncertainty arising from drift in the masses between calibrations. Temperature effects on the scale sensitivity of the balance. Off centre loading of the balance is assessed but is not included as a component of uncertainty as it is assumed that the calibrator will load it so this affect is minimised. Hysteresis is also assessed but is not included as a component of uncertainty as it is included in other contributions.
4 5 6
Type A Uncertainties 1 Repeatability is usually determined by making 10 measurements at representative points across the range. For this example we will
2 Determination of standard uncertainties 2.1 Repeatability of the balance. The standard uncertainty arising from random variation in 10 readings at a particular weight is calculated from the standard deviation of those readings. We shall assume a normal distribution with 9 degrees of freedom.
Ur = 0.00012 g
27
2.2 Uncertainty in the masses The uncertainty in the 50g reference mass is given as 0.000015g. The k factor is 2 so we will assume 30 degrees of freedom (see reference 1). Then:
Uref = 0.0000075g
Usens =
2 x 10-6 x 2 x 50 2x 3
Usens = 0.000058 g
2.3 Resolution of the balance The balance reads to 0.1mg on the 50g range. Treating this as a rectangular distribution with infinite degrees of freedom:
Ures = 0.0001 x 2 2 3
2 2 2 2 2 2 Uc = Ur + Uref + Ures + Udrift + Usens
2 Uc =
(.12)2 + (.0075)2 + (.0408)2 + (.00433)2 + (.0577)2 Uc = 0.14mg = 0.00014 g
Ures =
0.0000408g
where the 2 in the numerator accounts for the resolution being a difference between the load reading and zero. 2.4 Drift in the masses between calibrations. Where reference masses are used frequently in the field there is a risk that their certified values may change by an amount approaching the uncertainty. We will assume that occurs in the worst case and treat the uncertainty as a rectangular distribution:
Udrift = 0.000015 2 3 g = 0.00000433g
4 Expanded uncertainty To calculate the expanded uncertainty, U , we need the coverage factor k for a level of confidence of approximately 95%. Because we have a large number of degrees of freedom we can assume k = 2. The expanded uncertainty is thus:
U = 2 x 0.14 mg = 0.28 mg.

This is the uncertainty in the indication of the balance at 50 g. The uncertainty will need to be evaluated at other points in the other ranges of the balance as well. The limit of performance or best accuracy is found by adding the worst linearity error to this value for the 50 g range. The worst error, L, is found to be 0.0001 g, making the best accuracy, 0.00038 g or 0.38 mg. 5 A model and the uncertainty budget
2.5 Drift in sensitivity due to temperature We need either the actual change in scale sensitivity between the beginning and end of a calibration or we assume the coefficient relating scale sensitivity drift to temperature (See MSL Balances and Weighing Workshop Notes). The two possibilities are:
Usens = sens drift during cal x range max value 2 3
EM = IM IO + Mref + Esens where

EM IM IO Mref Esens
= error in balance indication at load M = the balance indication at load M = the balance indication at no load (usually zero) = the mass of the standard weights used = the error due to sensitivity drift (usually assumed to be zero with an appropriate uncertainty assigned).
or a minimum of:
Usens = 0.2 x 10-6 x (max - min temp) x (range max)
We assume a sensitivity variation with temperature of 2x10-6 g/C. We will use the latter estimate as a worst case value and thus assume a large number of degrees of freedom and a rectangular distribution. We get, for a 2C temperature change during the calibration:
The spreadsheet for this example is shown below.
28
Symbol Ur U res U ref U drift U sens
Source Repeatabili Resolution Mass unce Mass drift Sens variat
Uncertainty in Indication of an Analytical Balance ci Type Units Value Dist Divisor A B B B B mg mg mg mg C 0.12 Norm 0.070711 Rect 0.015 Norm 0.0075 Rect 2 Rect 1 1.73 2 1.73 1.73 1 1 1 1 0.05
ui
vi
0.120 9.00E+00 0.041 1.00E+10 0.008 3.00E+01 0.004 1.00E+10 1.155 1.00E+10
0.140 2.000 1.65E+01 0.279 mg
6 Best Measurement Capability The best measurement capability is calculated as the total uncertainty combining all sources attributable to the reference instruments and the calibration procedure. This includes only uref and udrift which combine to give a best measurement capability of 0.02 mg at 50 g (95% L.C).
temperature controlled laboratory and the GB and micrometer have come to equilibrium. 4 5 6 Resolution limitations in the digital readout of the micrometer will be considered. Drift in the gauge blocks will be considered negligible for this calibration. Uncertainty due to lack of parallelism and flatness of the micrometer anvils will be considered.
Example 4 Calibration of an external micrometer against gauge blocks.

1 Sources of uncertainty in the calibration The uncertainty we want is the uncertainty in the correction to the micrometer at a point. We will assume the reference is a set of grade 2 gauge blocks and the micrometer is a 0-25 mm digital indicating type.
Type A Uncertainties 1 Repeatability is determined by making 30 measurements at 25 mm. This will give us a standard uncertainty with 29 degrees of freedom. Type B Uncertainties 2 The reference gauge block (GB) set is calibrated by an accredited laboratory. These calibrations have the uncertainty reported on the certificate at a level of confidence of 95% with k = 2. We will assume 30 degrees of freedom for this value.
3 Temperature will affect both the gauge block and the micrometer. Using the formula in Example H.1 in reference 3 (GUM) we assume that the micrometer and block have the same thermal expansion coefficient and are at the same temperature. We shall also assume the measurements are being carried out in a
2 Determination of standard uncertainties 2.1 Repeatability of the micrometer. The standard uncertainty arising from random variation in the correction to 30 readings at 25 mm is calculated from the standard deviation of those readings which we will assume is 0.17 m. There are 29 degrees of freedom. We normally only make 5 measurements during a calibration so we divide this figure by 4 where 4 = number of readings minus 1 to obtain the experimental standard deviation of the mean (used to be called standard error).
Uduc = 0.17 4 m
i.e.
Uduc = 0.085 m
2.2 Uncertainty in the Gauge Block length The 25 mm GB has an uncertainty of 0.21 m on the calibration certificate. The k factor is given as 2 so we will assume 30 degrees of freedom. Then:
29
Uref =
0.21 2
m = 0.105 m
2 2 2 2 2 2 Uc = Uduc + Ures + Uref + Utemp + Ualign
2.3 Resolution of the micrometer The micrometer resolution is 1 m. The standard uncertainty is found by treating it as a rectangular distribution of size 2a = 1 m.
Ures = 0 .5 3 m = 0.289 m
2 Uc = (0.085)2 + (0.289)2 + (0.105)2 + (0.033)2 + (0.139)2
Uc =
0.349 m
We will assume a large number of degrees of freedom. 2.4 Uncertainty due to temperature effects on the GB and the micrometer.
Utemp = L x x U(t ) 3 m
4 Expanded uncertainty To calculate the expanded uncertainty, U, we need the coverage factor k for a level of confidence of approximately 95%. Because we have a large number of degrees of freedom we can assume k = 2. The expanded uncertainty is thus:
U = 2 x 0.35 = 0.70 m .
Remember this is the uncertainty in the correction to the micrometer being calibrated at 25 mm. The uncertainty will need to be evaluated at other points as well. 5 A model and the uncertainty budget
Cmicr = Lmicrometer LGB + Ctemp + Calign
where L = the GB length = the thermal expansion of the GB U(t) = the uncertainty in t between the GB and micrometer We shall assume there is 0.2C between them and treat this as a rectangular distribution. Other contributions from the GB not being at 20C are considered negligible. Then: utemp = 0.033 m 2.5 Uncertainty due to parallelism and flatness of the micrometer anvils. We make a number of measurements on the flatness and parallelism of the micrometer anvils and take the worst case figure as the full width of a rectangular distribution:
Ualign = 0.24 3 = 0.139 m
where
Cmicr Lmicr LGB Ctemp Calign
= = = = =
the micrometer correction indicated length of the GB certified length of the GB Correction for temperature effects Correction for alignment effects
The contributions of parallelism and flatness are each taken to be 0.17m and these are squared and combined to give 0.24 m. The result is taken to be a rectangular distribution.
6 Best Measurement Capability The best measurement capability is calculated as the total uncertainty combining all sources attributable to the reference instruments and the calibration procedure. This includes only uref, utemp and ualign which combine to give a best measurement capability of 0.35 m (95% L.C.). The spreadsheet for this example is shown below.
30
Symbol Ur U res U ref U temp U align
Source Repeatabili Resolution Reference Temp effec
Uncertainty in Correction of an External Micrometer ci Type Units Value Dist Divisor m 0.085 Norm 1 A B B B m m C m 0.5 Rect 0.21 Norm 0.2 Rect 0.24 Rect 1.73 2 1.73 1.73
ui 1 1 1
vi
0.085 2.90E+01 0.289 1.00E+10 0.105 3.00E+01 0.115 1.00E+10 0.139 1.00E+10
3E-01 1
Parallelism B flatness effects
0.349 2.000 8.09E+01 0.698 m
Example 5 Calibration of Glassware.

1
One
Mark
Volumetric
influence factors are critical in a laboratory doing its own glassware calibrations and deduce the sensitivity coefficients from physical principles. As always with an uncertainty analysis blunders are not allowed for so we will assume the laboratory has: read the meniscus correctly has allowed for the correct delivery time (where relevant) has cleaned the glassware has soaked both the vessel and the water in the ambient conditions for long enough.
Sources of uncertainty in the glassware calibration We will assume a testing laboratory is carrying out its own determination of the accuracy of its volumetric glassware. The easiest way to do this is to weigh the vessel full to the reference mark with water and convert this to a volume knowing the density at the water temperature. The equation relating volume to the weight indicated on the balance is approximated by:
V = 0.99985 xWb x( 1 + ( 20 Tw )) dw 0.0012
where
Wb
Tw
= weight of water indicated by the balance
dw
= volumetric expansion of vessel = water temperature = density of water
Type A Uncertainties 1 Repeatability is determined by making 10 measurements for each glassware type (preferably for a number of different staff members). If a lesser number of repeats are routinely made at each glassware mark during a normal calibration, then this number should be used to determine the experimental standard deviation of the mean i.e. instead of dividing the repeatability standard deviation by 10 we divide by n, where n is the number of repeats.
Type B Uncertainties The equation relating the balance indication is now:
Vm = Wb + Ct x Vn 1000
If we were really keen we could follow the GUM and partially differentiate this with respect to the important influence quantities such as the weight of water, the temperature of the air and water, the buoyancy, relative expansion of the glass and water and the density. Using the sensitivity coefficients obtained from this process, we would then evaluate the type B components associated with a correction for each of these influences and combine with the type A contributions to find the total uncertainty. Fortunately the equation above is rarely used to find the volume using the gravimetric method. Standards such as AS 2162.1 provide tables of corrections that include the important factors. We will work out what
Where: Vm Ct Vn
= The measured volume = the correction from the tables = the nominal value
Ct includes corrections arising from the buoyancy, relative expansion of the vessel and the water and also the density of the water. We will assume a worst case temperature uncertainty for the water being weighed and
31
take this as the total range of a rectangular distribution. Note that the 1000 divisor is included as the AS 2162.1 table gives corrections for a 1000 ml volume only. 2 Uncertainty in the correction. We will assume the water temperature changes during the calibration by 4C. The half range is then 2C for which the correction factor changes by 0.51 ml for a 1000ml vessel at 25C. Uncertainty in the balance indication of the water weight. We take this as the best accuracy of the balance and assume it is at a 95% level of confidence. Uncertainty in the density is assumed to be included in the correction factor uncertainty. Uncertainty in the thermometer used to read the temperature of the water and air.
Ubalance =
0.08 2
g = 0.04 ml
Although this figure is in grams the sensitivity coefficient is effectively 1. 2.4 Uncertainty in the thermometer As we have allowed a large variation in the correction factor with temperature this contribution will be negligible. 3 Combination of Uncertainties The uncertainties are combined using the root sum square method. The combined uncertainty is thus:
2 2 2 2 Uc = Ur + Ucorr + Ubalance
4 5
2 Uc =
(0.1)2 + (0.29)2 + (0.04)2 Uc = 0.31ml.
2 Determination of standard uncertainties 2.1 Repeatability of the balance. The standard uncertainty arising from random variation in 10 readings for a 1 litre flask. We shall assume a normal distribution with 9 degrees of freedom.
Ur = 0.1ml
4 Expanded uncertainty To calculate the expanded uncertainty, U, we need the coverage factor k for a level of confidence of approximately 95%. Because we have a large number of degrees of freedom we can assume k = 2. The expanded uncertainty is thus:
2.2 Uncertainty in the volume correction The uncertainty in the volume correction factor is given as 0.51 ml for a rectangular distribution with high degrees of freedom. Then:
Ucorr = 0.51 3 ml = 0.29 ml
U = 2 x 0.31 = 0.63 ml.

5 A model and the uncertainty budget The model (equation) for this example is shown above in the discussion about type B uncertainties. 6 Best Measurement Capability The BMC is not relevant in this case as we have assumed a testing laboratory is calibrating its own glassware. BMC is only applicable for calibration laboratories. The spreadsheet for this example is shown below.
2.3 Uncertainty of the weight indication The balance has a best accuracy of 0.08 g. We treat this as normal with 30 degrees of freedom thus:
32
Symbol Ur U corr U balance
Source Repeatability Resolution Mass uncert
Uncertainty in Indicated Volume of Standad Glassware ci Type Units Value Dist Divisor A B B ml ml mg 0.1 Norm 0.51 Rect 0.08 Norm 1 1.73 2 1 1 1
ui 0.100 0.294 0.040
vi 9.00E+00 1.00E+10 3.00E+01
0.314 2.000 8.63E+02 0.627 ml
33

As LAB C5 Metrology &amp; Calibration

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

As LAB C5 Metrology &amp; Calibration

Hochgeladen von

Copyright:

Verfügbare Formate

specific criteria for accreditation

Metrology and Calibration Laboratories 5

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

Laboratory Accommodation and Safety

Metrology and Calibration Laboratories

Equipment Management and Calibration

Metrology and Calibration Laboratories

Computer-Controlled Calibration Equipment

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

Compliance with Specification

Metrology and Calibration Laboratories

APPENDIX 1 Classes of Test

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

(l) (m) (n) (o)

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

APPENDIX 2 Equipment Calibration Intervals

Accelerometers Acoustic calibrators Anemometers Angle gauges - reference - working Attenuators

One One One five four three

three two five* three three five

three five five two two five

Time and spatial (temperature variations, recovery time, rate of ventilation).

five ten five two

Length bars - reference - working Levels - precision

three two three

five four five

three five three

five five three

three three one See IANZ Technical Guide No one

five five three

five five one six months two three

ten three five five five five five

five five two two five six monthly.

five five one

Thermometers (i) reference liquid-in-glass (ii) working liquid-in-glass*

Transfer standards, ac-dc

three five two two

two With regular intercomparisons - intervals to be based on history of performance.

Metrology and Calibration Laboratories

APPENDIX 3 Example: Scope of Accreditation

Metrology and Calibration Laboratories

APPENDIX 4 Approved Signatories and Other Staff

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

APPENDIX 5 Examples: Evaluation of Uncertainty of Measurement

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

Using the values from the uncertainty table (below) we get

(0.0117)4 (0.009) / 5 + (0.0038)4 / 15

U = 2.14 X 0.0117 = 0.025C.

Metrology and Calibration Laboratories

Metrology and Calibration Laboratories

Symbol U duc U spatial U ref U hyst

Source Repeatabili Spatial var. Reference Hysteresis

0.009 5.00E+00 0.003 1.00E+10 0.004 1.50E+01 0.006 1.00E+10

Combined Uncertainty: Expanded Uncertainty: k =?

0.012 2.140 1.41E+01 0.025 C

Uduc = 0.045 MPa Uduc = 45 kPa

Metrology and Calibration Laboratories

U = 2.09 x 47.5 kPa = 99.3 kPa.

As LAB C5 Metrology & Calibration

As LAB C5 Metrology & Calibration