 |
 |
|
||||
|
||||
|
||||
|
|
|
||||||||||||||||||||||||||
|
|
|
|
||||||||||||||||||||||||||
|
|
|
|
||||||||||||||||||||||||||
|
|
Frequently Asked QuestionsMechanical Metrology FAQ1. How many pounds are in an Imperial Ton? 1. How many pounds are in an Imperial Ton?The answer is "some". The avoirdupois ton is spelled "ton" and comprises the Imperial system and the Canadian conventional system. The ton comes in two sizes: The short ton is equal to 2000 pounds (symbol for pound is "lb") Since all avoirdupois units are traceable to the kilogram, it is useful to know that one pound is EXACTLY equal to 0.453 592 37 kilograms (kg). The so-called metric tonne (spelled that way) has the symbol "t" and is merely permitted to be used with the Système International. Prefixes to t may be used only for multiples of the tonne. (eg. 1 kt is 1000 tonnes). One tonne is equal to 1000 kg.
2. Are measurements made by the Mechanical Metrology Group traceable?Traceability is defined as the property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons all having stated uncertainties. (Ref. International Vocabulary of Basic and General Terms in Metrology, ISO, 1993). The National Research Council (NRC) is the agency of the Federal Government of Canada which, under the legal authority of the Weights and Measures Act and the National Research Council Act, is responsible for the establishment, maintenance, development and improvement of the National Standards of Canada. The Mechanical Metrology Group of the Institute for National Measurement Standards (INMS) performs calibrations and measurements on standards of mass, length, force, pressure, density, volume, mass flow, and acoustical attributes such as sound intensity. All calibrations and measurements performed by the Mechanical Metrology Group are directly traceable to the National Standards of Canada, which are realizations of the base and derived units of the International System of Units (SI) maintained by the National Research Council of Canada. See also general FAQ on traceability and INMS' International Cooperation page.
3. What kind of uncertainties are provided with calibration reports issued by the Mechanical Metrology Group?In the past, it was the practice of the Mass Standards Group and its successor, the Mechanical Metrology Group, (MechMet) to specify uncertainties associated with values of variables as one standard deviation with an assumed normal or Gaussian distribution. This statement corresponds to a confidence level of approximately 68%. With the acceptance and publication of ISO Guide to the Expression of Uncertainty in Measurement (First Edition 1993) and with the growing international consensus on the provision of uncertainties with confidence levels of approximately 95%, this practice has changed accordingly. Unless otherwise explicitly stated, the uncertainties provided in calibration reports issued by MechMet shall correspond to a confidence level of approximately 95%. All such reports will bear a statement such as the following: Uncertainties specified in this report are expanded uncertainties representing a confidence level of approximately 95% obtained by multiplying the combined standard uncertainty (one standard deviation) by a coverage factor of k=2. For more detailed information, refer to ISO Guide to the Expression of Uncertainty in Measurement (First Edition 1993). It must be emphasized that this change in format does not in any way alter the accuracy of calibration values reported by MechMet. The uncertainties formerly reported are simply multiplied by the coverage factor k=2. When the reported uncertainties are to be used by the client in subsequent calculation of variances or uncertainties combined with information from other sources, it is important that single standard uncertainties be used. That is, the uncertainties provided by the Group should be divided by 2 and treated as Type B standard uncertainties for recombination, as outlined in the cited ISO Guide.
4. How often should my mass set be recalibrated?The optimum recalibration cycle for full sets of weights depends upon a number of factors which are out of the direct control of the INMS and, indeed, of the client in some cases. These factors include frequency of use, effective control over and training of the personnel having access to the weights, security and cleanliness of storage, the environment in which they are used and the quality assurance procedures that the client may have in place. Because of these many variables, a cycle time is chosen by agreement between the client and INMS Mechanical Metrology Group (MechMet) personnel. Since mass calibration is costly, both to the client and the INMS, it is in everyone's interest to ensure that any calibration performed by MechMet is both necessary and sufficient to maintain confidence in the accuracy of the client's reference weights. Many years of experience accumulated by MechMet suggests that calibrations of well maintained sets of stainless steel weights of NBS Classes M, S, and S-1 and OIML weights of Classes E1, E2, F1, and F2, are secularly stable; they change slowly with time. It is tempting, in light of this, to recommend very long recalibration periods for all weights. Such a course of action is open to the valid criticism that, for example, accelerated mass loss due to increased usage might not be detected until the next scheduled recalibration. Therefore, MechMet recommends the following calibration routine to all clients as a matter of policy. 1. When first submitted by a client, any set of weights will be calibrated in its entirety (all weights labelled A, B, C, and D in Table 1) and a recalibration period will be decided upon in consultation with the client. This period will typically be one year. 2. At the first recalibration, only the weights in group A will be submitted for recalibration and inspection. At this time any changes in the weights will be noted and may result in reevaluation of the calibration interval or, if necessary, recall of the entire set. 3. If the weights in group A are found to be stable, then at the second recalibration only the weights labelled B in Table 1 will be submitted for recalibration. 4. Steps 2. and 3. will be repeated for the weights in groups C and D after the third and fourth calibration intervals, respectively. 5. At the fifth recalibration, the process will begin again with the weights in group A. This recalibration scheme offers many advantages, including the following : 1. A considerable saving in calibration costs accrues to the client. 2. The client obtains a continual reassessment of any changes that may occur in the condition of a representative sample of the members of the set, including wear, accidental damage and density changes. 3. The client is never wholly deprived of what may be its only reference weight set. Sums of weights can substitute for any given weight in the most often encountered series. 4. Sampling can be done at somewhat shorter intervals to increase confidence while still ensuring substantial savings to the client. 5. The client is assured of a complete recalibration, including every weight in the set, at the end of the fifth calibration period. 6. The INMS will be able to make better use of its resources by optimizing the overall number of calibrations for each set of weights. 7. It is expected that the reduction in the required number of calibrations will result in shorter turn-around times and fewer scheduling problems for the client population. This calibration scheme does not alter MechMet's advice to clients to submit any weight for recalibration without delay when it is believed to have been dropped or to have experienced any other accident that may have altered its mass. Verification of single weights from calibrated sets can be arranged very quickly, with minimum disruption to the client's mass metrology operations. See Table 1: Groups of Weights for Three Common Series (10 kbyte pdf format)
5. How should I ship my dead weight tester to NRC/INMS for calibration?Clients submitting piston-cylinder dead weight pressure balances for calibration by the Mechanical Metrology Group (MechMet) of the INMS are requested to ship the piston-cylinder assemblies in a disassembled state, separate from the frame of the balance. The assemblies must be clean and, in particular, free of oil. Failure to comply with this request will be considered grounds for refusal by the INMS to accept the apparatus for calibration. This condition is made partly to reduce the unproductive effort required to carry out routine disassembly and cleaning operations, but, more importantly, to minimize the chance of damage occurring to the submitted components. Clients are reminded always to ensure that all components of any apparatus shipped to the INMS are carefully packed and that all components are securely fastened in place. Particular care should be taken with the piston-cylinder assemblies and, if at all possible, the packing of the equipment should be carried out under the supervision of the user/experimenter. For further information, contact Anil K. Agarwal, Telephone: (613) 991-0615
6. What is a control chart and how do I use one?A control chart is employed to monitor the time rate of change of the mass of a particular weight. It is a plot of the corrections to the nominal mass and their associated standard uncertainties as a function of time. Such a plot is used to detect the presence and magnitude of drift due to accretion of contaminants on the surface of the weight (positive) or loss of mass due to wear or other damage (negative). It is particularly useful in monitoring the stabilization of a weight after cleaning. Control charts should be kept for all reference weights and particularly for check weights. Check weights are well characterized weights which are calibrated along with unknown weights in order to objectively assess the statistical control of the measurement process. It should be understood that two or three measurements do not constitute a control chart, but the start of one. As time progresses and more points are added, the usefulness of the chart becomes apparent. When a sufficiently large number of points are logged, the chart provides an objective base from which to assess whether or not an additional point may be regarded as being in statistical control. This assessment may be made as follows. Measurements of the corrections, which may be made as conventional or absolute masses, but not mixed, are plotted along with their standard uncertainties, represented as "error bars". In addition, as points are accumulated, the mean of the measured values is also plotted. The uncertainties associated with these mean values are the sample standard deviations (sigma) or some appropriate multiple of them. The multiple or coverage factor is often chosen as 2 or 3 so that the uncertainties reflect approximate confidence levels of about 95% or 99% respectively. Once such a mean and uncertainty are established, the latter may be selected as an objective criterion against which each new value may be assessed. That is, if a new datum lies within, say, three standard deviations of the cumulative mean then it may be regarded as being in statistical control. It may then be added to the data stream and contributes to the next computed mean. It is important to remember that the uncertainty of the mean is the standard deviation of the sample, s, and not the standard deviation of the mean, sm. If n is the number of measurements, these two estimates are related by sm = s/(n)1/2 . (1) As n increases, sm becomes progressively smaller, but s will reach a stable value that should not change much as long as the measurement process is stable. The value of sm, on the other hand, better characterizes the current knowledge of the mass of the weight. As an example of the use of control charts, consider the control chart which monitors the stability of an NRC 10 g weight which is used as a check (closure) weight and as a reference for the calibration of weights in the 1 g to 5 g range. The diamonds (Series 1) represent individual determinations of the absolute mass of the weight in terms of 100 g reference weights. The within group error bars have been omitted from these points for clarity. The squares (Series 2) represent the mean values of all preceding measurements and the associated error bars represent the standard deviation of the sample. Twenty four measurements have been made and four, or 16%, lie outside the one standard deviation interval. This is consistent with what should be expected of a normal distribution for a sample of this size, so it may be concluded that the measurement process is in statistical control. The straight line is a linear regression fit to the individual measurements. Its slope, compared to the standard uncertainties of the mean values, is insignificant. This implies that no significant drift exists for this weight. This graph and the associated calculations were generated using commercial spreadsheet software which easily handles the generation of the required statistics, plotting the results and performance of linear regression whenever necessary. This chart will be maintained into the indefinite future, alerting the NRC to any changes in the standard or the measurement process.
|
||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||
General FAQ
