The selection of an efficient weighing design is a common problem in mass metrology. The calibration of a set of nominally equal masses in terms of one or more reference masses of the same value does not normally present any difficulty, but designs for extending the calibration of a kilogram mass to multiples and submultiples of the unit are more complex. A selection rule that is very widely used is that the design matrix be orthogonal. However, orthogonality is advantageous only when all the weighings in the design have the same variance. In accurate work this is generally not the case, as the variance of balances is load-dependent and furthermore it is often necessary to use more than one balance. Designs which make better use of weighings of non-uniform variance have been studied, and a practical approach to the construction of such designs is given.

No abstract

Long-distance transport of high quality mass standards for the purpose of calibration or international comparison is becoming increasingly common, and the stability of these standards in transit is of universal concern. Despite efforts to minimize disturbance, significant mass shifts have occurred during bilateral intercomparisons involving the National Measurement Institute Australia using 1 kg travelling standards of a quality equal to national secondary standards. Mass shifts have also been observed at the E 1 level.This work explores the extent to which transport and storage during transport might bring about these mass instabilities. The weights, in their boxes or transit capsules, were subjected to controlled variations in temperature, pressure and humidity and were measured before and after these environmental stresses were applied. The stresses varied from relatively mild to the very worst that could be expected during transport and storage.This study indicates that, with modest precautions, weights of class E 1 or lower can be transported by courier. If significant mass shifts do occur at this level, one can be reasonably confident that they did not happen in transit. On the other hand, a higher quality kilogram weight exhibited a mass shift in the order of the measurement uncertainty (30 µg) during a period of rather extreme environmental stresses. The common practice of 'hand-carrying' such high quality standards is therefore recommended. The effect of vibration in transit was not studied here.