ABSTRACT
In 1946, E. Blade performed a limited study of mass assignment errors caused by convection forces in measurements related to chemistry. The M. Glaser experiment was performed with two 20-g masses and two 20-g tubes of nonmagnetic stainless steel with well-polished surfaces. One of the masses and one of the tubes were kept inside the balance chamber of a one-pan unequal-arm oscillating-beam balance. The mass kept inside the balance chamber was used as a reference. The balance used was a single-pan, electromagnetically-compensated mass comparator with a capacity of 1 kg, a standard deviation of less than 5 µg, and a resolution of 1 µg. The comparator was equipped with an automatic mass exchange device and a turntable with different positions. The physical models were used to extrapolate the experimental results to predict, for weights of similar shape but other nominal values, the change of apparent mass and waiting times for given residual deviations and residual temperature differences.
