Abstract :
Systematic errors, associated with coil reversals, in the determination of the absolute unit of resistance by Albert Campbell´s a.c. bridge method, originally noted by Hartshorn and Astbury and recently confirmed by Rayner, can, it is shown, be explained by uncompensated eddy-current coupling between supply and detector circuits. It is also shown that the errors are largely, but not necessarily completely, eliminated when a mean value is taken. A new bridge is proposed in which, by measuring an additional mutual inductance, the effect can be eliminated without external compensation: it is also indicated how this might be done, less conveniently, on the Campbell bridge. A model of the new bridge, operating at 1kc/s, has been used to demonstrate experimentally the introduction of systematic variations associated with coil reversals, the eddy-currents being produced under control in a tertiary circuit.
