Limits of accuracy in resistance measurement

The resolution that can be achieved in resistance measurement is shown to be limited by the amplitude of thermal noise relative to the input power of the resistance. The errors of measurement due to resolution and power loading are analysed and the minimum value of their sum is found to be proportional to the cube root of the circuit and detector noise factors, the detector bandwidth and the loadcorrection error. The noise factors of some common circuit arrangements are evaluated and the other errors of measurement in addition to resolution and power-loading errors are discussed. The total error is shown to have a limiting value near 1 part in 108 for existing high-quality standard resistors and this is illustrated by a practical example.