A method involving a computer is described for selecting resistors for precise analogue instrument circuits. A statistical analysis is carried out which provides formulae for calculating the probability of non-duplication of resistors, and also the predicted discrepancy for specified group and batch sizes. Theoretical and actual values are compared: the agreement is such that the theoretical formulae can be used confidently to specify batch sizes to obtain indicated discrepancies for given group sizes. The method is of practical importance in selecting closely matched resistor ratios, and matched resistors, for precise analogue circuits.
