ABSTRACT
To illustrate an inherent weakness in EVA using signed sensitivities, another approach to the EVA problem is shown. In this method, all possible tolerance combinations are used. To see how this is implemented in Mathcad, suppose one were to throw three coins into the air: a quarter, dime, and nickel. What are all the possible combinations of heads and tails for the separate coins? We have TTT, Till, TIIT, 11-IH, ••. etc. The reader might recognize that a 3-bit binary counter is being formed, with T = 0 and H = 1. Then to get all possible min and max tolerance combinations for a circuit with three components, there are 23 = 8 possible combinations. So we create a binary counter and usc the low tolerance for a zero and the high tolerance for a one.
