A heuristic thermistor model

Abstract

Negative temperature coefficient thermistors, traditionally characterized by a nonlinear relationship in the resistance-temperature space, are shown to be intrinsically linear when that characteristic is mapped into a space defined by an appropriate transform. Therefore, information formerly conveyed by scores of tabulated ordered sets is contained in a pair of first-order equations which define resistance as a function of temperature and the converse. The necessary transforms and equations are defined; and, procedures for calculating the coefficients from manufacturers data using regression techniques are delineated. The ambiguity inherent in estimating parameters by statistical methods is discussed. Maximum mean ambiguity (5 ° C) occurs when the model is applied over a span of 400 ° C ; this figure is reduced to 0.1 ° C for a span on 100 ° C . The ambiguity continues to decrease as the span decreases.