In 1975 Barlow & Campo presented the scaled TTT-transform and the TTT-plot, which was based on the total time on test concept introduced by Epstein & Sobel, as tools for identification of failure distribution models. Further applications are described in i.e. Bergman & Klefsjo (1984). When making a TTT-plot based on a set of censored data, one must estimate the survivor function using the Kaplan-Meier estimator (KME). In 1980, a new method for estimating the survivor function from a set of censored data, called the piecewise exponential estimator (PEXE), was proposed by Kitchin. These two estimators are compared when making a TTT-plot, to see which of them is best suited for identification of failure distribution models. The authors study the estimate of the optimal age replacement interval, when using the PEXE-technique to make the TTT-plot (6 refs.)