One method to characterize analogue to digital converters (ADCs) is to use a histogram, where Gaussian noise may be used as stimulus signal. However, a Gaussian noise signal that excites all transition levels also generates input values outside working range of the ADC. Modern signal generators can generate arbitrary signals. Hence, excluding undesired values outside the ADC full scale can minimize test sequences. Truncating the signal to the working range gives further advantages, which are explored in this paper. The Cramér-Rao lower bound and a minimum variance estimator for histogram tests with an arbitrary stimulus are derived. These are applied for truncated Gaussian noise and the result is theoretically evaluated and compared to untruncated noise. It is shown that accuracy increases for a fixed sample length and that variation over transition levels decrease.