Abstract Since they were first employed in nuclear pulse-height analysis, analog-to-digital converters show an evolution characterized by a continuous increase in their amplitude resolution. This trend looks as if it will continue in the immediate future but this forecast disagrees with the fact that the circuit complexity of the ADCs realized by conventional techniques rises sharply beyond 12 bit resolution. This paper describes and proposes a new A/D encoding technique. A successive-approximation ADC operating over a limited input range and having correspondingly a reduced resolution is employed. In order to adapt the input signal to this dynamic limitation, the signal is at first roughtly and quickly estimated by a flash converter and then, if necessary, an analog level is subtracted through a DAC. The input digit of this DAC is added to the result of the successive-approximations conversion to get the final correct result. The inherent differential non-linearities of the successive-approximations ADC and of the DAC are avoided by the combined actions of a sliding scale circuit and of an innovative circuit named “shaker”. The satisfactory performance of the encoder has been experimentally verified through a 13 bit prototype. The simplicity of the proposed technique in comparison to conventional ones leads to the possibility of further raising the standard performance of nuclear spectroscopy ADCs.