A/D converter circuits based on Hopfield neural network are studied in this paper.The besic operation mode of A/D conversion and the stable state solution of the circuit are compared.Then the relation between the output state and the reference input of the neurons is discussed.The reference input conditions for correct output states are derived.Two approaches of modifying A/D converter circuits are presented:1.Appropriate selection of reference inputs in the original circuit

2.A new circuit based on the concept of the Hopfield linear programming circuit.PSPICE computer simulation results are also given

and in accord with the theoretical analyses.