Switched-capacitor (SC) simulation of Hopfield type neural networks

A switched-capacitor (SC) structure simulating a general-purpose Hopfield-type neural network is proposed. Existing simulation networks use current summation to achieve the summation effect. In the proposed scheme, this effect is obtained by employing voltage summation achieved with passive SC three pots. The only active components used are the low-gain voltage buffers. Consequently, the structure is capable of operating at high speeds. The scheme is, however, best implemented in MESFET technology.<>