Abstract: A digital to analog converter for converting an N-bit digital word into its analog representation including means for splitting the N bits into n sections of N/n bits each. For instance a 12-bit word is split into an odd section and an even section which are processed independently and in parallel. This results in two partial results, V.sub.i and V.sub.p, respectively, representative of the odd and even bit sections. The last step of the conversion is the action of the two partial results V.sub.i and V.sub.p to provide the analog representation of the 12-bit word. Few operators are required to process each section because each bit is converted sequentially. This provides a low cost, compact and simple converter, moreover, since few operators are required, it may be advantageous to use high precision operators as disclosed.

Abstract: A constant biasing circuit is made to maintain a fixed difference of potential between two circuit nodes A and B otherwise at floating potential levels. The biasing circuit is made of series connected complementary transistors disposed between the circuit nodes A and B and having their gate electrodes respectively connected to opposite voltages of, e.g., +5 volts and -5 volts.