GB1190631A - Error Correction Circuits for Analog Signal Processing. - Google Patents

Abstract

1,190,631. Analogue-to-digital converters. GENERAL ELECTRIC CO. 2 Aug., 1967 [21 Oct., 1966], No. 35498/67. Heading G4H. An error correcting circuit for an analogue-todigital converter comprises an error storage register into which the lowest 5 bits are entered from the 11-bit register of the converter, a test source which produces a zero number in the test register when the converter is accurate, a digital-to-analogue converter responsive to the error storage register to generate a correcting signal which is applied to the converter, the test source being connected to the converter input alternately with the unknown analogue signals by a multiplexing circuit. The multiplexing circuit comprises switches 10-14, e.g. fieldeffect transistors, connecting the various inputs to a pulse length generator 30. The output of the generator 30 is differentiated to produce a sharp read-out pulse at the end of the pulse length to transfer the value then standing in a counter 20 (which counts clock pulses fo from the beginning of the cycle) to the corresponding register 21, 22, 23 &c., or error store 40, selected by the closing of appropriate switches 26, 27, 28. The input signal is biased by the negative reference voltage - VR which is set at half the full-scale value. The value - V R therefore gives a zero reading and +V R a full-scale reading. The sign of the input signal is shown by the highest digit (2<SP>10</SP>). This digit and the five lowest digits are entered into the error store 40. When the test signal is applied the error register should produce five 0's. If any of these digits is not zero a corresponding correction voltage is applied to the operational amplifier of the pulselength generator. The value in the error storage register is displayed on lamps.