High-precision auto-calibrating D=A converter - periodically tests output against standard and performs correction on a continuous basis

摘要

D-A converters convert the 11 bits of least weight of a 16 bit digital number into the sum of currents (L1, I2) with a precision of 1/2x2(power 17) of the current due to 1 of the weightiest of the 16 bits. A 5 bit DAC converts the weightiest 5 bits into a current with a precision of only 1/2(power 12). Whenever these 5 bits are at the mid-values 10000 or 01111 the output of a 5 bit counter (32) replaces them as input to the 5 bit DAC. The current now corresponds to each of the 31 non-zero test inputs in turn; it is compared with a reference current to ascertain whether it is a multiple. If not a multiple, any discrepancy is stored digitally in a RAM. These correction numbers provide a correction current via a correction converter, to the current to bring it to the required precision. A substitution current replaces the current whilst it is disconnected for calibration of the 5 bit DAC.