In A/D- und D/A-Umsetzern verwendete Netzwerke mit Korrektur von differentiellen Fehlern.

摘要

A digital-to-analog converter includes a decoding network and a plurality of output members such as capacitors. The decoding network receives a plurality of binary signals individually having logic levels respectively coding for binary "1" and binary "0" and individually coding for a binary value of an individually weighted significance and cumulatively coding for an analog value. The network decodes the logic levels of the binary signals and activates output members in accordance with such decoding. As the analog value coded by the logic levels of the binary signals increases, the output members previously activated in the plurality remain activated and other output members in the plurality become activated. The decoding network and the output members are disposed on an integrated circuit chip. Dependent upon their positioning on the chip, the output members have different characteristics which cause errors to be produced in the analog signal, particularly at low analog values. This invention compensates for such errors. The invention includes a second converter disposed on the chip with a construction substantially identical to the first converter and rotated on the chip substantially 180 DEG relative to the first converter. In this way, pairs of output members of the same binary significance may have, on the average, a median position in a first direction. Individual ones of the output members in the second plurality may have the same positioning, in a second direction co-ordinate with the first direction, as corresponding ones of the output members in the first plurality.