| 摘要 |
Improved methods for the calibration, in particular for self-calibration, of an A/D or D/A converter with weighted network (CN) are proposed. Only a relevant part of the weights (C 0 , C 1 , C 2 , Cn) is calibrated by measurement. In addition, by iterative repetition of measurements used for the calibration a noise is used for increasing a resolution. Finally, possibilities for dealing with the offset are illustrated. Complementary equations are set up and the offset is eliminated by subtraction. If an equation necessary for calibration cannot be directly set up because of an overflow, this is resolved by using special binary codes which indicate which weights are enabled and/or disenabled, and their conversion.
|
