A research on the fast correlation-based background calibration techniques for the pipeline ADCs

Abstract

The problem of slow convergence speed exists in the correlation-based digital background calibration techniques for the pipeline ADC. The proposed improved techniques reduce the signal-dependent interference amplitude by clipping MSBs off the ADC output. Meanwhile, the exponential averager is introduced into the LMS iteration. These means can not only calibrate capacitors mismatch in MDACs and the finite gain and bandwidth of the interstage opamp, but also converge more than 20 times faster than normal correlation-based calibration techniques. A behavioral model of 12bits pipeline ADC with 2% interstage gain error in the first stage is presented to verify the efficiency of the calibration techniques. With calibration, simulation results show that SFDR enhances from 69dB to 98dB and SNDR enhances from 53.8dB to 73.5dB.