ADAPTIVE NOISE CANCELLATION BASED ARCHITECTURE FOR CORRECTION OF GAIN AND OFFSET MISMATCH ERRORS IN TIME-INTERLEAVED ADC

Analog-to-Digital Converters (ADCs) represent a main bottleneck for realizing highspeed
telecommunication systems such as cognitive radio and Software-Defined Radio (SDR)
systems in addition to various requirements of transmission quality. Time-interleaved A-D
converter (TI-ADC) may be considered as an effective candidate to achieve high-speed ADC with
relatively slow circuits accounting for digital spectrum management. However, mismatch errors
cause a considerable degradation in the performance of TI-ADC. In this paper an adaptive
compensation technique is proposed for improving the overall performance in presence of offset
and gain mismatches. The proposed structure behaves independent from input signal. Here, Least-
Mean-Squares (LMS) algorithm has been exploited to adaptively estimate and correct mismatch
errors. The Proposed correction method is structurally very simple and hence suitable for
implementation on integrated circuits. Besides, proposed digital compensation algorithm is
computationally efficient. Considering a four-channel TI-ADC, the proposed compensated TIADC
provides an approximate improvement of 33.2dB in the performance compared to
uncompensated architecture. The proposed compensated TI-ADC would be an efficient high speed
ADC in most RF communication applications, and especially applications that deal with tuned
frequency.