Periodic Analysis of Adaptive LMS Filter in TIADC

A periodic analysis of LMS adaptive filter in TI-ADC is presented. TI-ADC is a efficient technique to reach high sampling frequency but mismatches among different sub-ADCs influence its performance. LMS adaptive filter is a finite impulse response (FIR) filter which adapt itself to different inputs once implemented. To analyse SFDR performance of LMS filter in time domain, converged weigh vector and its analysis frequency domain are done theoretically and emulationally. Simulation shows that our theoretical analysis is consistent with the practical results that the useful range of SFDR performance is only negatively related to filter length if sampling frequency is constant. Further, it is concluded that there is a trade-off between maximum of SFDR in LMS filter and its useful range in terms of frequency.