Comparison of Least Mean Square Based Techniques for Spur Cancellation in Fractional N-Frequency Synthesizer

Abstract

Owing to scarcity of frequency spectrum, the frequency allocation must be provided more precisely. This can be achieved by use of the systems that can maintain precise frequency control. Fractional N-frequency synthesizer is a controllable system to generate the accurately defined signals at different frequencies. In this paper, we present the LMS (Least Mean Square) based adaptive filtering methodology for spur noise reduction in digital fractional N-frequency synthesizer and investigates the noise performance response of the system against different variants of LMS algorithm. In literature only LMS was implemented, we have not only implemented different variants of LMS but also compare their performance. A digital fractional N-frequency system have been implemented in MATLAB using SIMULINK tools and noise performance of system have been analyzed in terms of SNR (Signal to Noise Ratio), spur noise and jitter by applying basic LMS, normalized LMS algorithms in adaptive filter. On comparing the performance results, normalized LMS is found to be better and efficient algorithm as it improved the system SNR by 28.09 dBc and reduces the spur noise by 9.56 dBc along with 2.9 ps improvement in jitter.