Blocker-Tolerant RF-to-Digital Linearization With a Very Nonlinear Auxiliary Path

This paper introduces a new linearization technique based on signal-component cancellation to remove the third-order lumped nonlinearities from the entire receiver chain using a nonlinear auxiliary path. The paper shows that the solution performance improves when the auxiliary path becomes more nonlinear, a rather counter-intuitive property of the proposed technique with several practical implications. The signal-component cancellation uses two coefficients that are found adaptively with a novel two-step LMS algorithm in the digital domain. In contrast with previous prominent approaches based on an auxiliary path, the proposed linearization has two significant advantages: 1) it does not estimate the signal nonlinearities used in other techniques for cancellation, and 2) upon convergence of the adaptive algorithm, the digital solution is input-signal independent and therefore is static in steady-state.