Architecture and algorithm for ultrahigh precision image rejection mixer and spurious rejection mixer using digital compensation

Abstract

A weaver-type ultra-high precision image rejection RF mixer architecture composed of analogue RF and digital IF mixers is proposed, along with the error compensating algorithm for various gain and phase mismatches. To estimate exact mismatch, we need the procedure to control the gain of one between I and Q signal path and to control the phase of quadrature LO used in the 2nd quadrature down-conversion. Also, for this precise and fast estimation, we induced global minimum equation. A prototype 2.4GHz image rejection mixer using this technique with an image rejection ratio > 80dB is demonstrated. The compensation structure satisfying high image rejection in all BW is proposed. The BW satisfying 70dB image rejection could be increased into more than 12MHz. The merits of the proposed architecture are the simple estimation algorithm, the higher image rejection ratio than that of conventional architecture, the conformation of global minimum equation, and simple compensation structure. Because our proposed algorithm had $2^nd$ order expression according to estimation variables, the gain and phase mismatches were estimated, independently. The demerits of the proposed architecture is the burden of increasing the power consumption in ADC for increasing BW, because our architecture needs more than 4 times over-sampling than maximum frequency in BW. The spurious rejection architecture, which is similar with principle of image rejection architecture, is proposed. The conventional transmitter doesn’t have any solution about the mismatch of quadrature modulator. But, our architecture includes the mismatch estimation block, and compensation block. Like as image rejection architecture, we induced global minimization equation in estimation block, and obtained high spurious rejection characteristic through estimating the exact mismatch from this equation. Our global minimization equation in estimation block has form of the $2^nd$ order polynomial with mismatch estimation variable...