Introduction

In the previous chapters, conventional M-phase filters that can provide a high-Q bandpass filter with a center controlled by clock frequency were introduced. The evolution of the conventional M-phase filter to other useful forms such as high-Q image-rejection filters and high-Q filters with their centers controlled by sum or difference of two clocks was also introduced. In this chapter, we will show how various structures of the previously described M-phase filters have been used in a fully integrated superheterodyne receiver.

Zero-IF or low-IF architectures are two of the dominant architectures for today's integrated receivers, because these two architectures are simple and offer the highest level of integration. For such receivers, image rejection is not a big concern [5, 42, 64, 66, 70], and channel selection is carried out through low-frequency low-pass filters after the downconversion mixer [Fig. 6.1(a)]. For multimode applications, the low-pass filter must be reconfigurable, and depending on the receiver mode, its channel bandwidth can vary by as much as a decade or more. The dominant structural choice for the low-pass filter is active-RC in which the area is a strong function of the filter order, its maximum allowed noise contribution, and the smallest channel bandwidth that the low-pass filter must cover. On the other hand, the bandwidth of the active-RC-based low-pass filter is dictated by the RC time constant. Therefore, to overcome the process variations, the capacitor units of the low-pass filter are composed typically of arrays of switched capacitors with a total capacitance close to 50% larger than the nominal value.