Efficient analog multiband channelization for bandwidth scaling in mm-wave systems

We consider analog multiband as a means for scaling up the bandwidth for millimeter (mm) wave communication. Analog multiband sidesteps the difficulty of scaling analog-todigital conversion (ADC) to higher sampling rates by channelizing the available bandwidth in the analog domain into subbands and using existing power-efficient ADCs to digitize each subband. In this paper, we address the issue of efficient channelization into subbands. A direct approach using a bank of mixers with independent frequency synthesizers has several disadvantages, including large power consumption and the potential for oscillator coupling. We explore an alternative approach based on polyphase sampling, using analog discrete Fourier transform (DFT) along with appropriately designed baseband filters. We quantify the inter-subband interference as a function of filter choice, and demonstrate that it can be handled using interference suppression strategies developed in our prior work. A natural comparison of such approaches, where analog channelization is followed by parallel ADCs running at slower rates for each subband, is with time-interleaved ADCs, which also use polyphase samplers, but use parallel sub-ADCs to digitize the entire band. We show that the proposed approach reduces the required dynamic range of the subband ADCs in comparison to the TI-ADC.