Dissertation Talk: New RF Transmitter Techniques for RFID, Cellular, and mmW Applications
Presentation: Dissertation Talk: Berkeley Wireless Research Center (BWRC): EE | May 7 | 3-4 p.m. | 2108 Allston Way (Berkeley Wireless Research Center), Rabaey room
Nai-Chung Kuo, University of California, Berkeley
Transmitter/reader designs for inductive-power-transfer (IPT) RFID systems have to power up the tag and also listen to it, and low power transfer efficiency (PTE) and high Tx-to-Rx interference are the two main challenges when a tiny tag is involved. To improve PTE, an analytical approach for optimizing the Tx coil and miniature rectenna has been developed. With the IPT frequency in the GHz range, the non-uniform current on the Tx coil and the radiation resistance have to be considered. The equation-based approach optimizes the IPT design rapidly, and the Tx-coil and the rectenna FOMs co-decide the optimal IPT frequency and the coil geometries. An optimized 2.2-mm IPT is realized at 4.8 GHz. Under a total RF power of 33.1 dBm, the designed CMOS rectenna, with a coil size of 0.01 mm2, harvests the designed dc power of 0.1 mW. On the other hand, in order to suppress the Tx-to-Rx leakage, we have invented a new TRx architecture with a two-tone Tx that simultaneously charges the tag and excites a third-order intermodulation (IM3) frequency from the tag. The IM3 tone is modulated by the tag data and transmitted back to the reader Rx via the same coupled-coils. The Tx/Rx frequency separation allows significant filtering on the Tx-to-Rx leakage and noise. This technique was adopted first in our IPT system, and our following works have extended it to far-field systems with a custom-designed tag and even a commercial UHF Gen2 tag.
The other key application of RF Txs is cellular communication, where the Tx has to accommodate multiple bands over a wide range of operation frequencies. We have fabricated an all-digital CMOS Tx core on three high-density-interconnection (HDI) interposers with high output power, >50% efficiency, and a collective bandwidth from 0.4 to 4 GHz. To further achieve frequency reconfiguration within a single-output module, a band-selecting HDI interposer was designed to combine three identical CMOS Tx cores. The band selection is carried out by reconfiguring the switching devices in the CMOS power amplifiers. Peak power higher than 23 dBm and drain efficiency better than 25% are achieved from 0.4 to 4 GHz, by rotating the three sub-Txs. 64-QAM, WLAN, and LTE modulation tests at 0.85, 2.1, and 3 GHz for the reconfigurable Tx package verify the functionality of universal standard adaptation. Finally, we will present a 1-GS/s E-band QPSK Tx element in 28nm bulk CMOS. The embedded PA has a saturation power of 15.7 dBm at 78 GHz and power added efficiency of 9%. The Tx element is a suitable building block for digitally-modulated phased array and high-speed communications. Attractive features of the digitally modulated phased array will be elaborated, including the improved efficiency, less heat dissipation, and the capability of synthesizing pattern null at a undesired direction. The leakage-suppression technique exploits the combination redundancy as the array synthesizes the desired waveform, so high-resolution magnitude and phase adjustments for the array elements, indispensable for conventional phased arrays to create pattern null, are not required.
CA, email@example.com, 7348460759