Last month, in this column, I focused on CMOS and BiCMOS technology making strong inroads into the microwave territory and extending its reach right into the millimeter wave region. In fact, another RF trend that was also the subject of discussion at last month's International Solid-State Circuits Conference (ISSCC) was the integration of various devices above the CMOS and BiCMOS RF ICs. And this is being driven by the fact that there is tremendous pressure to reduce external passive components and cut bill of material cost for transceivers and other communication circuits. Several presentations at the conference were indicators of this new trend in the RF IC arena.

Much of this development is being driven by advances in RF MEMS devices. In this column last year it was disclosed that a handful of developers have taken their respective RF MEMS devices (switches, duplexers, filters, etc.) to full production. And, a few were reading prototypes of integrated RF front-end modules and subsystems using their RF MEMS parts and combining them with power amplifiers and other passives in a single package. Agilent Technologies, Intel, IBM and XCom Wireless, among others, are key drivers of this integrated trend.

But with advances in RF foundry services for MEMS-based passives and packaging technology, some developers are taking the next step. Integrating on-chip MEMS-based high Q inductors and tunable varactors, together with an array of parallel cantilever beams combining mixers and channel filters in a single process above the traditional CMOS or BiCMOS chip. Consequently, such designs of wireless circuits incorporating above-IC bulk-acoustic-wave (BAW) resonators and filters opens the door to highly integrated transceiver architectures for multiband and multistandard applications.

Using MEMS, integrated high-quality-factor BAW resonators and filters have been developed at CSEM (Swiss Center for Electronics and Microtechnology), Neuchâtel, Switzerland and CEA (French Atomic Energy Commission), Grenoble, France. Using a process that integrates CMOS, bipolar and BAW devices, the researchers have shown that RF passive and active devices can be integrated on a single chip. In this design, the researchers described an RF front-end double-lattice BAW filter with balanced input and output at a center frequency of 2.14 GHz. The filter is integrated directly above a 0.25 µm BiCMOS RF IC. Insertion loss of -3 dB and out-of-band rejection of better than -50 dB was achieved. The BAW is integrated above an LNA chip. Likewise, in another paper at ISSCC 2005, researchers from STMicroelectronics in partnership with the French CEA and CSEM laboratories demonstrated the feasibility of a fully integrated RF front-end for wideband CDMA (WCDMA) applications using above-IC BAW filters. Designed and fabricated in a 0.25-µm BiCMOS SiGe:C technology that is enhanced with above-IC capabilities, this experimental chip is the first application of a filter on top of an RF IC. It achieves a gain of 31.3 dB, a 5.3 dB noise figure, and consumes a total power of 36 mW. The BAW filter area is 0.45 mm² and the total circuit area including the BAW filter is 2.44mm².

Also, scientists from the University of California in Berkeley, Calif. take advantage of the high-Q of MEMS-based film bulk acoustic resonator (FBAR) BAWs to implement a very-low-power super-regenerative transceiver for wireless-sensor-network applications. The receiver operates at 2 GHz and consumes only 450 µW from a 1 V supply, while achieving a -100.5 dB sensitivity at 5 kb/s for a 10^-3 BER. The circuit is implemented in a 130 nm CMOS technology and uses off-chip FBARs.

Meanwhile, researchers from Carnegie Mellon University showed that advances in RF foundry processes for micromachining have led to high inductor and capacitor quality factors, improvement in varactor tuning range, and creation of electromechanical mixer-filters that downconvert RF signals from GHz to MHz frequency band with built-in frequency selectivity.

In reality, though the above-IC approach offers a higher integration alternative, cost will ultimately determine whether this 3D technology will become a viable mainstream technology.