Cellular handsets comprise the largest consumer market in the world. It's even larger than markets for some of the most common electronic devices, such as wristwatches and televisions. Service providers, handset makers, and software and IC suppliers are all aggressively pursuing this high-growth market.
The persistent trend toward integration and cost savings is key in all communications applications, but it is intensely acute in the wireless industry, where bill of materials (BOM) costs in cellular handsets are expected to decline dramatically over the next several years. This begs the question, how will ICs continue to enable further integration and reduced BOMs?
As handset makers seek to maximize functionality while reducing costs, CMOS is emerging as the process technology of choice. ICs implemented in standard CMOS are quickly replacing products implemented in more exotic processes like silicon bipolar, biCMOS, and silicon germanium (SiGe). Generally two to three line-width generations ahead of other processes, CMOS lets designers integrate complex analog functions with digitally intensive circuitry in a monolithic, cost-effective, silicon solution.
Driven by the memory and microprocessor markets, a vast majority of the world's semiconductors are already implemented in CMOS. The digital baseband block of cellular handsets is no exception. Basebands have been implemented in CMOS for several product generations. As a result, they have benefited from the highly available technology and are riding the technology curve of a process used worldwide for digital ICs.
Despite the benefits of CMOS for digital ICs, many believed that implementing complicated mixed-signal RF circuitry in CMOS was impossible because it lacked the requisite horsepower. This perception isn't new. It has been around since the early '80s in other applications like telecom codes (1980), T1 line interface ICs (1986), Ethernet transceivers (1992), and magnetic read channels (1993). These applications were implemented in what critics considered "more suitable" processes than CMOS. It took the resolve of a few pioneering companies to develop these applications in CMOS. Once this eventually occurred, CMOS became the de facto standard for these applications.
History has once again repeated itself with the introduction of the first 100% CMOS RF synthesizers and GSM transceivers. CMOS RF solutions are being shipped worldwide to over 20 handset providers and already represent about 15% of the market.
CMOS enables the technology leap required to meet the challenging integration and cost savings goals of the handset manufacturers. Triple-band transceivers using alternative technologies for handsets still require as many as 60 discrete components in the RF front end. By comparison, a GSM transceiver in CMOS reduces that total component count to only 15 components.
Handset designers choosing a CMOS solution benefit from $1 to $2 BOM savings per handset by eliminating about 75% of the discrete components in addition to the incremental costs of selecting, designing, stocking, assembling, and testing all of those components. Extrapolate those savings over an entire product line, or the entire available market, and the potential cost savings created by CMOS solutions are immense.
CMOS use in cellular applications is a reality. The integration, size, and manufacturing cost advantages offered by this proven process technology mirror the goals of the products it was designed to enhance. Current CMOS solutions fold in the RF front end from antenna to digital baseband with only a handful of supporting external discretes. In addition to increasing functionality while containing system costs, these RF innovations in standard CMOS edge us ever closer to the long-held industry goal of a single-chip radio, creating endless possibilities for cost-effective, highly integrated cellular applications.