In spite of the proliferation of microcell basestations, cellular handsets and their requirements still get all the attention and have all the glamour. Semiconductor makers are climbing all over one another to provide RF and baseband devices with high levels of integration and lower power consumption. And suppliers are competing ferociously to show how their devices will reduce the size and increase both the functionality and battery life of the next-generation handset.
It shouldn't be any wonder: Something like 162.8 million cellular handsets were shipped to users in 1998, according to Matt Hoffman, cellular industry analyst at Dataquest, who presented at the CTIA Wireless '99 conference in New Orleans in February. Nokia became the No. 1 supplier of cell phones, having shipped 37 million. Motorola, the No. 2 supplier, shipped 32 million handsets, while Ericsson shipped 24 million. Panasonic delivered between 12 million and 13 million handsets, while Alcatel shipped between 7 million and 8 million, according to Hoffman.
The drivers for this phenomenal growth are the "digital one-rate" calling plans offered by service providers like AT&T Wireless and Cellular One-as well as burgeoning cell-phone use in Asia. China in fact accounted for the largest portion of Ericsson's 1998 business, according to Lars-Goran Lundlad, manager of sales for Ericsson Components AB (Kista, Sweden). Ericsson's business in cellular basestations amounted to $10.6 billion in 1998; wireline systems were $6.2 billion, while cellular handsets accounted for $5.6 billion in revenue. Of that, China accounted for 12 percent, the United States absorbed 9 percent of Ericsson's output, the U.K. took 8 percent, while Brazil and Italy (Ericsson's fourth and fifth largest customers) each took 6 percent.
Among the cell-phone suppliers-Nokia, Ericsson and Motorola-market share seems contingent on the ability to deliver increased functionality in an appealing package with increased battery life and reduced cost. The Nokia 6160, for example-the darling of the AT&T Wireless Digital One Rate plan-includes a built-in alphanumeric phone book, calendar, links to voice-mail services, caller ID and paging functions, and even the ability to entertain with some rudimentary games. The device fits in a shirt pocket, and provides a 50-hour battery life in standby.
The semiconductor manufacturers have done a stellar job here: Texas Instruments Inc. (Dallas), whose C54x DSP core is used by both Ericsson and Nokia, is the No. 1 supplier of digital baseband processors-and the No. 2 supplier (behind Motorola Semiconductor Products Sector) of cell-phone ICs, according to Dataquest.
New-generation DSPs promise increased processing power for the handset, as evidenced by StarCore's SC100 (a joint project of Motorola and Lucent Technologies Microelectronics Group), the Infineon (formerly Siemens' semiconductor group) Carmel, the Philips R.E.A.L., STMicroelectronics' ST100 and the DSP Group's TeakDSPCore and PalmCoreDSP, as well as iterations of the TI C6x and C54x family. They indeed will pave the way for the increased data rates and multimedia services (video and Internet browsing) promised with 2.5 and 3G wireless technology.
Yet there are indications that the integration provided by digital CMOS will only bring us so far. Of the four main cell-phone components identified by Ericsson's Lundlad-the EL display driver, the RF power amplifier, the RF ASICs and baseband ASIC-only the last item lends itself to higher levels of integration. (Power-management devices like voltage regulators, switches and controllers constitute a fifth major component. And this does not lend itself well to integration, either.) The reason is that special processes such as gallium arsenide (GaAs), silicon germanium (SiGe) bipolar or BiCMOS are needed to handle the power and linearity requirements of the RF ASICs, the amplifiers, display drivers and power regulators.
Manufacturers like Ericsson Components and Texas Instruments are trying to integrate RF transceiver elements-frequency synthesizer, mixer and oscillator-onto a single device using BiCMOS. Other manufacturers are attempting the integration in CMOS.
But so far the results have been limited because the RF transceiver cannot integrate or eliminate all the passive components-largely inductors and capacitors-required to tune the RF circuit and its intermediate-frequency (IF) stages. The miniaturization embodied in new-generation cell phones is a consequence of LTCCs-multilayer ceramic packaging, which embeds capacitors and spiral inductors on the substrates-not IC integration.
At this year's International Solid States Circuits Conference (ISSCC), participants of an evening panel on the single-chip phone were generally not optimistic about the prospect. Cell-phone builders like Werner Gruber, vice president for new technology sourcing at Nokia Mobile Phones (Salo, Finland), and Ken Hansen, vice president of the technical staff at Motorola's Wireless Integration Technology Center (Austin, Texas), hesitated to say that integration could be accomplished any time soon. Gruber said that he had to deliver a $50 phone, one that would support multiple standards, and that his customers didn't really care if it was a one-chip solution or not.
Despite a wealth of ISSCC papers exploring RF integration in CMOS and BiCMOS, Hansen hesitated to name a single technology that could integrate all the divergent requirements of RF/IF components. He said there were more than 400 parts in the typical cell phone, and that maybe 10 of them were ICs. "We can reduce the number of ICs down to two or three," he suggested. "What are we going to do about the other 390 components?"
Outside of the RF ASIC-essentially the tuner and synthesizer, upconverter and downconverter elements-the cell phone will also depend on antenna drivers and receivers. These are the transmitter power amplifier (PA) and low-noise amplifier (LNA) receiver components, currently fabricated in GaAs (though some manufacturers-most particularly IBM-are promoting SiGe). While GaAs FETs are inexpensive and highly efficient at GHz frequencies and cellular-handset power levels, they require a split-voltage power supply (i.e., a negative bias voltage)-a nuisance to cell-phone manufacturers. More significantly, GaAs offers a limited ability to integrate. Thus, cell phone makers looking for integrated RF ICs will still not be able to find PAs and LNAs and other IC components.
And, with cell phones, the current trend is to replace multicell NiMH rechargeable batteries with single-cell lithium-ion, whose voltage can vary from 4.2 to 2.5 V as it discharges. That means the voltage regulator must step down the battery voltage at the beginning of the cell phone's usage cycle to provide, say, 3.3 V for baseband processors and system logic, but then step up the voltage as the battery nears the end of its discharge cycle. Switching regulators (dc/dc converters) with "buck-boost" capabilities are increasingly used here, along with a number of low-dropout linear regulators.
The decreasing voltages for logic and baseband processors, combined with fixed voltages for display drivers and RF transceivers, will likely increase the proliferation of voltage regulator components in a cell phone rather than consolidate them. The situation does not bode well for a single-chip phone anytime soon.