When Texas Instruments last week laid down yet another gauntlet to the industry with the announcement of its intention to realize an Edge-capable, single-chip multimedia phone using its Digital RF Processor (DRP) all-CMOS technology, I got flashbacks to September 2002. It was then that TI boldly promised to realize the first single-chip GSM phone by sometime in 2004.
Skeptics had a field day. This wasn't a Bluetooth radio they were talking about here. This was a relatively complex GMSK modulated waveform with extreme sensitivity requirements on the receive side. CMOS would just not cut it, they proclaimed. But TI ignored the nay sayers.
Fast forward two years and the team managed to meet their deadline, albeit with just weeks to spare. The full story of that achievement was dealt with by EE Times in the feature Seven year odyssey nets single chip phone. Except for an external power amplifier and some off-board power management functions it was pretty much all there. That chip is now emerging onto mass market in the coming weeks as LoCosto.
That was then. This is now, and TI has yet again set itself up for potential disaster. The integration levels of the new OMAP V1035 have been well covered and are humbling indeed. But of particular interest is the method by which it plans to attack the thorny problem of handling Edge's 8-PSK modulation and the PA linearity issues that go along with such a dynamic, complex modulation technique. It's much more involved than the plain old GMSK that's used for GSM. To understand why, see Polar modulation ups efficiency in mobile PA designs. Many companies have tried many ways to improve the efficiency and reduce the linearity requirements of 8-PSK-modulated front ends using various polar modulation approaches as well as predistortion. Polar specialist Tropian Inc. was but one example. But few have had much success in all CMOS.
All this came to mind as I sat down with Bill Krenik, Advanced Architectures Manager for TI's Wireless Terminals Business Unit to discuss the architecture and how they planned to overcome the inherent obstacles. He was closed-mouthed as always and we danced a merry dance. But to no avail. He relied on the usual refrain that it is the 'digital tuning' ability that is at the heart of DRP that will allow them to achieve a single-chip 8-PSK Edge radio. Fair enough, I guess we'll just have to wait and see.
That said, he did make a good point when he added that the skeptics were negative when they announced their Bluetooth radio and then their LoCosto intentions and "we came through on those," he said. "And we'll come through again." Krenik expects to sample in early 2007 and be in production in 2008.
Having had fair insight into what the team did to achieve its first single-chip phone, I wouldn't be willing to bet against them. But they have a lot of work ahead of them. What do you think? Email me your thoughts at email@example.com.