The one constant in mobile-phone design is change. Recently, this fact was underscored with the introduction of an assortment of creative new phones designed and envisioned to be in vogue around the world.
These novel phone designs are changing the way designers tackle some of the most difficult challenges: How do you provide faster digital and analogue processing speeds? How do you deal with all sorts of convergence of services technologies and applications? How do you keep power consumption low? Should you integrate digital signal processing (DSP) and radio-frequency technology on the same design?
Dedicated versus shared processing
A hot issue in mobile-phone design centres on how to partition chip functions to maximise horsepower without increasing power consumption. One architectural approach involves an entry-level mobile phone with CD-quality music. This design integrates the DSP and microcontroller functions on a single chip (see box, next page). The microcontroller functions have separate duties. One, integrated with the DSP, is an engine that handles communications processing, such as the actual mobile-phone voice signals. The other microcontroller's sole responsibility is handling the processing of, music, multimedia, and Internet access.
Using a dedicated applications processor separate from the communications engine reduces mobile phone manufacturer's product development costs and boosts overall processing speed. This segmentation allows each processor to be fully dedicated to its particular function and eliminates complications, such as increased integration and debug times, that often occur with designs offering combined communications and applications processors. In addition, segmentation purges the interdependencies that result when domains share processing and memory space.
Also, the horsepower from the dedicated applications processor delivers several benefits:
- Facilitates phone calls and data downloads while playing music or video games
- Simplifies applications software code development, test, and verification for mobile phone manufacturers because it doesn't impact the core communications engine
- Reduces power consumption because the more power-hungry applications processor can be powered down most of the time a phone is on
- Liberates mobile-phone manufacturers to focus on developing applications and man-machine interfaces without worrying about impacting the communications engine
Widespread debate has heated up in recent months about what device will ultimately be the centre of the most mobile-phone convergence. Devices being considered include mobile phones, PDAs, Blackberries, wireless LANs, and music players.
One question is whether the music-player handheld device will become a mobile phone, or will the mobile phone become a music handheld device? The truth is, both will probably happen. From an engineering perspective, however, there should be no doubt that much architectural planning going forward will be centred on the mobile phone.
With all of the mobile-phone application convergence continuing at a frantic pace, the phone designer must be nimble and smart in deciding what to integrate and what to leave as separate chips. They will have to keep their sights set on building higher-performance, lower-cost, and lower-power-consuming devices.
Mobile phones are stealing communications and entertainment functions from PCs. Therefore, the designer would benefit by being familiar with the latest PC design trends and apps.
Designers will have to figure out how to integrate GPS navigators onto mobile phones, because that market is gathering momentum. Also rapidly picking up steam is the design of CD-quality music capability on mobile phones, tracking with the amazingly popular small and easy-to-use music devices among consumers. Mobile phones can already store hundreds of songs.
For consumers to bite on these phones, though, high-quality digital acoustics hardware and software must be developed to make CD-quality music. In addition to CD-quality music, consumers need to be able to find a specific song quickly and easily.
The amount of power consumed by a mobile phone lies at the heart of the architectural design-tradeoff issue. This trend will continue as new, more creative designs come to market.
Mobile phones must have extended battery life. Simultaneously, they require faster possessing speed to handle higher-bandwidth multimedia services, such as mobile TV and music.
The way to address this problem is no longer focused on the silicon chip at the circuit level. What's crucial now is addressing this problem from a system-level architectural platform point of view. That's because all of the mobile phone's hardware and software aspects must be factored into the design analysis to get more accurate, ideal solutions. Such a system-level approach opens up more possibilities for innovations in design to keep power low and bandwidth high.
The power-consumption issues relate to what seems like a surprising emerging trend: reduce the features on a mobile phone. More phones are being customdesigned and segmented to play high-quality music, do high-speed Internet browsing, or take digital camera photos—but not necessarily do all three on one mobile phone. This way, the mobile phone consumes less power.
Integrating baseband and RF on the same chip?
For at least the past 10 years, the semiconductor industry has been conceptualising and working to deliver a mobile-phone architectural nirvana: combining RF and DSP silicon into one system-on-a-chip (SoC). Such integration has already begun in the ultra-low-cost mobile-phone segment, owing to the SoC benefits of fewer components, lower costs, and lower power consumption.
Achieving this SoC for converged phones is probably too much to ask in the next few years, given the design complexity of new phones. No single-chip multimedia phone is likely to materialise on the market before 2008.
Other factors will push out this integration. Lack of industry-wide knowledge in baseband and RF-technology design will continue to slow integration progress.
Company relationships represent another factor. A company that has mobile-phone basebands but not RF may consider buying silicon from RF chip providers. So far, though, that doesn't look to be financially feasible as RF prices are generally too high.
If those prices decreased, the baseband provider would be more likely to collaborate with the RF provider. As a result, progress toward bringing to fruition the single-chip wonder would ultimately accelerate.