The challenges of integrating mobile phone, broadcast TV & radio and wireless datacomms radio technologies are significant, not least because any converged device aggregates the most technically demanding performance requirements of each radio application onto a single platform.
System designers must also bear in mind that a ‘global’ standard (e.g. DVB-T) may actually be deployed with significant nuances at a national or regional level (e.g. operational frequency, adjacent channel conditions, transmit power levels etc).
Whilst delivering increased functionality - without compromised performance - these new converged platforms must also adhere to the CE mantra of low cost, low power and small size, and this can only realistically be enabled by multi-standard, multi-function front-end solutions. Any multi-standard front-end must also strive to deliver a competitive solution even when deployed in a single- or dual-standard application to deliver economies of scale for the silicon vendor and system integrator.
Functional partitioning
Although all of the cellular, broadcast and connectivity radio technologies listed above may never be simultaneously featured in a single CE device, as an example, today’s cellular handsets already typically integrate one radio from each of these three functional categories, with manufacturers striving to absorb further functions. Increasingly, notebook computers, portable media players (PMPs) and other CE devices are following this integration trend.
Since there is a practical size limit and cost consideration to the number of individual integrated circuits (ICs) that can be physically accommodated onto a single device platform, chipsets are now available that integrate radio functions within one of the three functional categories defined above.
There are fundamental reasons for integrating radios of similar user function, rather than attempting cross-functional integration. Firstly, in CE devices featuring cellular capability, device integrators typically develop multiple product generations based upon the same cellular baseband platform to leverage the huge software development effort a new cellular baseband mandates; if new radio functions peripheral to the cellular functionality were routinely absorbed into the cellular baseband, the required software re-engineering would probably lead to a ‘latency’ in the deployment of new radio technology.
Therefore, multi-radio silicon integration does not typically cross the cellular boundary. Secondly, CE device manufacturers typically develop a range of products with different functions and features to address disparate consumer, geographic and market needs; a multi-radio, multi-function - and hence potentially higher cost - device will not therefore address all market or consumer needs.
Thirdly, there are technical reasons for grouping - or not - certain radio technologies together. These reasons may encompass fundamentals such as: the use of a similar modulation scheme (e.g. several digital TV and radio technologies use OFDM modulation); antenna limitations (e.g. inability to share an antenna due to bandwidth constraints or the need for simultaneous operation); co-existence problems related to interference (e.g. a sensitive GPS receiver may be ‘deafened’ if integrated with a cellular transceiver).
Multi-standard RF front-end challengesA multi-standard RF front-end may be a standalone IC or part of a larger system-on-chip (SoC) solution integrating RF and modem capability. In both cases, the key design requirements for the RF circuitry are principally the same, namely:
- Supporting a large frequency band of operation and wide range of channel bandwidths
- Supporting a high input signal dynamic range (low noise, high linearity)
- Co-existence with other RF functions and immunity to ‘real world’ interferers
- Accommodate antenna limitations i.e. size, antenna sharing, antenna isolation etc
- Low power to maximise battery life
- Enable low system cost
- Deliver a small solution size
Of course these considerations are typical to the system integration of any RF front-end, but the challenges become even more acute when co-location of multiple radios is attempted due to the aggregation of disparate requirements.
Even if the market requirement for an all encompassing product existed, a single front-end could not realistically enable support for the listed radio standards; the vast operational frequency range, various channel bandwidths and differing dynamic range requirements are simply beyond cost-effective implementation today.
A Multi-standard front-end example
The design of a multi-standard front-end that ‘only’ delivers broadcast TV and radio reception in a portable device covering FM, DAB and DVB-T (for EU market) is challenging in its own right. The receiver’s synthesizer must cover one decade in frequency from 88MHz (FM) to over 800MHz (DVB-T), support frequency step sizes down to single digit kilohertz (FM), deliver low integrated phase noise to support 64 QAM modulation (DVB-T), and also feature low single-side band noise to mitigate blocking effects.
Additionally, the receiver’s RF input stage must support a large dynamic range enabling reception of signal levels from as low as -105dBm (FM) to as high as 0dBm (DAB) often in the presence of interferers, whilst the baseband section must support channel bandwidths from 200KHz (FM) to 8MHz (DVB-T).
One method to address these distinct requirements would be to design multiple RF chains interfacing into multiple corresponding baseband circuits. However, such an approach would result in a die size so large as to be commercially uncompetitive if only one or two application standards were to be supported. A better approach is to employ a reconfigurable receiver architecture.
Robust performance in the presence of unwanted interferers is vital to ensuring a good user experience in the ‘real world’. In the case of this broadcast receiver example, typical sources of on-channel interference include high power FM signals, whose second harmonics may fall onto the wanted DAB signal, or harmonics of DAB signals falling onto the wanted DVB-T channel.
Additionally, harmonic mixing products of the wanted signal and oscillator may also fall in-band. Other interference scenarios which result in blocking of the wanted signal include the presence of strong adjacent channels, or reciprocal mixing of wideband interfering signals due to inadequate phase noise. Careful choice of down-conversion architecture (heterodyne versus homodyne), use of harmonic rejection mixing, and on-chip filtering can help to alleviate these problems.
Antenna integration must be considered in any multi-standard system. In the case of broadcast reception, di-plexing, tri-plexing and even quad-plexing techniques can be used to reduce the physical number of antennas that must be integrated into a CE device. However, wideband antenna frequency response and multiplexing insertion loss must also be considered.
There are also significant power and size trade-offs between optimal partitioning of signal processing between the RF front-end and the digital demodulator. Analogue filtering is a good example of this point: if an RF front-end provides too much analogue channel selectivity, the die size is negatively impacted; conversely if insufficient analogue selectivity is delivered, the demodulator’s analogue-to-digital converter faces unreasonable dynamic range requirements. Such carefully considered system trade-offs can deliver significant technical and commercial benefits.
Chet Babla is director of marketing at Mirics Semiconductor