Outlook 2010: Is there a future for analogue?

Digital is cleaning up its act; the dark secret harboured by digital technology, which is well understood by the engineering community but only now starting to be appreciated by end users, is that it is inherently 'dirty'. Quantisation errors exist throughout the signal path; be that at the hardware or software level. The truth, therefore, is that any signal represented digitally, but which started life in the analogue domain, has been corrupted by conversion. The same isn't technically true for purely digital signals and so it follows that by removing analogue from the equation, the overall effect is an increase in quality, albeit through raising the noise floor. While lossless is a modern misnomer applied to the mixed signal domain, digital and analogue have coexisted throughout their respective development, maintaining a perfunctory link through fairly coarse conversion methods. Improving those conversion methods has been pivotal in the development of digital technology, as it continues to rely on superior analogue sources in order to sustain the degradation of a digital signal path. However, as digital technology expands its reach, new developments imply that a closer coupling between the two domains may be on the horizon, promoting a new impetus for mixed signal advancements. The elegance of a constantly changing signal may deliver a seamless analogue experience, but it does nothing for the signal's storage, repeatability, reproducibility or alterability. What you see is what you get and if you want to carry on seeing it, you will need some form of signal processing solution. Today, that will inevitably be carried out in the digital domain. Bridging the two domains requires a converter technology that meets the application's requirements; historically, this has been either through high resolution or through a high number of samples. But neither of these options meets the emerging demands for greater resolution and more frequently updated results. If the digital domain is to evolve beyond its limitations in terms of quality, it needs to get closer to the source at both ends of the signal path. In other words, it needs to remove the limitation imposed by converter technology. Speed or samples? One overriding difference between analogue and digital signals is their relationship with time; analogue is closely associated with time, or frequency, such that if changed, it has a direct impact on what the signal represents. In the digital domain, frequency is essentially associated with how quickly a task is carried out. Accommodating the two is difficult and perhaps one of the biggest drawbacks of analogue to digital converters is that they often remove this quintessentially analogue artefact as soon as they can, in the form of a sample and hold circuit at the input. This introduces many compromises, both for the analogue and digital signals that respectively enter and leave the device. With good reason, therefore, one of the most recent developments in converter technology – which is likely to see significant further development in the near future – removes this limitation. Continuous time sigma-delta converters (CTSD) are now appearing from leading vendors. This approach challenges the limitations of existing converter technologies and promises a future of faster conversion at higher data rates. The concept behind CTSD isn't new, but the industry has only recently revisited it in order to meet growing demand, specifically from the 4G/LTE wireless infrastructure, for a conversion solution that can keep up with requirements. The benefits, however, are likely to be felt throughout the engineering industry, as the same underlying requirements are present in other sectors such as medical imaging and industrial control, not to mention consumer devices. Perhaps most importantly, CTSD represents a low power alternative to existing technologies, not least because it requires fewer external analogue accompaniments. For instance, CTSD converters typically don't require an antialiasing filter at the input. Similarly, because the input isn't sampled, there is no need for a sample and hold circuit on the input – such as a switched capacitor – and so the input signal is presented with a 'friendly' constant resistive input. This, in turn, removes the need for a high performance, high cost low noise amplifier at the input stage. Equally importantly, the technology is applicable to a cmos process and so is expected to scale with geometry developments much more readily than other conversion topologies. The technology is now being exploited by leading vendors in the analogue segment, significantly moving the concept of mixed signal out of an era of compromise and into the realms of compliance. D is for… Another maligned technology that could be seeing a revival, thanks to digital dominance, is the Class D amplifier. While some suppliers seem to have abandoned the technology, others are continuing to embrace it and the opportunity it represents. As with CTSD, Class D amplifier technology exhibits a 'digital friendly' aura. When developments in Class D amplifiers first started, vendors cited low power, low cost and low component count as its main benefits. While that remains true, its not so low complexity has driven some to despair. However, for those who have persevered, market conditions could be converging to create even greater opportunity. Again, the underlying process technology is also 'digitally friendly'. One such example is STMicroelectronic's sixth generation BCD6S (bipolar/cmos/dmos) process, which was used to create its TDA7492 analogue input Class D amplifier, which delivers 50W+50W stereo amplification. Perhaps more indicative of the potential this topology offers in an increasingly digital world is Analog Devices' recently announced digital input Class D audio amplifier; the SSM2517, which combines a filterless mono Class D amplifier integrated with an audio d/a converter featuring a sigma delta modulator. It uses a pulse density modulation input which it claims is the same digital interface supported by digital microphones in many handset and laptop applications. There is little doubt that the electronics industry anticipates greater demand for increased quality in the digital experience, which extends beyond the consumer market. The quality in audio and video is in the detail. Existing methodologies are striving not only to deliver that detail, but also to accommodate increased detail. High definition A/V will not be limited to the living room; once established, we will expect – and ultimately demand – HD in all walks of life. A relevant example is in medical imaging, which would immediately benefit from improvement for apparent reasons. When presented with a question posed as 'either' and 'or', the respondent will invariably say 'both'. When it comes to analogue conversion – in both directions – the question has always been 'do you want quality or quantity?'. In the pursuance of an improved A/V experience, the answer in future needs to be 'both'. The challenge facing manufacturers is incorporating 'both' into future developments. Arguably, CTSD converters and Class D amplifiers could represent a seismic change in design methodology; a disruptive shift in their approach to system design. Their use could lead to simpler system level design, as it brings many of the benefits inherent in digital technology. But as this new strain of mixed signal devices gains momentum, could it also lead to a decline in the elegant art of analogue design?