ADCs tuned to digital radio

A breakthrough discovery by DSTO scientists will have a significant impact on future technology in many areas of sensing and communications (including mobile telephones), bringing the potential for lower power, higher fidelity, increased user density and dramatically improved wideband RF surveillance.

This innovation paves the way for high-performance digital radio receivers, sounding a final knell for the coils, transistors and previously unmatched performance of their expensive and physically complex analog counterparts.

High performance radio communications have until recently relied on narrowband radio transmitters and receivers, each constructed and tuned to its specific signal transmission or reception task.

These analog systems do their job extremely well, with high signal-to-noise ratio (equals an ability to work well in the presence of interference), even at weak signal strengths.

However, traditional analog radios have limitations in some situations, such as when the received signals are extremely weak and there is a high noise component (eg, HF communication systems); when the signals are from multitudinous sources (as in mobile telephone networks); or when many different types of data are to be extracted from complex signals for analysis by multiple users.

Technologists have responded to these limitations in conventional communications systems by moving to 'software-defined' radios. These radios adapt to the environment and maximise the information rate that can be obtained at any given time.

To implement these 'software-defined' radios, the speed and power of computer algorithms are needed.

But computers live in the 'pseudo' world of binary or digital information (two states only, 'on' or 'off') while radio signals are analog (or variable continuous).

Enter the ADC, or analog-to-digital converter to change the continuous radio signal to a binary or computer representation.

Major limitation

Analog-to-digital converters have been around for many years. However, all have a major limitation - during the conversion process they generate large numbers of relatively low-level spurious signals.

These signals create an artificial 'noise floor' that swamps actual low-level signals, which are then not seen by the receiver.

The 'spurious-free dynamic range' (SFDR) is the usable range of an ADC device before these spurious signals distort or interfere with the signals of interest.

The SFDR has therefore become one of the most important specifications used to qualify the performance of an ADC.

The key to improving the conversion process came with the realisation by DSTO scientists that these spurious signals can be separated from real signals when an array of sensors is used, but cannot be separated where only one sensor is used.

"Conventional ADC technology can perform much better than anyone ever knew," says Dr Warren Marwood, leader of the DSTO scientific team researching this area.

"Most research has been into ways of improving the performance of individual ADCs. We have discovered that by using a number of ADCs in an array, and with proper processing of the signal, we can eliminate almost all the spurious signals and increase the spurious-free dynamic range by some 40 dB."

Researcher Angus Massie explains "Since most of the spurious signals don't actually come from any direction in real space, we refer to them as 'ghosts'.

"If we use an array of ADCs, rather than a single ADC, these 'ghosts' are suppressed with conventional beamforming techniques that cause the array to 'see' signals from particular directions only.

"For spurious signals that do correspond to real directions, more advanced signal separation techniques are used to suppress the unwanted signals." The SFDR of a typical 100 MHz analog-to-digital converter is currently about 84 dB. However, applying these techniques to an array of about 32 antennas would enable an SFDR of up to 120 dB to be obtained.

As long-term improvements in ADC performance have averaged only 10 dB every eight years, this is the equivalent of a 30 year improvement in ADC capability in a single bound.

Setting a de-facto standard

The use of ADC-array technology will bring benefits across a wide spectrum of surveillance, communications and instrumentation technologies.

Virtually all future wireless communication systems can be expected to use multi-channel techniques. This evolution has been for a variety of technical reasons unrelated to the performance of the ADC.

If the DSTO technology is applied to this new multi-channel infrastructure, the ADC performance can be significantly enhanced, resulting in a substantial improvement in overall system performance. Superior receivers for multi-access systems are one main area of application for ADC arrays, but the concept can also be adapted to obtain higher performance in many other areas, such as scientific instruments and radars.

Almost everything with a computer in its system has a requirement for improved analog-to-digital conversion.

"ADC arrays are of immediate value in defence surveillance applications," says Massie.

"They are also going to be very important in mobile communications but will probably not come on stream until the mobile phone networks begin reinvesting in their infrastructures with the deployment of 4G PCS technology between 2006 and 2012.

"Techniques such as the exorcising of 'ghost' signals will assist designers in making trade-offs between a reduction in the number of base stations, an increase in the data rates to users or an increase in the battery life of mobiles."

Marwood adds "This is an excellent idea that is being proved through the current Capability Technology Demonstrator funding with Ebor Computing. DSTO has patents pending in Australia, the USA and Europe for the fundamental technique.

The ubiquitous use of such devices in the future means that DSTO's research has the potential to become the de-facto standard for high performance digital radio receivers."

Reprinted from 'Australian Defence Science'.