A new electronics research centre has opened in Belfast's docklands with great expectations, writes Dick Ahlstrom
A new research centre in Belfast's historic harbour area is going out of its way to pack as many eggs as possible into a single basket. It wants to build unprecedented levels of functionality into a single microprocessor to produce a "system-on-a-chip".
The idea is to combine wireless technology, high-speed data handling, local processing and more, hard-wiring the lot into a single chip, says Queen's University Belfast (QUB) professor Vincent Fusco.
"It is really the philosopher's stone. If it can be done we will have done something very substantive," says the electronic engineering specialist and director of the university's new International Centre for System-on-a-chip and Advanced Microwireless (Socam).
QUB vice-chancellor Prof Peter Gregson officially opened the €7.7million centre yesterday. Funded through a public-private partnership, Socam is the first major element going into the university's Institute for Electronics Communications and Information Technology (ECIT) a purpose-built €57 million research centre opening later this year in the Northern Ireland Science Park in the harbour area.
Socam's goal sounds disarmingly simple, to engineer a silicon chip in such a way that just one compact processor will provide high speed wireless communications for phones, video streaming and intelligent sensor applications.
Getting there however requires high-level basic research in a whole range of technologies from chip production and materials science through cryptography and signal processing, explains Fusco. Each application arguably might require its own chip but Fusco and his team of 35 researchers want to hard wire the lot into just one.
"There are lots of fundamental technical challenges there in terms of basic physics and engineering. It is very research oriented, long-term research oriented," he says. "The whole thing would ultimately be combined into a single integrated chip."
Socam's name explains what the team is trying to do. It is system-on-a-chip, but also wireless communications.
Individual elements such as digital signal processing, analogue to digital and digital to analogue converters and internal data processing capacity are "hard wired" directly into the chip circuitry. Hard wiring means that processing speeds are much higher than when using conventional software driven computer architectures, Fusco explains.
The team develops mathematical algorithms that can deliver say digital signal processing and then converts the algorithms into actual circuits that can be built into the chip. Similar work is under way for the essential cryptography needed to protect wireless digital signals, Fusco says.
The system-on-a-chip must then be able to communicate with the outside world, and this represents yet another technological challenge, building radio frequency (RF) signal reception, broadcast and processing into the same chip.
RF signals do not do well in conventional silicon, Fusco explains. The RF energy tends to dissipate as heat before the information carried in the signal can be captured. The Socam team's response is to "play tricks with the material", he says, "make structural modifications to the material itself".
One approach involves producing pits on the silicon surface, known as "electromagnetic band gap structures", he says. The work also involves fundamental physics and mathematical modelling, the latter to help anticipate how various chip architectures will perform when handling RF signals.
As if all this wasn't enough, Fusco also wants the chip to be "cognitive", to have a degree of intelligence that would enable the system's RF receiving antenna to be "aware of its surroundings".
A basic assumption is that the chip will be used in a wireless environment, but broadcasting and receiving RF signals is dynamic depending on the local environment. The signal quality and strength varies depending on which way say a mobile is held and the system will have to reconfigure itself to optimise antenna performance.
Fusco sees this self-configuring antenna forming part of the chip itself and possible antenna designs are being tested in Socam's large anechoic chamber pictured here. Anechoic chambers block off sound or RF signals leaving a quiet test area where antenna performance can be measured.
In effect the Socam team is developing the next generation of digital signal processing device based on a wireless environment. It doesn't really matter what the digital signal contains, whether it is voice, music, video or data.
"It is built like a Swiss army knife," says Fusco. It has lots of functionality but at a low cost, with yet another challenge being to ensure it runs on the lowest power-level possible to ensure long battery life.
Fusco is philosophical about whether the goal of a single chip can be achieved, suggesting that one way or another the work will produce new ideas, innovations and products. "We are particularly excited by the fact that aside from providing a platform for world leading research, the Centre will also generate a substantial amount of new intellectual property which will undoubtedly produce both new products and new companies," he says.