Cmos nano-wire communication (TCS.5791)
Project nummer:
tcs5791
Omschrijving van het onderzoek
Summary
Since the seventies, CMOS IC technology features continuously down shrinking transistor dimensions, bringing us the fast complex chips required for a wide range of communication and multi-media applications. While transistors still become faster when CMOS Technology scales down into the nanometer region, wires DO NOT, maily because of the drastic increase in the resistance of the thin nano-wires. Hence, on-chip data communication is becoming a key speed and power consumption bottleneck in new generations of CMOS ICs, even in optimistic scenarios. Recently, copper wires were introduced to replace Aluminium wires, despite of the large risks and research efforts associated wirth such drastic changes in IC technology. Nevertheless, as copper only has 30% less resistivity, this shifts the speed problem only a few years to the future. While others are searching solutions in the technological field, we propose to introduce mixed analog/digital data communication techniques by using specially designed modems and data lines. In this way the data transmission capacity of non-chip nano-wires cab be increased significantly (more bits per second) at the cost of additional transistors used for high performance line drivers/receivers and modems. As transistors consume ever smaller chip area compared to wires, this apporoach fits well to the CMOS IC technology trend. On the other hand, current modems take far too much area and power to be practical even in downscaled future technologies. Hence innovative solutions, taking into account the embedded IC design boundary conditions, are needed.
Summary of Utilisation Plan
CMOS IC's are ubiquitous in applications requiring large yet cheap computational power, for instance in personal computers (PC's), mobile phones and various up-coming multi-media applications. In PC's, the huge intrinsic computational power of Silicon is exploited to obtain flexibly programmqable high-speed microprocessors. However, pushing speed and flexibility to the limit comes at the cost of high power consumption and renders excessive heat production. Hence cooling fans are required and portable battery-based operation becomes impractical. Fortunately, smart application specific IC architectures cabn typically bring 2 to 3 orders of magnitude power-savings, which is very much wanted for the fast growing market of portable consumer electronics. In these low-power architectures, power dissipation is already now often dominated by data communication. the nano-wire problems will aggrevate this, and innovative approaches to the on-chip data communication bottleneck are very much wanted.
Unfortunately, European industry is lagging behind compared to the USA in the field of fast digital design (e.g. almost all high speed microprocessors are developed by USA companies (INTEL, AMD, MOTOROLA, IBM). On the other hand , efforts are being made to obtain a strong position in the market of portable consumer electronics. European industry in general and Philips in particular
Has a strong position in telecommunication and in mixed analog-digital design. The project build on this expertise and explores the possibilities to cope with on-chip data communication bottlenecks using mixed signal communication techniques. This can contribute to maintain and improve the position of Dutch industry in the semiconductor market, which has become the largest industry worldwide (even larger than the automobile industry). This project mainly develops expertise in transmission line design, line driver/receiver electronics and in mixed-signal modem techniques. Apart from on-chip communication, this expertise is useful in may other applications where large data capacity is wanted, while only few wires and low energy budget are available. Thus the application potential of techniques developed in this project is very high, especially as we use mainstream standard CMOS IC-technologies that are directly industrially available.
Resultaten van het onderzoek
Er zijn nog geen resultaten bekend.
Gebruikers
Er zijn 5 bedrijven bij dit project betrokken.
Projectleider
| Dr.ir. E.A.M. Klumperink |
Universiteit Twente
Elektrotechniek Wiskunde en Informatica
MESA+ |
Postbus 217
7500 AE Enschede
|
Status van het project
| Gestart | : 15-12-2002 |
| Einddatum | : 01-03-2007 |
Trefwoorden
ASIC, CMOS, Communicatiesystemen, Demodulation, IC-technologie, Interconnect, Mixed-signal IC, Modulation, Nano, Transmission lines.