|Home | Articles | Forum | Glossary | Books|
Fastest micro wave transistor
In conjunction with the University of Illinois, the American General Electric Company has developed what it claims is the fastest ever microwave transistor. It's a mod-fet (modulation doped f.e.t.) with a maximum cut-off frequency of 230GHz-a shade faster than a similar device produced recently by MIT Lincoln Laboratory, the details of which were released earlier this year*.
The increased speed is a function both of the semiconductor material itself and also of the fabrication technology. Electrons travel faster in indium gallium arsenide than in both gallium arsenide and silicon, but unfortunately the former material is not readily available. Researchers at Illinois therefore used molecular beam epitaxy to create 40 atomic layers of indium gallium arsenide on a gallium arsenide substrate. As for device geometry, GE managed to fabricate with a gate width of one quarter of á micron (2.5 x 10^-7m), significantly reducing electron transit time.
A noise figure of around 2.5dB at 60GHz and an efficiency of 25 to 30% has been achieved with the new transistor, but GE say they have no plans to market the device, and expect to refine it for incorporation into their own products.
Communications Commentary, October 1986, page 9.
Conductive windows improve r.f. immunity
Recent tests at ERA Technology have shown that thin metal coatings on glass can provide an effective electromagnetic screen to radio-frequency interference.
Glass panels coated with a thin layer of conducting material have potential application in the reduction of emissions from computing devices and in providing protection from incoming high-intensity radiation from radar transmitters.
The results of ERA's tests, carried out as part of a sponsored research program on electro-magnetic compatibility, agree reasonably well with calculated values based on accepted screening theory, and some simple expressions have been derived to allow designers to obtain a rapid assessment of screening performance from the sheet resistance value.
Two types of coated glass were tested, one having a layer of indium tin oxide deposited by a d.c. reactive sputtering process and the other having a tin oxide/ copper/tin oxide sandwich layer deposited by magnetron sputtering. Tests were made on 1m^2 panels forming the front of a copper box which covered the range 1 to 30MHz (magnetic mode) and 30 to 1000MHz (electric mode).
Comparisons between the results with the coated glass panels in place and with the glass re moved i.e. from the open box (the box is a typical computer terminal enclosure) showed that the 1m^2 enclosure on its own made little difference to the free-space attenuation in the electric mode except at 50-60MHz, due to multipath interference within the cabinet. However, because of such existing multipath loss, the additional benefits of adding the coated glass panel were negligible at these frequencies.
Elsewhere in the spectrum the glass samples gave attenuation values of 20-30dB. (Separate tests using ridged-guide horn aerials showed that these results held good up to 10 GHz at least.) These values agree reasonably well with theory and may be calculated according to S(dB) = 20log10 (5+-) where S is screening effectiveness and R5 the sheet resistance in ohms per square. Results of tests in the magnetic mode showed that the attenuation of coated glass is minimal. ERA Technology say therefore that the application of such glass to v.d.u. enclosures is only of value at higher frequencies. It would not, they add, be much use in protecting line-scan radiation, when a mesh would be more appropriate.
First error-free microprocessor
Ferranti Electronics has produced samples of what it claims to be the world's first microprocessor with guaranteed error-free design. Known as Viper (verifiable integrated processor for enhanced reliability) the new device is a 32-bit microprocessor designed by the Royal Signals and Radar Establishment for applications requiring high operational integrity such as aircraft autopilot systems, missile systems and nuclear power plants.
Viper's operation can be formally specified and verified using mathematical techniques to ensure that a completely predictable system can be implemented for such safety-critical applications. This 'provably' correct operation has not been possible with previous microprocessor systems. The 5000-gate logic design was simulated and implemented in silicon by Ferranti using the u.l.a. 'silicon compiler' software system.
Last November, Ferranti Electronics was one of two companies selected to manufacture the first Viper chips, now delivered to RSRE for evaluation.
Although the device is designed to operate in a military environment and is resistant to high radiation levels, it is expected to have many civil applications, and Ferranti will be marketing the microprocessor commercially as the VIP1.
The device is fabricated using Ferranti's latest 1.5 micron double-layer metal advanced bipolar process and will be sup plied in a Jedec standard 84-pad chip carrier.
Large-screen l.c.d. gets closer
A race is on between British and Japanese groups to produce the world's first commercially viable fast :switching liquid-crystal display. Here in Britain a collaborative program under Professor George Gray FRS of Hull University's Department of Chemistry is expected to have a demonstration display running in two year's time. Flat-screen television is however just one application of the University's new work on liquid-crystal technology.
Since liquid-crystal screens do not flicker, their use for visual displays in word processors would considerably reduce the eye-strain often associated with those machines. They could also be used to make a very fast camera shutter.
Professor Gray's team, which includes Ken Toyne, David Lacey and Mike Scrowston, is continuing research into the tilted smectic liquid crystals required.
These respond many times faster than ordinary smectic liquid crystals used for large panel displays or the nematic liquid crystals used for calculators, which are simply not fast enough for television applications.
Already the Hull team have come up with compounds that respond in less than 1000-about a hundred times faster than the response time of nematic displays. Part of the secret lies in the viscosity of the material, but the real advantage comes from the tilted arrangement of the molecules. Unlike those of a nematic system which have to rotate and point in the opposite direction, the molecules of a tiled smectic system have only to sweep out half a cone.
First all-optical regeneration
The first all-optical light re-generator for use in optical communications has been demonstrated by British Telecom Research Laboratories at Martlesham Heath. The regenerator still in the experimental stage--both amplifies and retimes light pulses directly without converting them from light to electricity, as occurs in conventional repeaters.
All-optical regenerators, when developed commercially, promise considerable savings in the cost of optical communications links which presently need re generators every 30km or more, especially for undersea systems.
Regenerators will be significantly cheaper and simpler to make, and their power requirements will be reduced.
The all-optical regenerator was developed by two British Telecom research engineers, Rod Webb and John Devlin. Its key component is a micro-laser which under certain conditions can behave like an optical logic switch. An optical clock signal is fed to the laser to hold the switch state just in the "off' condition.
When a pulse of light from the incoming fiber arrives at the laser it has sufficient energy to switch on the laser but only when the optical clock signal is present. This triggers the laser to generate a more powerful burst of light in synchronism with the clock which is then injected into the outgoing fiber.
The BT all-optical regenerator differs from previously demonstrated optical amplifiers in two important respects. Bistable operation leads to a signal output level that is relatively constant over a range of input levels; and secondly the signal is retimed by an optical clock.
It is based on the principle that a Fabry-Perot semiconductor laser has nonlinear transfer characteristics because its effective refractive index varies with optical power level. At some wavelengths this nonlinearity leads to bi-stability.
To form a regenerator, an optical clock waveform consisting of a train of pulses with peak power just below the bistable threshold is combined with the data stream and coupled into the amplifier. When a pulse is 'low', a slightly amplified clock pulse appears at the output, and when 'high' the additional power is sufficient to exceed the threshold and the output jumps to a higher level, which is insensitive to the data power, and reverts to low only at the end of the clock pulse. The output is the regenerated data in return-to-zero form, re-timed by the clock.
Results obtained recently from the Nuclear Structure Facility (NSF) at the Science and Engineering Research Council's Daresbury Laboratory represent the final step in a long search for atomic nuclei with a super-deformed shape. Theory predicts that under conditions of extreme stress caused by rapid rotation some nuclei will suddenly adopt a fixed, super-deformed shape.
This shape, which is similar to a rugby ball, has a 2:1 major-to-minor axis ratio and is stable under these extreme stress conditions.
Over the past 15 years, confirmation has been sought by many groups worldwide and previous reports by scientists from Daresbury and Liverpool University have already hinted at a breakthrough. In a recent series of experiments at the NSF using high-resolution spectroscopy, scientists have now, for the first time, produced a spectrum showing the sequence of discrete gamma rays emitted as a rapidly rotating super-deformed nucleus (dysprosium 152) slows down.
Almost 2% of the dysprosium 152 nuclei, produced by bombarding a palladium 108 target with beams of calcium 48, were formed in the super-deformed shape. The spectrum shows a series of 19 gamma rays, each separated by an energy of 47keV, which slow down the nucleus from an angular momentum of 60 units to 22 units. This observation of nuclear states' up to 60 units is itself a great leap forward: the previous record was 46 units and previous advances have been in steps of only a few units; in addition, 60 units is very close to maximum possible angular momentum for any nucleus before it breaks up under rotational stress.
In super deformed nuclei the emitted gamma rays, which arise from transitions between a sequence of states, are predicted to have very short lifetimes caused by the large quadrupole moments of the highly deformed shape. The lifetimes of these gamma rays have now been measured by scientists from Daresbury and Liverpool University.
These measurements established that the quadrupole moments are indeed extremely large, con firming that this nucleus is the most deformed nuclear shape found so far.
These observations open up new possibilities for studying the nucleus- nature's only strongly interacting, many-body quantum system- under novel conditions. Such studies are already being started at several European and American Laboratories as well as at Daresbury.
Thunder clash in the new Mexico
A bizarre experiment to try and create lightning involved stretching an electrified wire 2km long between two peaks in the Magdalena Mountains. A re port by researchers at the New Mexico Institute of Mining and Technology claims not only to have caused clouds to form but to have started local thunder storms. But spectacular though this may have been, it's nowhere near as significant as the renewed storm it has sparked off over the origin of cloud electrification.
The group believe their work supports a convection theory in which charges are carried into the atmosphere on rising plumes of hot air. This theory is, however, discounted by a rival group of scientists working at the same institute. Their theory, known as the particle-charge theory, suggests that charges are built up within clouds as frozen moisture particles collide with each other.
On this hypothesis, the more heavily negatively charged lighter particles rise. Charge separation is thus accomplished.
Which of the theories more nearly represents the true state of affairs remains to be seen; perhaps, as some workers have suggested, it could be a combination of both. Whatever else, these New Mexico experiments represent a spectacular and successful attempt to control the development of thunderstorms.
Radio telescope bigger than the earth
Aperture synthesis, common enough on a small scale as a means of reducing the beam-width of a transmitting or receiving antenna, has recently been applied on the grand scale by radio astronomers in Butan and the USA. Not content with using radio telescopes in different countries linked by radio, they've taken advantage of a temporarily redundant tracking and data re lay satellite (TDRS) put into orbit to keep shuttle vehicles in con tact with Houston. This TDRS, with its 5m dish, has been hooked up by radio link to the Tidbinbilla tracking station in Australia and a similar receiver at Ousuda in Japan. Together they form an antenna with an equivalent revolving power of a dish 178,000 km in diameter. In practice this means that it should be capable of resolving radio objects no more than 2 arc sec. in diameter.
The experiment, coordinated from the Jet Propulsion Laboratory in California, is extraordinary because of the extremely fine degree of control needed on the satellite. For the array to work it was necessary to measure the distance between the earth bound tracking stations and the satellite to better than 13cm--the wavelength being used for the astronomical observations.
In practice, the satellite also had to be controlled so that the 5m dish did not move by more than 1cm during the observations.
Remarkable though this is as a technical feat, it's by no means the biggest array envisaged. A project code-named Quasat is designed to make use of a 50m dish in space.
Also see: APPLICATIONS SUMMARY
(adapted from: Wireless World , Jan. 1987)