L Band is a generic name in radio engineering parlance, applied to radio frequencies between 1 and 2 GHz (1000-2000 MHz); the upper end of the band is licensed for personal wireless phones called PCS. Depending on the country and licensing arrangements, the L-Band frequency range for PCS is 1.7-2.4 GHz. PCS phones are hybrids: working like cordless and cellular, but also capable of mediating computer to computer communications allowing paging, position locating, electronic banking, email, etc. etc. Some say a PCS phone call will cost no more than a fraction of a penny within a decade and phones will be so cheap they will be disposable. Maybe we can apply Moore's law here..... The cost of putting in a copper wire telephone line four or five years ago was about $4000. Now it's about $1000. Putting in an equivalent wireless line working at L Band is about $800-1000. Arraycom Inc. says that it can bring the cost to $600. Following the past pricing declines in semiconductors and electronic systems, the price of an L Band wireless line could be in the $100-200 range sometime around 2000/2002. As mentioned in my last posting, a number of projections and studies forecast about 200 million PCS phones in service by 2000. These numbers are being seriously considered by Ericsson, Nortel, Motorola, etc. Whatever the electronic protocols or signaling techniques that these phones will use, the two things they will all have in common will be antennas and L-Band RF amplifiers.
EHF also has a generic connotation. It refers to radio frequencies having a wavelength of about a half an inch or less. The frequencies around 30 GHz are typically called EHF though I have heard the term applied to frequencies between 20 and 40 GHz. What is special about these frequencies is that the licensed bandwidths are very wide: 500-3000 MHz. Because EHF is easily formed into beams by small high gain antennas, it becomes possible to set up very high speed data links over short ranges ( 1-5 miles). These links can carry hundreds of millions and even billions of bits per second. This sort of technology is ideal for connecting together the cells that make up the 800/900 MHz cellular and L Band PCS systems. As wireless phone use grows, the cells become smaller, and more intercellular links will be required. Significant growth is also expected with those EHF systems that are used to establish links between buildings in highly congested urban areas. In Europe (especially in UK and Germany) about 10,000 to 15,000 38 GHz terminals are in use. One nice attribute of EHF is that it can provide point to point data communication at rates only achievable with fiber optics, at less cost. PCMS (a CLTK customer), is a major developer of such point to point systems. Another area of similar application is called LMDS, which provides multi channel TV and bi-directional data transmission using a cellular type of organization. LMDS competes with direct to home satcom and coax cable TV. It is expected that the LMDS market will mushroom in the next few years also, with the FCC expected to grant first licenses within a year. In Canada 3 GHz of LMDS bandwidth has already been licensed. CVUS in NEwYOrkCity is running a highly interesting wireless Internet/TV system using LMDS. The major LMDS companies today are TI an HP; but lots of others will materialize as the market develops. All of these systems need EHF amplifiers.
L and EHF band amplifier technology can only be implemented using GaAs semiconductor technology. Barring any developments with high speed silicon/silicon germanium, the expectation is that GaAs will be the high speed semiconductor technology of preference and production of GaAs is expected to increase four fold in the next 3 years.
Marie, with regards to your comments on the production capacities of the various companies, it certainly would be nice to have some reliable numbers and breakdowns. It would make investing a lot easier. Many large companies produce GaAs only for their internal consumption; some like Celeritek, produce for their own consumption and sell into the merchant market. Celeritek currently has a capacity of 500 four inch wafers per week (500X4 in/wk), according to information I received from an inquiry I made several months ago. This in comparison to what you stated as 750 X 4 in /wk for ANAD and 250 X 4in/wk for TQNT. In the 1996 GaAs survey in Microwave Journal (August 1996), ANAD is listed as having a GaAs wafer production capacity 180X4in/wk, CLTK is listed at 250X4in/wk and TQNT at 100X4in/wk. In that survey Celeritek projects having a 1000X6 in/wk capacity by 1999 (seems like they are expecting quite a bit of growth). One has to take these numbers cautiously. Having a capacity to produce does not necessarily mean production at these quantities nor does it say anything about the yield. These mixed numbers seem to indicate that no one company is overwhelmingly larger than the other.....that's why I say they are on par. There is no argument however that Triquint and Anadigics currently have greater revenues.
An investment factor that should be considered in evaluating GaAs companies is what their current production is vis a vis future growth. TQNT and ANAD get about 50-70% of their GaAs revenue from 800/900 MHz RF products, RF switches, optoelectronics devices, TVRO products, etc. It is my understanding that CLTK makes about 20% of its revenue from 800/900 MHz products while the rest comes from L and EHF band products. CLTK seems to be better positioned for L and EHF band wireless growth than either TQNT or ANAD. This is reflected in their products: Anadigic's and Triquint's product lines at L Band and EHF are considerably smaller than CLTK's. When it comes to L Band and EHF amplifier technology, Celeritek is way ahead.
CLTK is a recent arrival as GaAs Semi merchant house scene. The company came out too late to be a significant player in the 800/900 MHz field. It's making the transition from a high quality military equipment company to a GaAs semi merchant supply house. The transition is still not complete: a good portion of CLTK's revenue continues to come from communications and military GaAs products. The management made a decision a couple of years ago to get heavily involved in L Band mobile satellite since this was one of the remaining market niches left to them. Though CLTK didn't see its fortunes rise (primarily because poor market timing by the satcom providers), this exercise was not totally wasted. This work gave Celeritek significant expertise, intellectual property, and plant capacity at L-Band. TQNT and ANAD have to yet to invest to the same degree in L-Band. CLTK has gone through this. The overhead is paid for.
CLTK is a company for bottom feeders. On the surface its looks mundane; it has fewer glossy folders and less of a media profile than either Anadigics or Triquint. It produces SEC filings that resonate with doom. However, their numbers look nice, and despite the hard knocks they took with mobile satellite , they still managed to have respectable growth. They also have about $4-5 per share in cash floating about. But most significantly, the company has a lot of respect in the RF design community for producing top notch L Band GaAs amplifiers having superlative performance. Its this that will carry them over the top when wireless takes on global proportions.
The information given above was taken from WWW and print published sources . It may be correct/correct me if I'm wrong. I do not work for any of the companies mentioned or for any other semiconductor/communications company. /Use the information at your own risk/ JDjzuba/ |
