September 24, 1998

Satellite-based telephones punch past reality's impediments

Bill Schweber, Technical Editor

Users of cellular phones and pagers may think they have the ability to talk, receive messages, or send data from just about anywhere, but that's not the case. There are large areas on every continent that don't have a cellular base station within range, due to the area's relative isolation or to local signal conditions. And, you can easily be out of range if you're on a boat.

But there's an alternative that's appearing on the horizon for those who like to roam. Direct-to-satellite systems will let you use cellularlike handheld phones to link up from approximately 98% of the Earth's surface. Half-a-dozen competing multicompany consortia are implementing these systems, using different satellite configurations and system architectures and providing total or partial worldwide coverage. Each of the proposed systems is incompatible with the others, and each phone (handset) will work only with its designated system.

Unlike systems that started small and eventually grew much larger—such as the standard telephone system and even the Internet—global phone systems are designed to be large and complete from day one of their existence. They are both a challenge and an opportunity for the system architects. Because the designers are starting with a proverbial "blank sheet of paper," they are less burdened with system scalability and backward compatibility. But, the system is hard to test piecemeal and thus is viable only when it is fully operational.

Will that be LEO, MEO, or GEO?

At first glance, you may say that worldwide access is no big deal: Just put up a few geosynchronous-earth-orbit (GEO) satellites at 35,800-km altitude, and the rest is straightforward. The Thuraya system (which links the Arab states, Central Asia, India, and East Europe) and the Asian Cellular System (which links the Asia Pacific region) have adopted this approach. Because these satellites illuminate such large footprints, you can cover the Earth, except for the polar extremes, with just three of them.

High orbits bring some extreme challenges. The first challenge is their high latency of about 260 msec (round trip) due to the path propagation delay—a condition both irritating for voice conversations and inefficient for data protocols. The second challenge is the inherent path loss of about 183 dB at 1 GHz, which is caused by the distance. (Remember, free-space loss increases with the square of the distance.) The high path loss thus requires higher power or higher effective antenna gain at both the handset and satellite nodes of the link to maintain an adequate link margin.

One alternative to GEO is to use low-earth-orbit (LEO) satellites, which circle at 500 to 1500 km above the Earth. Both the 780-km Iridium and 1400-km Globalstar systems use this approach. Iridium's altitude cuts latency to about 10 msec and cuts path loss to 155 dB. But these attractive numbers come at a price. Because these satellites have small footprints and orbit every few hours, Iridium requires 66 satellites plus six in-orbit spares for coverage. Globalstar, a higher altitude LEO system, needs 48 satellites for coverage.

The large number of LEO satellites and their fast movement relative to the user makes system management complex. A single call connects through a shifting array of satellites, and the network management and routing is complex as calls are handed off through moving satellites.

As a compromise between the comparative virtues and vices of LEO and GEO designs, two of the global groups are using medium-earth-orbit (MEO) configurations with orbits at 5000 to 12,000 km above the Earth. These systems need 10 to 14 satellites for full coverage. Whether such an intermediate siting of satellites is the best choice between the extreme alternatives is a never-ending discussion between system architects and managers.

Satellite is the visible key

There are many similarities as well as sharp differences in the elements of each of the competing LEO, MEO, and GEO systems. For several reasons, let's concentrate on the Iridium system (www.iridium.com), which Motorola proposed in June 1990 and now owns 25% of its 18-member consortium. First, it's the first system that was announced with a committed financial backing of several billion dollars. Second, it's the closest, by far, to reality; Iridium was slated to begin service this month. Third, it's the best funded (translation: most expensive) and most complex system, thus it also requires a new way of thinking about satellite-based links. Iridium's large number of satellites, for example, changes the traditional "one-at-a- time" satellite build/launch approach.

By last month, Iridium had launched 72 satellites in a year and a half; at that time, 65 units were still operating, and seven units had failed for various reasons. The satellites occupy six near-polar orbit belts, at a 780-km altitude. The typical Iridium satellite measures 4.5m long, with a triangular base measuring about 1m across. This triangular shape allows you to pack the satellites together for a multiple-unit launch. Close-packing is an important factor, given the large number of satellites that the Iridium system needs and the continual need for the satellites to be replaced. This need for replacement is unavoidable because atmospheric drag, fuel consumption, and the effects of the inner Van Allen radiation belts give these LEO satellites an expected lifetime of five to eight years (see sidebar "Something else to worry about").

Although the satellites are "hand-crafted" at Motorola's Satellite Communications Group plant in Chandler, AZ, the large number needed also means that the design and assembly benefit from techniques used for moderate-volume assembly. There are many tangible and intangible benefits and implications when a quantity of units has the same design and construction—a situation that the satellite business has been unable to take advantage of until this program.

Even a completed and tested satellite still has a long way to go—both literally and figuratively—to be useful. Motorola packs each satellite in an environmentally controlled, EMC-shielded crate with an integral electronic data recorder and then ships the unit via Federal Express (Fedex). This shipment involves a lot more than just putting a label on the crate, dropping it off in the mail room, and having the local Fedex route truck pick it up! Along with the satellite itself goes support equipment, including electrical and mechanical satellite prelaunch- preparation systems, batteries and hydrazine fuel, safety and security equipment, office and medical supplies, forklifts, and small trucks. The total package weighs 160,000 lbs (72,500 kg), although the satellite itself, powered by GaAs solar cells, weighs about 1500 lbs (700 kg).

Fedex can take these packages only as far as its delivery range extends; the packages are then loaded onto trucks and driven to their final launch site. The Iridium consortium launches satellites from the Taiyuan Satellite Launch Center in China, the Baikonaur Cosmodrome in Kazakhstan, and the United States, because of limited worldwide launch capacity, the consortium's practical desire to not rely too heavily on any single site, and cost considerations. The Russian Proton rocket used at Baikonaur can simultaneously boost seven satellites, the Chinese Long March vehicle can lift two, and the US Delta II rocket handles five.

All the competing systems use channels in the 1.610- to 1.6265-GHz band for uplink and channels in the 2.4835- to 2.500-GHz band for the downlink. Most of the user handsets will support both voice and data, and internal buffer memory supports data-port connectors (a necessity given the latencies and hard-to-predict routing paths of any satellite system). The Iridium handsets resemble second-generation cellular phones and are about twice the size of the cellular-only phones now available. They will be made by Motorola as well as by other vendors, such as Kyocera; phones and pagers from various vendors will support different sets of features and options.

Most of the systems plan to offer dual- or even tri-mode phones (satellite plus either or both TDMA and CDMA) with 1/2W output power and omnidirectional antennas; these phones currently cost around $1000. When you're in proximity to a conventional cellular system base station, you'll link through that base station; when you're out-of-range, you'll link via the satellite. This dual-mode operation reduces the load on the satellite system and uses an infrastructure that's already in place. It's puts you at an advantage when you're in a large building that excessively attenuates the signal between the handset and satellite, causing signal strength to fall below the 16-dB link margin. (You can also move closer to a window to get to a reduced-attenuation situation.) Satellite ground stations link the satellite-ground-station system to the public switched telephone system and cellular-phone providers.

One of the other factors that makes Iridium complex, in addition to the large number of LEO satellites, is that Iridium allows intersatellite linkages and can connect users without the path going through a ground station. Depending on the link start and termination points, a satellite may directly connect to an adjacent one, and even hand off calls as the satellites pass in and out of range. In effect, each satellite is an orbiting equivalent version of the cellular-phone system's mobile-telephone switching office (MTSO), handling as many as 1100 full-duplex calls. Unlike ground-based MTSOs, though, the Doppler effect in this case adds significantly to the signal-capture and recovery challenge.

In contrast, MEO and GEO systems link each call through a ground-control station and associated gateways. This method simplifies the satellite's functions, but it does so at the expense of terrestrial links among these earth-based stations.

Success is assured—or not

Although it's likely that at least several of these systems will become technically active, some may fail in profitability. These complex systems are expensive to set up and operate, and the true size of the market for their capability is, at best, an educated guess. Various satellite systems are coming online, and their business viability varies (Reference 1). It remains to be seen how many users will be willing to pay the relatively high handset cost and the estimated $1.50-per-minute usage cost for worldwide coverage, especially when a viable regional alternative of conventional cellular service is available. But the integration of such advanced systems and technologies and into a seamless global system is an impressive accomplishment and a testament both to carefully planned system architecture and the potential of satellite-based systems.

References

1. Paul Gibson, "ET phones home," Electronic Business, May 1998.
2. Barry Miller, "Satellites free the mobile phone." IEEE Spectrum, March 1998.
3. Brian Taptich, "Wireless communications Star Wars," Red Herring, July 1998.

Something else to worry about

Although long-term reliability is every engineer's concern, most designers do not have to worry about the unavoidable and debilitating effects of radiation on their products. Not so for the designers of the satellites used in these global systems. The components within are bombarded with radiation from the Van Allen belts as well as solar and other space-sourced radiation. (Radiation is not the only life-ending factor; depletion of onboard steering fuel, used for the positioning and steering rocket engines, is another.)

The differing radiation factors cause hard and soft single-event upsets and gradual degradation in parametric performance. This degradation is typified by shifts in IC thresholds, increases in operating currents, and shifts in propagation delays. Because of their different altitudes, LEO, MEO, and GEO satellites endure various combinations of the various radiation sources, which are somewhat unevenly distributed in relation to altitude. In general, LEO satellites last five to eight years; MEO units, 12 years; and GEO units, 12 to 15 years.

One way for a satellite designer to deal with the effect of radiation is to use radiation-hardened (rad-hard) devices. Component vendors have developed various sophisticated process techniques to provide this performance. However, such devices are generally too expensive for the commercial world in which these global systems compete, and the availability of these devices in quantity is uneven. Also, you just can't get many of the desired and needed latest generation ICs in rad-hard designs, or the qualification period for these special devices is too long for them to be competitive.

As a result, the designers of the global communication satellites use this combination of techniques:

• Using commercial ICs that have a few extra process steps added, which adds some rad-hard characteristics at relatively low incremental cost.
• Using circuit design with extra guardbands and tolerancing, to allow for parametric drift.
• Adding redundancy, error detection and correction, and resets to accommodate hard and soft failures.
• Selectively using tungsten/copper radiation shielding, which reduces the electron (but not the proton) aspect of radiation but adds a weight penalty.
• Acknowledging that radiation will limit the operating lifetime of a satellite and, therefore, putting standby spares in orbit and planning to regularly launch new satellites.

You can reach Technical Editor Bill Schweber at 1-617-558-4484, fax 1-617-558-4470, or bill.schweber@cahners.com.

EDN Access | Feedback | Table of Contents |