Communications satellites

Communications is the oldest and still the largest commercial use of space. The architecture has changed twice in sixty years: from wide-beam to spot-beam GEO, and then from GEO to LEO constellations. Both transitions were driven by the same underlying quantity.

Capacity is about reuse, not bandwidth

The naive view is that a satellite’s capacity is set by how much spectrum it has. That is wrong, and understanding why explains the whole industry.

A satellite with a single beam covering a continent must share its allocated bandwidth among every user in that footprint. Give it 500 MHz and every user in Europe divides 500 MHz between them.

Now replace the single wide beam with a hundred narrow spot beams. Beams that are geographically far apart can reuse the same frequencies without interfering, exactly as terrestrial mobile networks reuse channels between distant cells. Total capacity becomes

where $N_{\text{colours}}$ is the reuse pattern size, typically four, formed by two frequency sub-bands crossed with two orthogonal polarisations.

A hundred beams in a four-colour scheme reuse the spectrum twenty-five times over. The satellite’s capacity has increased twenty-five-fold with no additional spectrum at all. This is the single idea behind high-throughput satellites.

The cost of narrow beams is that each one covers less ground. A 0.5° beam from GEO illuminates a spot roughly 400 km across, so hundreds of beams are needed to cover a continent, each requiring its own feed and amplifier chain. The payload becomes large, heavy and power-hungry, which is why HTS satellites are among the most complex spacecraft built.

A second consequence of narrow beams is that capacity is now placed geographically. A beam over open ocean earns nothing while a beam over a city saturates. Early HTS satellites had fixed beam patterns baked in at manufacture, so a satellite designed for one demand map became poorly matched as demand moved. This drove the shift to software-defined payloads with digital channelisers that can reallocate bandwidth and power between beams in orbit, and to steerable beams that can be repointed at new markets.

GEO architecture

A conventional GEO communications satellite is a bent pipe: it receives on one frequency band, amplifies, shifts frequency, and retransmits on another. It performs no processing on the content.

Bandwidth is sold in transponder units, historically 36 MHz, which became the industry’s trading unit. Uplink and downlink use separate bands so the satellite does not hear its own transmission, and the gap between them is why band pairs such as 14 GHz up and 12 GHz down exist.

Regenerative payloads instead demodulate to baseband, route packets and remodulate. This allows onboard switching between beams, error correction on each hop separately, and efficient multicast, but it commits the satellite to a specific waveform for its fifteen-year life, which is a long time in communications standards. The industry has been slow to adopt it for exactly this reason, and the compromise of digital channelisation without full regeneration is now the common middle ground.

LEO broadband

The LEO architecture inverts almost every GEO assumption.

The user terminal is the technology that made LEO broadband commercially possible, and it is often underappreciated relative to the satellites. A LEO satellite crosses the sky in minutes, so the terminal must track it continuously and hand over to the next one without interrupting the session. A mechanically steered dish that can do this reliably for years at consumer price points does not exist. Electronically steered phased arrays do, and their cost fell far enough during the 2010s to make the architecture viable. The satellites were always the easier half.

The ground segment is the LEO architecture’s hardest structural problem. A satellite can only relay traffic if it simultaneously sees a user and a gateway connected to the terrestrial internet. Over an ocean or a sparsely populated region there is no gateway in view, and service stops.

Two solutions exist. Build many gateways, which requires landing rights and infrastructure in every region served and is impossible mid-ocean. Or fit inter-satellite links so traffic hops across the constellation to a satellite that can see a gateway. Optical inter-satellite links solve the problem completely and turn the constellation into a mesh network in orbit, at the cost of a pointing and acquisition problem between platforms moving at kilometres per second relative to each other. See Distributed satellite systems.

Direct-to-device

The newest architecture attempts to connect unmodified mobile phones directly to satellites, with no special terminal at all.

The obstacle is link budget arithmetic, and it is severe. A phone has an antenna gain near 0 dBi, transmits a fraction of a watt, and was designed to reach a cell tower a few kilometres away. Reaching a satellite 550 km away needs roughly a hundred thousand times more link budget in the uplink direction, and the phone cannot supply it.

Since the phone cannot change, everything must be done on the satellite:

Very large antennas

AST SpaceMobile flies phased arrays of tens of square metres, the largest commercial communications arrays ever deployed, purely to compensate for the phone.

Very narrow beams

Concentrating power into small cells raises flux density at the phone and allows spatial reuse.

Low data rates

Early services target messaging and voice rather than broadband, because the link supports kilobits and not megabits.

Terrestrial spectrum

The service operates in bands licensed to mobile network operators, which requires partnering with them and raises novel regulatory questions about satellite use of terrestrial allocations.

The regulatory dimension is as hard as the engineering. Terrestrial mobile spectrum is licensed nationally and was never coordinated for space use, so satellites transmitting in it risk interfering with terrestrial networks in neighbouring countries that never agreed to the arrangement. The frameworks are being written now, and they will shape which services are viable.

Regulation

Nothing in this field launches without a regulatory position, and the process is long enough to belong on the programme schedule alongside the hardware.

The ITU regime is first come, first served by filing date, which created a long-standing problem of paper satellites: filings made to reserve priority with no intention to launch. Milestone requirements now force a constellation to deploy a defined fraction within set periods or lose its priority, which has substantially reduced speculative filing while making the schedule pressure on genuine operators severe.

Geostationary slots are additionally coordinated under a separate planning regime that reserves capacity for every country, on the principle that the orbit is a shared resource and not a first-mover prize. See Policy and regulation.

The economics

The unit that matters is cost per bit delivered, and it has fallen by orders of magnitude, driven by frequency reuse rather than by cheaper satellites.

The strategic question the industry has not resolved is whether LEO broadband economics close at scale. The capacity is real and the latency advantage is real, but the constellation must be replaced continuously, and the subscriber base required to fund that is very large. The answer depends on launch cost and satellite manufacturing cost falling and staying low, which ties the business directly to reusable launch and to the launch market.

References

• Maral, G., Bousquet, M. and Sun, Z. Satellite Communications Systems: Systems, Techniques and Technology, 6th ed., Wiley, 2020.
• Elbert, B. R. The Satellite Communication Applications Handbook, 2nd ed., Artech House, 2004.
• Del Portillo, I., Cameron, B. G. and Crawley, E. F. “A technical comparison of three low earth orbit satellite constellation systems to provide global broadband”, Acta Astronautica, 159:123, 2019.
• ITU Radio Regulations, Articles 9, 11 and 22, and Appendices 30, 30A and 30B.
• ETSI EN 302 307, the DVB-S2 and S2X standards for satellite transmission.