Electronics Guide

Satellite communication systems carry signals between points on Earth by way of a relay in orbit. They deliver broadcast television, broadband internet, mobile voice and data, maritime and aeronautical connectivity, and precise timing to regions that terrestrial cable and cellular infrastructure reach poorly or not at all.

A satellite system has three segments. The space segment comprises the satellites themselves. The ground segment comprises gateway earth stations and user terminals. The control segment comprises the telemetry, tracking, and command facilities that keep satellites healthy and correctly positioned. Traffic flows from an earth station to the satellite on the uplink, and from the satellite back to receiving terminals on the downlink, almost always at a different frequency so that the two directions do not interfere.

This article covers the physics and engineering that shape those links: orbital selection, transponder architecture, link budgets, antenna and beam design, spectrum allocation, propagation impairments, modem and coding technology, terminal and ground station hardware, constellation design, and the regulatory framework that governs access to orbit and spectrum.

Satellite Orbits: GEO, MEO, and LEO

The choice of orbit determines coverage geometry, latency, link budget, terminal complexity, and system economics more than any other single decision. Three orbital regimes dominate commercial and government communications.

Geostationary Orbit (GEO)

A geostationary satellite orbits 35,786 kilometers above the equator in a circular, zero-inclination orbit. Its period equals one sidereal day, about 23 hours, 56 minutes, and 4 seconds, so the satellite appears fixed against the sky and a ground antenna can be aimed once and left alone.

The advantages follow directly from that geometry. One satellite illuminates roughly one-third of the populated surface of Earth at usable elevation angles, ground equipment needs no tracking, and the constellation is a single spacecraft rather than a fleet. The costs also follow from the geometry. The one-way path from the ground to the satellite takes about 119 milliseconds at the speed of light, so an end-to-end hop from one earth station to another is roughly 240 to 280 milliseconds depending on elevation angle, and a round trip approaches half a second. That delay is invisible to broadcast video and tolerable for most web traffic, but it degrades interactive voice, online gaming, and any protocol whose throughput depends on round-trip acknowledgment. Coverage also degrades badly above roughly 70 degrees latitude, where the satellite sits very low on the horizon and is easily blocked, and fails outright near the poles.

Typical applications include direct-to-home television, satellite radio feeds, maritime and aeronautical broadband, cellular backhaul, and enterprise networking. The geostationary ring, sometimes called the Clarke Belt after Arthur C. Clarke's 1945 proposal, is a finite resource. Orbital slots are assigned through the International Telecommunication Union, and satellites sharing a frequency band are typically spaced two degrees apart in longitude so that a properly sized earth station antenna can discriminate between them.

Medium Earth Orbit (MEO)

MEO spans altitudes between the lower Van Allen belt and the geostationary ring, conventionally 2,000 to 35,786 kilometers. The regime trades some of the coverage footprint of GEO for lower delay and some of the constellation size of LEO for simpler handover.

Navigation systems dominate the regime. GPS operates near 20,200 kilometers with a 24-slot baseline constellation, and Galileo, GLONASS, and BeiDou use comparable altitudes. At those altitudes the orbital period is roughly half a sidereal day, satellite geometry changes slowly enough for straightforward tracking, and the spread of visible satellites across the sky gives the geometric diversity that trilateration requires.

MEO is also used for broadband. The O3b system operates near 8,000 kilometers, where a single ground-to-satellite hop costs about 27 milliseconds and a full round trip through the satellite runs near 130 to 150 milliseconds, roughly a quarter of a GEO round trip. Terminals must track moving satellites and hand over between them, so MEO terminals carry two steerable apertures or an electronically steered array.

Low Earth Orbit (LEO)

LEO extends from roughly 160 to 2,000 kilometers, and most broadband constellations operate between 500 and 1,200 kilometers. Proximity is the whole advantage: free-space path loss falls by more than 35 decibels relative to GEO, propagation delay through the satellite drops to a few milliseconds, and the reduced link budget allows small, low-power user terminals. Measured service latency for consumer LEO broadband typically falls between 25 and 60 milliseconds, dominated not by propagation but by scheduling, processing, and terrestrial routing beyond the gateway.

The costs are equally direct. A satellite at 550 kilometers sees only a few hundred kilometers of ground at useful elevation angles, so global coverage demands hundreds or thousands of spacecraft. Satellites cross the sky in minutes, requiring continuous beam steering and frequent handover. Residual atmospheric drag limits orbital lifetime to a few years without propulsion, and constellation capacity must be continually replenished. Doppler shift is substantial: a satellite passing at 550 kilometers produces shifts on the order of tens of kilohertz at L-band and hundreds of kilohertz at Ku-band, which terminals must track and pre-compensate.

Starlink, Eutelsat OneWeb, and Amazon Leo are the leading broadband constellations, and Iridium and Globalstar have operated LEO voice and data services since the late 1990s.

Satellite Transponder Design

The transponder is the communications payload chain that receives an uplink signal, translates it in frequency, amplifies it, and retransmits it on the downlink. Transponder count, bandwidth, and power determine how much traffic a satellite can sell.

Bent-Pipe Transponders

A transparent, or bent-pipe, transponder performs no demodulation. The received signal is low-noise amplified, down-converted, channel filtered, amplified to power level, and radiated at the downlink frequency. Because the payload never interprets the waveform, operators can change modulation, coding, or even the entire air interface from the ground without touching the satellite. The penalty is that uplink noise is amplified along with the signal, so uplink and downlink carrier-to-noise ratios combine and the weaker link dominates end-to-end performance.

Classic commercial transponder bandwidths are 36, 54, and 72 MHz, with 24 MHz and narrower units used for some services and wideband channels of several hundred megahertz appearing in Ka-band high-throughput payloads. Radiated power per transponder ranges from a few watts on mobile-satellite payloads to roughly 100 to 200 watts on direct-broadcast satellites. The power stage uses either a traveling-wave-tube amplifier (TWTA), which remains the efficient choice at high power and high frequency, or a solid-state power amplifier (SSPA), which offers better linearity and reliability at lower power. Both are backed off from saturation when carrying multiple carriers, because a saturated amplifier generates intermodulation products that fall on adjacent channels.

Processing and Software-Defined Payloads

A regenerative transponder demodulates the uplink, decodes and corrects errors, switches or routes the resulting packets, and remodulates for the downlink. Breaking the link into two independent hops means uplink noise is not carried through, which typically buys several decibels of end-to-end margin. On-board switching also enables single-hop mesh connectivity between terminals that would otherwise need a double hop through a hub.

Digital transparent processors go further, channelizing the entire received spectrum in the digital domain and routing arbitrary sub-bands between beams. Software-defined payloads add on-board beamforming, so frequency plans, beam shapes, coverage areas, and capacity allocation can be reconfigured in orbit over a fifteen-year lifetime. That flexibility matters commercially, because demand rarely stays where it was when a satellite was ordered five years before launch.

Channelization and Filtering

An input multiplexer (IMUX) splits the received band into transponder channels ahead of the power amplifiers, and an output multiplexer (OMUX) recombines the amplified channels into a single feed. These are high-Q waveguide filter networks. They must provide steep skirts to pack channels closely, low insertion loss because every decibel after the power amplifier is lost radiated power, and controlled group delay so that wideband digital carriers are not distorted. Multipaction and passive intermodulation set hard limits on OMUX design at high power in vacuum.

Uplink and Downlink Systems

Satellite links operate at enormous path loss and small margins, so system design reduces to a disciplined accounting of every decibel gained and lost. The link budget is that accounting.

Link Budget Fundamentals

Free-space path loss in decibels is 92.45 plus twenty times the base-ten logarithm of the frequency in gigahertz plus twenty times the logarithm of the distance in kilometers. For a geostationary satellite at 35,786 kilometers, that gives roughly 196 dB for a 4 GHz downlink, 199 dB for a 6 GHz uplink, 210 dB for a 20 GHz Ka-band downlink, and 213 dB for a 30 GHz Ka-band uplink. Slant paths to satellites low on the horizon add another decibel or so.

The received carrier-to-noise ratio follows from transmit equivalent isotropically radiated power (EIRP), path loss, receive figure of merit, and the noise bandwidth. Designers then subtract margin for equipment aging, antenna pointing error, polarization mismatch, adjacent-satellite interference, intermodulation, and atmospheric losses. Clear-sky margin of 3 to 5 dB above the minimum required carrier-to-noise ratio is typical for C-band and Ku-band services; Ka-band systems either carry far more margin or, more commonly, use adaptive coding and modulation so that throughput rather than availability absorbs the fade.

Uplink Considerations

The uplink must deliver a specified power flux density at the satellite antenna. Because the satellite amplifier operates at a fixed gain, a fade on the uplink directly reduces the downlink carrier level, which is why uplink power control is used at Ku-band and above. Earth station EIRP ranges from roughly 45 to 55 dBW for a small VSAT with a 1.2-meter antenna and a few watts of transmit power, to more than 90 dBW for a gateway combining a 13-meter antenna with a kilowatt-class amplifier. Regulatory off-axis EIRP density masks limit how much power a terminal may radiate toward adjacent satellites, and these masks, rather than amplifier capability, often set the minimum usable antenna size.

Downlink Considerations

Downlink design maximizes satellite EIRP within the constraints of solar array power, thermal dissipation, and regulatory power flux density limits that protect terrestrial services sharing the band. Direct-broadcast satellites radiate 50 to 60 dBW or more per transponder so that an unattended 45-centimeter consumer dish can close the link.

Receiver quality is captured by the figure of merit G/T, the antenna gain in decibels minus the system noise temperature in decibels relative to one kelvin. System noise temperature combines the receiver's own noise, antenna spillover into warm ground, and sky brightness, which itself rises during rain. A handheld satellite phone has a negative G/T, a 0.6-meter consumer Ku-band terminal achieves roughly 13 to 17 dB/K, a 1.2-meter enterprise VSAT reaches about 21 to 23 dB/K, and a large gateway antenna exceeds 35 dB/K. Every decibel of G/T is directly interchangeable with a decibel of satellite power or a proportional increase in data rate.

Spot Beam and Shaped Beam Antennas

Satellite antenna design sets coverage, power efficiency, and the degree to which spectrum can be reused. It is the single largest lever on total system capacity.

Global and Regional Beams

Early satellites used a global horn illuminating the full 17.4-degree disc of Earth visible from geostationary orbit, giving roughly 20 dBi of gain. Regional beams narrow that to a continent. Wide beams spread limited satellite power thinly, so they demand high transmit power and large receive antennas, but they remain valuable for mobile and maritime services where a terminal may be anywhere in the footprint and uniformity matters more than efficiency.

Spot Beams

A spot beam concentrates power into a footprint typically 200 to 1,000 kilometers across, subtending a few tenths of a degree to about 1.5 degrees from geostationary orbit. Narrowing a beam from the full Earth disc to one degree raises gain by roughly 25 dB, which translates directly into smaller terminals or higher data rates. More importantly, beams that do not overlap can reuse the same frequencies.

High-throughput satellites (HTS) exploit this aggressively, deploying dozens to hundreds of spot beams in a cellular-like layout. Most use a four-color frequency plan built from two frequency sub-bands and two orthogonal polarizations, so each block of spectrum is reused once in every four beams. A payload with a hundred beams therefore reuses its spectrum roughly twenty-five times. The Hughes JUPITER 3 satellite carries more than 300 spot beams for about 500 Gbps of capacity, and each ViaSat-3 satellite is designed for more than 1 Tbps. Capacity on this scale is not limited by spectrum but by the number of beams, the power available to feed them, and the isolation achievable between beams sharing a color.

Shaped Beams

Shaped beams conform to a geographic or political boundary, concentrating power inside a service area and suppressing spillover where the operator holds no landing rights. Shaping is achieved either by a reflector with a deliberately distorted surface, by a cluster of feed horns whose amplitudes and phases are combined to synthesize the contour, or by a direct-radiating phased array. Applications include national coverage matching country borders, coastal beams for maritime service, and contours weighted toward population density.

Frequency Bands: L, S, C, X, Ku, Ka, and V

Band selection balances available bandwidth, rain sensitivity, antenna size, regulatory allocation, and equipment cost. The letter designations below follow the IEEE radar band convention; the specific satellite service allocations within them are narrower and are set by the ITU Radio Regulations.

L-Band (1 to 2 GHz)

L-band propagates almost unaffected by rain and works with small, low-gain, near-omnidirectional antennas, which makes it the natural home for mobile satellite service. Inmarsat operates around 1525 to 1559 MHz downlink and 1626.5 to 1660.5 MHz uplink, Iridium uses 1616 to 1626.5 MHz, and GNSS signals occupy adjacent allocations. The band is narrow and heavily congested, so capacity per satellite is limited and spectrum is expensive, but nothing else supports a handheld terminal as reliably.

S-Band (2 to 4 GHz)

S-band offers similarly robust propagation with somewhat more bandwidth. It carries mobile satellite services, satellite digital audio radio (2320 to 2345 MHz in the United States), and telemetry, tracking, and command links for many spacecraft. Terrestrial weather radar and wireless broadband share parts of the band, creating coordination constraints in some regions.

C-Band (4 to 8 GHz)

C-band has been the workhorse of fixed satellite service since the 1960s. Standard allocations are 3.7 to 4.2 GHz for the downlink and 5.925 to 6.425 GHz for the uplink, with extended-band variants reaching 3.4 GHz and 6.725 GHz in some regions. Rain attenuation is nearly negligible, which is why C-band still carries broadcast contribution feeds and services in tropical regions that require very high availability. The drawbacks are large antennas and shared allocation with terrestrial fixed microwave links. In the United States the FCC reallocated 3.7 to 3.98 GHz to terrestrial 5G beginning in 2020, compressing satellite downlinks into 4.0 to 4.2 GHz; comparable reallocations have followed elsewhere.

X-Band (8 to 12 GHz)

X-band satellite communication is reserved for government and military use. The military allocations sit at 7.25 to 7.75 GHz for the downlink and 7.9 to 8.4 GHz for the uplink, which straddle the nominal C-band and X-band boundary. Restricted access reduces interference and simplifies coordination, and the propagation characteristics sit conveniently between C-band robustness and Ku-band compactness. The band supports systems such as the Wideband Global SATCOM constellation and comparable allied programs.

Ku-Band (12 to 18 GHz)

Ku-band dominates direct-to-home television and enterprise VSAT networking. Fixed satellite service uses roughly 10.7 to 12.75 GHz downlink and 13.75 to 14.5 GHz uplink, while broadcast satellite service allocations vary by ITU region; North American direct-broadcast systems use 12.2 to 12.7 GHz downlink with 17.3 to 17.8 GHz feeder uplinks. Higher frequency means higher antenna gain for a given aperture, so consumer dishes shrink to 45 to 75 centimeters. Rain fade becomes material, typically requiring a few decibels of margin or adaptive coding for high availability.

Ka-Band (27 to 40 GHz)

Ka-band enabled the high-throughput satellite era by offering far more contiguous bandwidth than Ku-band. Satellite allocations commonly used commercially span roughly 17.7 to 21.2 GHz downlink and 27.5 to 30 GHz uplink; the lower half of that range technically falls in the IEEE K band but is universally marketed as Ka. The short wavelength permits small, high-gain satellite apertures and therefore many narrow spot beams. Rain fade is severe, exceeding 10 to 20 dB in heavy rain, so Ka-band systems rely on adaptive coding and modulation, uplink power control, and gateway site diversity. Consumer Ka-band broadband typically targets 99.5 to 99.9 percent availability, the latter corresponding to about nine hours of outage per year.

V-Band (40 to 75 GHz)

V-band, usually paired with Q-band (33 to 50 GHz) for feeder links, offers several gigahertz of spectrum for gateway and inter-satellite use. Attenuation from rain and atmospheric gases is severe, and the 60 GHz oxygen absorption complex makes part of the band unusable for Earth-space paths. The practical approach is to confine V-band to gateway feeder links, where large antennas, high power, and many geographically diverse sites can be justified, and to keep user links at Ku-band or Ka-band.

Rain Fade and Atmospheric Losses

Above about 10 GHz, precipitation is the dominant physical limit on satellite link availability. Predicting and mitigating it is central to system design.

Rain Attenuation Mechanisms

Raindrops both absorb and scatter microwave energy, and attenuation rises steeply with frequency because drop diameters approach the wavelength. At Ku-band, a heavy rain rate near 50 millimeters per hour produces roughly 5 to 8 dB of excess path loss; at Ka-band the same rain can produce 15 to 20 dB or more. Rain also depolarizes the signal, degrading the cross-polarization discrimination that dual-polarization frequency reuse depends on.

Recommendation ITU-R P.618 provides the standard method for predicting Earth-space rain attenuation statistics from local rainfall rate statistics, frequency, polarization, and elevation angle. Elevation angle matters greatly: a low look angle lengthens the slant path through the rain cell, so high-latitude stations viewing a geostationary satellite near the horizon suffer disproportionately. Fades develop and clear over seconds to tens of minutes, which is fast enough to require closed-loop mitigation rather than static margin.

Other Atmospheric Effects

Atmospheric gases absorb continuously, with a water vapor resonance near 22 GHz and a broad oxygen complex near 60 GHz. Clear-air gaseous absorption contributes a few tenths of a decibel at C-band and typically 0.5 to 2 dB at Ka-band, rising with humidity and falling elevation angle. Clouds and fog are negligible below Ku-band but can add one to several decibels at Ka-band and above. Dry snow and ice crystals attenuate far less than liquid water, though wet snow accumulating on a reflector or feed cover can be worse than the storm itself. Tropospheric scintillation, caused by refractive index turbulence, produces rapid amplitude fluctuations of a decibel or more at low elevation angles. Ionospheric scintillation affects the lower bands, particularly near the geomagnetic equator after sunset, and is a recognized hazard for L-band mobile and GNSS links.

Fade Mitigation Techniques

Adaptive coding and modulation (ACM) adjusts the modulation order and code rate for each terminal on a frame-by-frame basis. In clear sky a terminal may run 32APSK with a high-rate code; as the link degrades the system steps down through 16APSK, 8PSK, and QPSK with progressively stronger coding, trading throughput for robustness while keeping the link up. Because only the faded terminals step down, ACM converts what would have been a system-wide margin into usable capacity.

Uplink power control raises earth station transmit power to hold the received flux density at the satellite constant during a fade. Stations measure a satellite beacon to estimate fade depth and adjust accordingly. Ten to fifteen decibels of control range is common, limited by amplifier headroom and by regulatory flux density limits.

Site diversity exploits the limited spatial extent of intense rain cells. Gateways separated by 10 to 50 kilometers rarely experience deep fades simultaneously, so traffic can be routed through whichever site is clear. Diversity gain grows with the depth of the single-site fade: ITU-R P.618 models predict several decibels for moderate fades and more than 10 dB for the deep fades that dominate Ka-band and Q/V-band availability budgets. Large Ka-band systems accordingly build more gateways than capacity alone requires, purely to buy availability.

Satellite Modems and Coding

The satellite modem converts data into a waveform matched to a power-limited, bandwidth-limited, high-latency channel, and recovers it at the far end. Modem technology has advanced further than any other part of the system over the past two decades, and most of the gain has come from coding.

Modulation Schemes

Constant-envelope and near-constant-envelope phase modulation dominates because satellite amplifiers operate near saturation, where amplitude variation produces intermodulation. BPSK offers maximum robustness at one bit per symbol. QPSK doubles that at the same energy per bit and remains the default for difficult links. Above QPSK, systems use 8PSK and then amplitude-and-phase keying, 16APSK and 32APSK, whose ring constellations tolerate amplifier nonlinearity better than square QAM of the same order.

DVB-S2 (ETSI EN 302 307-1) defines the modulation and coding combinations used by most commercial satellite links, and DVB-S2X (EN 302 307-2) extends the set with 64APSK through 256APSK for very strong links, finer granularity between operating points, roll-off factors as low as 0.05 for tighter carrier packing, and very-low-signal-to-noise modes usable well below 0 dB for mobile and small-terminal applications. DVB-RCS2 standardizes the return channel from terminals to the hub.

Forward Error Correction

Retransmission is expensive when the round trip is half a second, so satellite links rely on forward error correction. Concatenated convolutional and Reed-Solomon coding, as used in the original DVB-S standard, has been superseded by iteratively decoded codes. Turbo codes and low-density parity-check (LDPC) codes operate within a fraction of a decibel of the Shannon limit at practical block lengths.

DVB-S2 pairs a long LDPC inner code, with code rates from 1/4 to 9/10, with a BCH outer code that removes the residual error floor. The move from convolutional and Reed-Solomon concatenation to LDPC bought roughly 2 to 3 dB, which is directly interchangeable with smaller antennas, lower transmit power, or higher throughput at the same availability.

Access Methods

Frequency division multiple access assigns each terminal a continuous carrier, which is simple and efficient for constant high-rate traffic such as trunking or broadcast, but wasteful for bursty users. Time division multiple access gives terminals timed bursts on a shared carrier, which suits bursty traffic but requires accurate network timing and burst-mode demodulators. Code division multiple access separates users by spreading code and is used where interference resistance or very low power spectral density matters.

Most VSAT networks use multi-frequency TDMA on the return link, combining frequency and time division so that the network can assign both a carrier and a set of slots to each terminal. Demand-assigned multiple access allocates that capacity dynamically in response to measured demand, which is what allows a shared transponder to serve far more terminals than its peak rate would suggest. Forward links are almost always a single wideband ACM carrier per beam, statistically multiplexed across all terminals in that beam.

VSAT Terminal Design

Very small aperture terminals made satellite connectivity affordable outside the broadcast and carrier markets. A modern VSAT is a compact outdoor antenna and radio assembly plus an indoor modem, installed in hours and serving an enterprise site, a cell tower, a ship, or a household.

Antenna Subsystem

VSAT reflectors range from 0.6 to 2.4 meters, with 1.2 meters common for enterprise Ku-band service. Offset-fed geometry dominates because it removes the feed and support struts from the aperture, improving efficiency to roughly 55 to 65 percent and reducing sidelobes, and because the steeper physical tilt of an offset dish sheds rain and snow. Regulatory off-axis emission masks demand tight pointing, so installation requires accurate azimuth, elevation, and polarization alignment; assisted-pointing tools using GNSS position and inclinometers have largely replaced manual peaking. Mobile installations replace the fixed mount with a stabilized platform or an electronically steered flat panel.

Antenna figure of merit sets what the terminal can achieve. A 1.2-meter Ku-band VSAT reaches roughly 21 to 23 dB/K, while 0.6 to 0.75-meter consumer terminals achieve about 13 to 17 dB/K. Dual-band feeds covering both Ku-band and Ka-band, and flat-panel arrays capable of tracking non-geostationary satellites, are increasingly common as operators run hybrid GEO and LEO fleets.

Outdoor Radio Unit

The outdoor unit mounts at the feed and contains the receive and transmit electronics. On the receive side, a low-noise block downconverter combines a low-noise amplifier with a mixer and local oscillator, translating the received band to a 950 to 2150 MHz L-band intermediate frequency carried indoors on a single coaxial cable. Ku-band LNBs typically achieve noise figures below 1 dB, corresponding to noise temperatures under about 75 K, and that figure dominates the terminal noise budget.

On the transmit side, a block upconverter translates the modem's L-band output to the uplink band and amplifies it. Output power ranges from 1 to 10 watts for consumer and small enterprise terminals to 40 watts or more for high-throughput professional installations. Gallium nitride devices have displaced gallium arsenide and traveling-wave tubes in this class, offering higher power density, better efficiency, and enough thermal headroom for fanless outdoor enclosures.

Indoor Unit and Acceleration

The indoor unit houses the modem, which performs modulation, coding, timing recovery, and network termination, and presents Ethernet to the customer. Modern units integrate routing, quality of service, traffic shaping, encryption, and remote management.

Latency remains the defining characteristic of a GEO link, and unmodified TCP performs poorly across it because slow start and congestion control are driven by round-trip time. Performance-enhancing proxies split the TCP connection, terminating it locally at each end and running a satellite-optimized protocol across the space link. Combined with HTTP object prefetching, DNS caching, and header and payload compression, these techniques improve perceived web performance substantially. Widespread end-to-end encryption limits what a proxy can inspect, so newer terminals rely more on transport protocols such as QUIC that tolerate long paths, and on congestion control algorithms designed for high bandwidth-delay products.

Mobile Satellite Services

Mobile satellite service (MSS) connects platforms that move: ships, aircraft, vehicles, and handheld terminals. Motion, small apertures, and unpredictable blockage make MSS the most demanding class of satellite link.

Maritime Communications

Maritime systems serve merchant shipping, offshore energy, fishing fleets, and yachts, and they underpin the Global Maritime Distress and Safety System. Inmarsat, part of Viasat since 2023, provides L-band services with FleetBroadband terminals delivering roughly 150 to 432 kbps depending on terminal class, alongside Ka-band Global Xpress and, increasingly, LEO broadband delivering tens to hundreds of megabits per second. Stabilized three-axis pedestals keep the antenna on the satellite through pitch, roll, and yaw, and vessels operating at high latitude must account for geostationary satellites sitting very low on the horizon or below it.

Aeronautical Services

Aircraft connectivity divides into passenger broadband and safety-of-life operational communications, which are certified separately and use different links. Fuselage-mounted Ku-band and Ka-band antennas, historically mechanically steered and increasingly electronically steered flat panels, serve passenger cabins; per-aircraft throughput has risen from a few megabits per second on early Ku systems to tens or hundreds of megabits per second on modern Ka-band and LEO services. Engineering constraints include aerodynamic drag and structural certification of the radome, maintaining pointing through banked turns, beam and satellite handover across oceanic routes, and coordinating transmissions across national borders and over regions with differing licensing regimes.

Land Mobile Terminals

Vehicle-mounted and transportable terminals serve military units, emergency responders, utilities, resource industries, and satellite news gathering. Auto-acquire antennas deploy and lock onto a satellite in minutes without a skilled operator. Manpack terminals trade data rate for weight, and the design problem is balancing aperture, amplifier power, battery capacity, and mass within what one person can carry. Blockage from terrain, buildings, and foliage is the dominant impairment, and it is far more disruptive than fading because it is total rather than gradual.

Personal Handheld Systems

Satellite telephones close a link with an antenna of roughly unity gain, which forces either a low orbit or a very large satellite antenna. Iridium uses 66 active satellites in six near-polar planes at about 780 kilometers, providing genuinely global coverage including both poles, with Ka-band crosslinks that relay traffic between satellites so a gateway need not be in view. Globalstar uses a bent-pipe LEO architecture requiring gateway visibility. Inmarsat and Thuraya provide handheld service from large L-band geostationary satellites, with regional rather than polar coverage. Voice runs through low-rate vocoders near 2.4 kbps; data services have grown from a few kilobits per second to several hundred kbps on modern service classes such as Iridium Certus.

Direct-to-Device and 5G Non-Terrestrial Networks

The newest branch of mobile satellite service connects ordinary, unmodified smartphones directly to satellites, without a dedicated satellite terminal. The technical premise is that a satellite carrying a very large antenna and substantial transmit power can close a link to a handset whose antenna gain is near 0 dBi.

Standardization

3GPP Release 17, frozen in 2022, introduced the first normative support for non-terrestrial networks, covering both 5G New Radio over satellite and narrowband IoT and LTE-M over satellite. The specifications address the problems that separate a satellite cell from a terrestrial one: propagation delays far beyond the terrestrial timing advance range, Doppler shift and Doppler rate from a moving satellite, cells tens to hundreds of kilometers across, and cell boundaries that sweep across the ground. Release 18 extended coverage, mobility, and band support, and Release 19 adds regenerative payload architectures that place a complete base station on the satellite rather than relaying to a ground-based one.

Architectures and Spectrum

Two approaches coexist. Standardized NTN operates in mobile satellite service spectrum with satellite-specific network functions. Supplemental coverage from space instead reuses a terrestrial mobile operator's own licensed spectrum from orbit, so any handset already supporting that band can connect without new hardware or new spectrum; the FCC established a regulatory framework for this arrangement in 2024. The Starlink direct-to-cell payload, operating with T-Mobile in the United States and with partner operators elsewhere, is the largest deployment of this approach, with several hundred capable satellites in orbit.

Capabilities and Limits

Capacity per beam is modest compared with terrestrial cells because a single satellite beam covers an area that would hold hundreds of terrestrial sites, and that capacity is shared. Services have therefore launched in stages: emergency and short messaging first, then low-rate data and machine-to-machine traffic, then voice. Direct-to-device is best understood as a coverage layer that eliminates dead zones for low-rate traffic rather than a substitute for terrestrial capacity, and it complements rather than replaces dedicated satellite terminals for high-throughput links.

Satellite Constellation Design

Constellation design determines coverage, revisit time, capacity distribution, and the cost of both deployment and replenishment. Most designs derive from a small number of well-characterized geometric patterns.

Walker Constellations

The Walker patterns describe uniform constellations with the notation i: T/P/F, where i is the inclination, T the total number of satellites, P the number of equally spaced orbital planes, and F a phasing parameter setting the relative offset between satellites in adjacent planes. In a Walker Delta, or rosette, pattern the ascending nodes are spread over the full 360 degrees of right ascension at a common inclination below 90 degrees. This gives efficient coverage of populated mid-latitudes and is the basis of most inclined broadband shells, which typically use inclinations between 45 and 55 degrees.

A Walker Star pattern instead distributes ascending nodes over 180 degrees with near-polar inclination, so the planes converge over the poles and counter-rotating planes meet at a seam. This maximizes polar coverage at the cost of redundant satellite density near the poles and sparser coverage at the equator. Iridium and OneWeb both use this form.

GPS is often described in Walker terms but is not a strictly uniform Walker constellation: it places a 24-slot baseline across six planes at 55 degrees inclination with deliberately non-uniform slot spacing within each plane, chosen to improve worst-case geometry for users on the ground. Since 2011 it has operated in an expandable 24-slot configuration, with three slots widened to hold two satellites each, and the operational fleet typically numbers about 31 satellites.

Coverage and Altitude Trade-offs

Raising altitude enlarges each satellite's footprint and reduces the number of satellites needed, but increases path loss, delay, launch cost per satellite, and radiation exposure through the Van Allen belts. Lowering altitude does the reverse and adds atmospheric drag, which shortens lifetime but also guarantees rapid natural deorbit at end of life. Inclination sets the highest latitude served; a constellation inclined at 53 degrees provides no direct coverage of the poles, which is why several operators add a small number of near-polar planes to an inclined core.

Mixed-Altitude and Hybrid Systems

Several operators now run more than one orbital regime, pairing geostationary satellites for broadcast and steady regional capacity with LEO or MEO satellites for low latency and coverage at high latitudes. Hybrid terminals with two apertures, or a single electronically steered array capable of switching between orbits, let a customer draw from whichever layer suits the traffic. The cost is added complexity in network management, terminal design, and spectrum coordination between an operator's own systems.

Inter-Satellite Links

Inter-satellite links connect satellites directly, turning a constellation into a network in space and reducing dependence on a dense global network of ground stations.

Radio-Frequency Links

Radio-frequency crosslinks typically operate at Ka-band or V-band, where allocations exist for inter-satellite service and antennas remain small. Iridium has used Ka-band crosslinks since its first generation, with each satellite maintaining links fore and aft within its plane and to satellites in adjacent planes. Steerable or electronically scanned antennas retrack as relative geometry changes, and link budgets must cope with ranges of thousands of kilometers. Radio links acquire quickly and tolerate pointing error, which makes them robust, but they are bandwidth-limited and consume regulated spectrum.

Optical Links

Laser crosslinks offer orders of magnitude more bandwidth with no spectrum allocation required, and their extremely narrow beams make interception and interference difficult. The engineering challenge is pointing, acquisition, and tracking: beam widths measured in microradians demand precise attitude knowledge, fine steering mirrors, and an acquisition sequence that finds a moving target across a large uncertainty cone. Terminals must also survive thermal cycling and maintain alignment over years. Optical crosslinks are now operational at scale, most visibly in the Starlink constellation, and are baselined for several government relay architectures.

Benefits and Applications

Crosslinks let a constellation carry traffic between arbitrary endpoints without a gateway in view of each satellite, which is what makes service possible over oceans, polar regions, and territories where an operator cannot build ground infrastructure. They also reduce latency on long routes, because light travels nearly 50 percent faster in vacuum than in silica fiber, whose group index is close to 1.47, and a great-circle path through space is shorter than a cable route. Finally, a meshed constellation degrades gracefully: traffic reroutes around a failed satellite or a weathered-out gateway rather than dropping.

Ground Station Equipment

Teleports, gateways, and earth stations aggregate traffic, provide the high-performance end of every link, and house the network operations that keep services running.

Large Aperture Antennas

Gateway antennas typically range from about 6 to 13 meters, with larger apertures at specialized facilities. Cassegrain and Gregorian dual-reflector geometries place the feed near the vertex, which shortens waveguide runs, keeps the low-noise amplifier close to the feed, and points the feed spillover at cold sky rather than warm ground. Efficiencies of 65 to 75 percent are achievable at these sizes.

Surface accuracy governs high-band performance through the Ruze relation, in which gain loss grows with the square of the ratio of RMS surface error to wavelength. At 30 GHz an RMS error of 0.3 millimeters already costs roughly 0.6 dB, so Ka-band reflectors are held to a few tenths of a millimeter and are protected against solar-induced distortion. Drive systems hold pointing within roughly 0.01 to 0.05 degrees against wind loading and thermal drift. De-icing, either by heaters embedded in the reflector or by a radome, preserves the surface in winter weather at the cost of some insertion loss and additional noise temperature.

High-Power Amplifiers

Gateway transmitters use klystrons, traveling-wave tubes, or increasingly solid-state amplifiers, with kilowatt-class output at Ku-band and Ka-band. Tubes retain an efficiency advantage at the highest powers and frequencies; gallium nitride solid-state amplifiers have closed much of the gap while improving reliability, mean time between failures, and maintainability. Redundancy is arranged in N+1 or N+2 configurations with automatic waveguide switching, because a gateway outage removes an entire beam or region from service.

Monitoring and Control

A gateway continuously monitors antenna pointing, forward and reflected power, spectral occupancy, carrier quality, and environmental conditions. Spectrum monitoring receivers detect interference and verify that transmitted carriers conform to their assigned masks. Automatic gain and frequency control hold the chain at its designed operating point. A network management system correlates ground segment telemetry with satellite payload status and terminal statistics, and drives the capacity allocation and ACM decisions that keep the network efficient.

Satellite Tracking Systems

Tracking keeps an antenna pointed as the satellite moves, as the platform moves, or as thermal and mechanical effects shift the antenna itself. The required approach depends on beamwidth and on how fast the geometry changes.

Program Tracking

Program tracking computes the satellite's predicted position from orbital elements and drives the antenna open loop. It suits geostationary satellites, whose apparent motion is a slow figure-eight caused by residual inclination, and it requires periodic ephemeris updates from the operator. Because there is no feedback, accuracy depends entirely on the quality of the elements and the mechanical calibration of the mount.

Step Track

Step tracking closes the loop on received signal strength. The antenna is nudged in azimuth and elevation, the resulting change in beacon or carrier level is measured, and the mount moves toward the maximum. It is inexpensive, requires no special feed, and corrects for mount misalignment and orbital drift, but it dithers the pointing slightly, responds slowly, and can be confused by a fade that changes signal level for reasons unrelated to pointing.

Monopulse Tracking

Monopulse tracking derives a pointing error signal from a single received sample by comparing outputs of a multi-mode feed or a cluster of feed horns, producing independent azimuth and elevation error terms. It is fast, continuous, and does not perturb the main beam, which makes it the choice for narrow-beam gateways, for tracking fast-moving LEO satellites, and for stabilized shipboard and airborne antennas.

Beacon Tracking and Electronic Steering

Most satellites radiate an unmodulated beacon, often at a band edge, that terminals use as a stable reference for tracking, polarization alignment, and fade measurement. Because the beacon is independent of traffic, tracking continues even when the terminal carries no carriers. Electronically steered phased arrays eliminate the mount entirely, switching beams in microseconds, which is what makes practical the fast satellite-to-satellite handovers that LEO constellations require, and what allows a flat antenna on a vehicle or aircraft to track without moving parts.

GPS and GNSS Technologies

Global navigation satellite systems are a specialized branch of satellite technology: one-way broadcast links whose payload is precise time. They share hardware, orbital, and propagation engineering with communications satellites, and communications systems depend on them heavily.

GPS System Architecture

GPS operates a 24-slot baseline constellation in six orbital planes inclined 55 degrees at approximately 20,200 kilometers altitude, with an orbital period near 11 hours 58 minutes, or half a sidereal day. The operational fleet is normally larger, around 31 satellites. Each satellite broadcasts direct-sequence spread-spectrum signals on L1 at 1575.42 MHz and L2 at 1227.60 MHz, with modernized civil signals L2C and L5 at 1176.45 MHz on newer spacecraft. A receiver measures pseudoranges to four or more satellites, solving simultaneously for three position coordinates and its own clock offset.

Rubidium and cesium atomic frequency standards on board provide the stability that ranging demands, and the control segment uploads clock and ephemeris corrections that keep broadcast GPS time closely aligned with UTC as maintained by the U.S. Naval Observatory; the Standard Positioning Service performance standard commits to holding that offset within 40 nanoseconds 95 percent of the time. Civil receivers under open sky typically achieve horizontal accuracy of a few meters, and differential and real-time kinematic techniques using carrier-phase measurements and reference stations reach centimeter level.

Other GNSS Constellations

GLONASS (Russia) uses 24 satellites in three planes at 64.8 degrees inclination, historically distinguishing satellites by frequency rather than by code, with CDMA signals added on newer spacecraft. Galileo (European Union) uses a nominal 24 satellites plus spares in three planes at 56 degrees, and broadcasts an authenticated open service message that lets receivers verify the navigation data originated with the constellation. BeiDou (China) combines medium Earth orbit satellites with geostationary and inclined geosynchronous satellites for enhanced regional performance. QZSS (Japan) and NavIC (India) are regional systems augmenting global coverage. Multi-constellation receivers routinely track well over twenty satellites at once across several systems, which improves availability in urban canyons and under foliage and hardens the solution against interference.

GNSS in Satellite Communication Systems

Satellite networks depend on GNSS at almost every layer. Terminals use their own position to select the correct beam, to apply the regulatory rules of the jurisdiction they are in, and to compute pointing angles automatically at installation. GNSS time disciplines TDMA burst timing, network synchronization, and billing records across an entire network without a terrestrial timing distribution. On board, GNSS receivers give LEO satellites continuous, autonomous orbit determination, greatly reducing the ground tracking effort a large constellation would otherwise require.

Spectrum Regulation and Orbital Coordination

Orbit and spectrum are shared international resources, and access to them is administered rather than purchased outright. Regulation therefore constrains satellite system design as tightly as physics does.

The ITU Framework

The ITU Radio Regulations, revised at World Radiocommunication Conferences held roughly every four years, allocate frequency bands to services and set the procedures by which administrations file for, coordinate, and register satellite networks. A filing must be brought into use within a defined period or it lapses, which prevents indefinite warehousing of orbital and spectrum rights. Geostationary networks coordinate primarily against neighbors in the same band, and the two-degree orbital spacing convention in the C-band and Ku-band fixed satellite service exists because earth station antenna sidelobe performance can be relied on to discriminate at that separation.

Protecting Incumbents

Non-geostationary systems sharing spectrum with geostationary networks must respect equivalent power flux density limits, which cap the interference an NGSO constellation may cause into geostationary earth stations. Meeting them typically requires the constellation to avoid transmitting when a satellite passes near the geostationary arc as seen from a protected receiver, a maneuver known as arc avoidance. Separate power flux density limits at the surface of Earth protect terrestrial fixed services that share bands such as C-band and Ku-band. Off-axis EIRP density masks perform the same function in the uplink direction by limiting what a terminal may radiate toward adjacent satellites.

Deployment Milestones and Debris Mitigation

Because large constellation filings can be speculative, WRC-19 adopted milestone-based deployment rules for non-geostationary systems, requiring an operator to place a defined fraction of its constellation in orbit within set intervals after bringing the filing into use or lose the corresponding rights. End-of-life disposal is regulated separately: long-standing international guidelines called for removing LEO spacecraft from orbit within 25 years of mission end, and the FCC adopted a stricter five-year rule for satellites it licenses in 2022. Low deployment altitudes help by making atmospheric drag itself the disposal mechanism, and operators supplement this with propulsive deorbit, conjunction screening, and coordination with space situational awareness providers.

Interference, Security, and Resilience

A satellite is a wide-open receiver in the sky, visible to everything within its footprint. Managing unintended interference and deliberate attack is a standing operational requirement.

Unintentional Interference

Most satellite interference is accidental: a misaligned VSAT illuminating the wrong satellite, a terminal transmitting on a stale frequency plan, an amplifier driven into intermodulation, or a terrestrial microwave link sharing a C-band channel. Adjacent satellite interference and cross-polarization leakage set practical floors on how tightly carriers and orbital positions can be packed. Carrier identification, in which a low-level spread-spectrum identifier is embedded beneath the traffic carrier, allows an operator to determine which terminal is causing an interfering transmission and is now standard practice for video contribution uplinks.

Jamming and Spoofing

Deliberate uplink jamming can deny a transponder to legitimate users, and downlink jamming can deny reception locally. Mitigations include spread-spectrum waveforms, nulling antennas on the satellite, on-board processing that discards unrecognized uplink signals, and geolocation techniques that fix an interferer's position by comparing signals received through two satellites. GNSS is especially vulnerable because its received signals are extremely weak; jamming denies service, while spoofing substitutes counterfeit signals to induce a false position or time. Countermeasures include controlled-reception-pattern antennas, receiver autonomous integrity monitoring, cross-checks against inertial sensors, multi-constellation and multi-frequency operation, and signal authentication.

Link and Network Security

Commercial systems encrypt traffic over the air, and conditional access systems control which receivers can decrypt broadcast content. Government systems add transmission security to conceal traffic patterns and protect against traffic analysis, along with authenticated command links so that no unauthorized party can command a spacecraft. Ground segment security has become at least as important as link security, because a compromised gateway, network management system, or terminal management platform can affect an entire network.

Emerging Mega-Constellations

Low-cost launch and mass-produced spacecraft have made constellations of hundreds to thousands of satellites practical. These systems dominate current investment in satellite communications and are reshaping the economics of the industry.

Starlink

SpaceX's Starlink is by far the largest constellation deployed. The first-generation system is authorized for roughly 4,400 satellites, and the FCC authorized the second-generation system in two tranches of 7,500, in December 2022 and January 2026, for 15,000 Gen2 satellites in total; SpaceX has filed for substantially more. Most operational satellites fly in shells near 540 to 570 kilometers. Satellites use phased arrays for user and gateway links and optical crosslinks for the network backbone, which removes the requirement for a gateway in view of every satellite. Consumer service typically delivers tens to a few hundred megabits per second with latency in the 25 to 60 millisecond range. Vertical integration, with SpaceX manufacturing its own satellites and flying them on its own vehicles, supports a launch cadence no previous operator has matched.

User terminals are electronically steered phased arrays that acquire and hand over between satellites with no moving parts. Satellites are designed for a service life of roughly five years, which keeps the fleet technologically current and ensures that failed or retired spacecraft deorbit quickly from these altitudes. A separate direct-to-cell payload variant carries an LTE base station serving unmodified handsets in partner operators' terrestrial spectrum.

Eutelsat OneWeb

OneWeb's constellation comprises 648 satellites, including spares, in twelve near-polar planes at about 1,200 kilometers, a Walker Star arrangement that provides coverage to the poles. It uses Ku-band user links and Ka-band gateway links, and it relies on ground gateways rather than crosslinks, so its ground segment is comparatively large. The company merged with Eutelsat in 2023, creating an operator with both geostationary and LEO fleets. Its commercial focus is wholesale: cellular backhaul, enterprise and maritime connectivity, government service, and community access in underserved regions, sold through distribution partners rather than direct to consumers.

Amazon Leo

Amazon's constellation, developed as Project Kuiper and renamed Amazon Leo in November 2025, is authorized for 3,236 satellites across shells at 590, 610, and 630 kilometers. It uses Ka-band user links with phased arrays at both ends, offers several terminal classes from compact consumer units to enterprise-class antennas, and is positioned to integrate with Amazon Web Services for enterprise and government customers. Deployment began in 2025, and several hundred satellites were in orbit by early 2026 against a regulatory deadline requiring half the constellation to be launched by mid-2026.

Challenges and Concerns

Mega-constellations concentrate risk in ways single satellites do not. Collision risk and debris generation scale with the square of object density, and a fragmentation event at a popular altitude would affect every operator there. Radio astronomy faces both direct out-of-band emission and aggregate interference across large fields of view, and optical astronomy faces satellite trails in long exposures. Regulatory oversight is fragmented across national administrations even though the physical effects are global. Operators have responded with lower operating altitudes and rapid disposal, propulsive collision avoidance driven by automated conjunction screening, darkened coatings and sunshades to reduce optical brightness, and coordination agreements with radio and optical astronomy organizations. Whether these measures scale to tens of thousands of satellites is an open question, and it is being tested in orbit rather than on paper.

Spacecraft Bus and Support Electronics

The communications payload depends on a platform that supplies power, holds attitude, manages heat, and remains under ground control for the whole mission. These subsystems consume much of the spacecraft's mass and nearly all of its reliability engineering effort.

The Space Environment

Multiple Access and Network Operation

A transponder or beam is a shared resource, and the method by which terminals share it determines efficiency, latency, and terminal cost.

Future Directions

Satellite communications continues to advance along several parallel lines, driven as much by launch economics and manufacturing as by radio technology.

Higher Frequency Bands

Q-band and V-band feeder links free the Ka-band spectrum for user traffic and are already flying on high-throughput satellites. W-band (75 to 110 GHz) is under study. At these frequencies propagation impairments are severe enough that link availability must be bought with gateway site diversity and adaptive transmission rather than with static margin, which shifts the design problem from the radio to the network.

Software-Defined Satellites

Fully reconfigurable payloads with digital channelizers and on-board beamforming allow a satellite's frequency plan, beam layout, and capacity distribution to be redefined after launch. This decouples a fifteen-year asset from a forecast made twenty years before it retires, and it lets an operator follow demand from one region or market to another.

Non-Geostationary Fixed Services

Regulatory and technical frameworks have matured enough for non-geostationary systems to compete directly in markets that geostationary satellites once held alone, including broadcast contribution, cellular backhaul, and enterprise networking. Operators increasingly run multi-orbit fleets and sell service rather than transponder capacity, with orbit selection becoming an internal routing decision rather than a customer-visible one.

Integration with Terrestrial Networks

The clearest long-term trend is the disappearance of the satellite link as a distinct network element. 3GPP non-terrestrial network specifications place satellite access inside the same architecture, authentication, and quality-of-service framework as terrestrial 5G, and regenerative payloads put the base station itself in orbit. In that model a satellite becomes another radio access node, chosen by the network when it is the best available path, and the user never knows which layer carried the traffic.

Conclusion

Satellite communication systems have grown from a handful of experimental relays into infrastructure carrying broadcast television, broadband access, mobile connectivity, and the precise time that underpins much of the world's other networks. The engineering spans orbital mechanics, radio propagation, microwave hardware, coding theory, phased-array antennas, and network protocol design, and a working system requires all of them to hold simultaneously.

The economics have changed more than the physics. Free-space path loss, rain attenuation, and the speed of light are the same constraints they always were, but cheap launch, mass-produced spacecraft, on-board processing, and near-Shannon-limit coding have shifted what those constraints permit. The result is a field where geostationary broadcast, medium Earth orbit navigation, and low Earth orbit broadband coexist, each best at something the others do poorly.

Whatever the orbit, the discipline is the same: account honestly for every decibel, respect the regulatory limits that make shared spectrum work, and design for the fade and the failure rather than the clear-sky case.

Related Topics