Mobility satellite links are RF paths that have to hold lock, timing, and spectral discipline while the terminal is moving. The motion can be slow — a maritime VSAT rolling through a few degrees of pitch — or fast, like a low-earth-orbit (LEO) tracking antenna on an aircraft crossing a beam boundary every 90 seconds. The design problem isn’t just a static link budget with a few dB of margin. It’s a dynamic system where antenna pointing error, Doppler shift, blockage, and handover all eat into the margin at the same time. For engineers working C-band through Ka-band, the practical question is how to allocate those losses without ending up with a 3-meter dish on a drone.
This article covers the RF chain, antenna selection, tracking loops, Doppler management, and the link budget structure for mobile terminals. It assumes you already know the basics of free-space path loss and G/T. What follows is the part that usually gets learned the hard way: the moving platform’s effect on every dB in the chain.
Start with the Motion Profile, Not the Modem
Before selecting a modem or coding rate, define the platform’s motion envelope. A maritime terminal on a stabilized pedestal has different pointing errors than a land-mobile terminal on a vehicle roof. An airborne terminal has Doppler rates that a ship will never see. The motion profile drives three parameters: maximum angular acceleration, maximum angular velocity, and the expected blockage duration.
For a ship in sea state 5, a 1-meter antenna on a 3-axis stabilized pedestal might see residual pointing error of 0.2 to 0.5 degrees RMS. For a vehicle-mounted flat panel with electronic beam steering, the residual error can be under 0.1 degree if the array is well calibrated, but the blockage from bridges and buildings is far worse. The link budget must treat these as separate loss terms, not a single “pointing loss” number.
Pointing Error Loss
Pointing error loss for a parabolic antenna is approximately:
Lpoint (dB) ≈ 12 × (e / θ3dB)²
where e is the RMS pointing error in degrees and θ3dB is the half-power beamwidth. A 1.2-meter Ka-band antenna at 30 GHz has a 3 dB beamwidth of about 0.6 degrees. A 0.3-degree RMS pointing error costs roughly 3 dB. That is not a small number. It is the difference between a 16APSK link and a QPSK link, or between closing and not closing at all.
For phased-array terminals, the loss is not a simple parabolic function. Array pointing error manifests as scan loss and beam squint, and the effective aperture shrinks as the scan angle increases. A flat panel scanned to 60 degrees off boresight loses roughly cos²(60°) in effective aperture — about 6 dB. That is before atmospheric loss. Designers who treat a phased array as a fixed-gain antenna are making an expensive mistake.
Doppler: The Moving Target Problem
Doppler shift for a LEO satellite at Ka-band can exceed ±200 kHz. For a GEO satellite, the satellite itself is nearly stationary, but the terminal’s motion still creates a shift. A vehicle moving at 30 m/s toward the satellite at 20 GHz creates a Doppler shift of about 2 kHz. That is small, but the Doppler rate matters more. An aircraft accelerating through a turn can produce a Doppler rate of hundreds of Hz per second. The modem’s carrier recovery loop must track that rate without cycle slips.
The practical rule: if the Doppler rate exceeds about 10% of the symbol rate per second, the demodulator will struggle. For a 1 Msps carrier, that means a Doppler rate above 100 kHz/s is a problem. LEO systems handle this with pre-compensation from the terminal’s ephemeris data. GEO systems usually ignore it, but high-speed airborne terminals should not.
Pre-Compensation vs. Closed-Loop Tracking
There are two ways to handle Doppler. The first is open-loop pre-compensation: the terminal knows its own position and velocity from GNSS, knows the satellite ephemeris, and calculates the expected Doppler shift. It then offsets the transmit frequency so the signal arrives at the satellite on the correct frequency. This is standard practice for LEO terminals and works well when the ephemeris is fresh.
The second is closed-loop tracking: the modem measures the received frequency error and adjusts the transmit frequency accordingly. This works for GEO links where the Doppler is small and slowly varying. For LEO links, closed-loop tracking alone is usually too slow. The loop bandwidth needed to track a fast Doppler rate introduces phase noise that degrades the link. Most serious LEO terminals use both: open-loop pre-compensation for the bulk of the shift, and a narrow closed-loop for residual error.
Antenna Selection: Parabolic vs. Phased Array
The choice between a parabolic antenna on a stabilized pedestal and a phased-array panel is not just about cost. It is about the motion profile, the frequency band, and the regulatory environment.
Parabolic antennas offer higher aperture efficiency and better cross-polarization isolation. A well-built 1-meter Ka-band dish can achieve 60% efficiency and 30 dB of cross-pol isolation. A phased array at Ka-band might achieve 40% efficiency and 20 dB of cross-pol isolation. That 20 dB cross-pol number matters for regulatory compliance. Many satellite operators require 25 dB or better cross-pol isolation to avoid interfering with the adjacent polarization. A phased array that cannot meet that spec will not be approved for use on a shared transponder.
Phased arrays win on mechanical simplicity. No motors, no slip rings, no pedestal maintenance. But they lose on scan loss, thermal management, and cost per element. A Ka-band phased array with 1,024 elements is not cheap, and the beamforming ICs dissipate real power. A 30 GHz array with 10 dBm per element output power and 20% power-added efficiency will draw over 50 W just for the transmit chain. That is before the modem, the upconverter, and the cooling fan.
Hybrid Approaches
Some terminals use a hybrid: a mechanically steered parabolic antenna for the elevation axis and electronic beam steering for the azimuth axis. This reduces the number of phase shifter elements while keeping the aperture efficiency of a dish. The tradeoff is mechanical complexity in one axis and a slower tracking response. For maritime terminals, where the roll and pitch rates are moderate, this is often the best compromise.
Link Budget Structure for Mobile Terminals
A mobile link budget differs from a fixed link budget in three places: pointing loss, blockage margin, and handover margin. The rest of the budget — free-space path loss, atmospheric loss, rain fade, modem implementation loss — is the same as a fixed link.
Here is a worked example for a Ka-band maritime terminal on a GEO satellite:
- Frequency: 20 GHz downlink, 30 GHz uplink
- Satellite EIRP: 52 dBW
- Free-space path loss at 20 GHz, 38,000 km: 210.1 dB
- Atmospheric loss (clear sky): 0.4 dB
- Rain fade margin (99.5% availability): 4.0 dB
- Terminal G/T: 16 dB/K (1.2 m dish, 120 K system noise temperature)
- Pointing loss: 1.5 dB (0.3° RMS error, 0.6° beamwidth)
- Blockage margin: 2.0 dB (occasional mast shadowing)
- Modem implementation loss: 1.0 dB
- Required C/N for 16APSK 3/4: 12.5 dB
The resulting link margin is about 2.5 dB. That is workable, but it leaves little room for antenna misalignment, radome loss, or a hot LNB. If the terminal is a phased array with 3 dB of scan loss, the margin disappears. This is why many maritime Ka-band terminals still use parabolic antennas on stabilized pedestals.
Blockage and Shadowing
Blockage is the most under-budgeted loss in mobile links. A ship’s mast, a vehicle’s roof rack, an aircraft’s wing — all create periodic shadowing that can last from milliseconds to seconds. The link budget must include a blockage margin, but the real solution is at the protocol layer. TCP does not handle 2-second outages well. A mobile terminal needs a modem with deep interleaving, or a link-layer retransmission scheme, or both.
For LEO systems, blockage is also a handover problem. When the terminal switches from one satellite to another, there is a brief period where neither satellite is at the optimal pointing angle. The handover margin — typically 1 to 3 dB — accounts for the reduced gain during the transition. If the constellation has overlapping coverage, the handover can be make-before-break, and the margin is smaller. If not, the margin must be larger.
Spectrum Monitoring and Interference in Mobile Contexts
Mobile terminals are more likely to cause and receive interference than fixed terminals. A moving terminal can point its transmit beam at an adjacent satellite if the tracking loop loses lock. A phased array with poor sidelobe control can spray energy across the geostationary arc. Regulatory bodies take this seriously. The ITU and national regulators require mobile terminals to demonstrate that they can maintain pointing accuracy and shut down transmit if tracking is lost.
Spectrum monitoring on the terminal side is also important. A mobile terminal passing through a port or an urban area will encounter terrestrial interference that a fixed terminal never sees. A spectrum analyzer integrated into the terminal’s receive chain can log interference events and correlate them with position. That data is valuable for troubleshooting and for filing interference reports with the satellite operator.
Mute-on-Loss Requirements
Most satellite operators require mobile terminals to mute the transmit carrier within a specified time after losing the receive signal. The typical requirement is 100 ms to 1 second. This is not a suggestion. A terminal that keeps transmitting while its antenna is pointed at the wrong satellite can cause interference across an entire transponder. The mute circuit must be independent of the modem — a hardware timer that cuts the HPA bias if the tracking loop reports a loss of lock for more than the allowed duration.
Testing and Verification
You cannot fully test a mobile terminal on a bench. The motion environment matters. The standard approach is a three-phase test program:
- Static testing: Verify the link budget, modem performance, and antenna patterns on a fixed mount. This catches most RF chain problems.
- Motion simulation: Mount the terminal on a motion platform that reproduces the expected angular rates and accelerations. Verify that the tracking loop maintains lock and that the pointing error stays within budget.
- Field testing: Install the terminal on the actual platform — ship, vehicle, or aircraft — and run the link through the full motion envelope. This is where blockage, multipath, and thermal issues show up.
Field testing is expensive, but it is the only way to validate the blockage margin and the handover behavior. A terminal that works perfectly on a motion simulator can still fail on a ship because the mast shadowing pattern is different from the simulated one.
Practical Takeaways
If you are designing a mobile satellite link, here is what I would put on a sticky note above the bench:
- Define the motion profile first. Pointing error, angular rate, and blockage duration drive the entire design.
- Budget pointing loss separately from atmospheric loss. Do not lump them together.
- For LEO terminals, use open-loop Doppler pre-compensation. Closed-loop alone is not enough.
- Check the cross-pol isolation spec before committing to a phased array. Many arrays do not meet the regulatory requirement.
- Include a mute-on-loss circuit that is independent of the modem. This is not optional.
- Test on a motion platform before you test in the field. The bench will not show you the real problems.
The difference between a mobile link that works and one that works only on paper is usually 3 dB of pointing loss and 2 seconds of blockage. Budget for both, and the link will survive the real world.
Frequently Asked Questions
What is the biggest mistake in mobile satellite link design?
The most common mistake is treating the terminal as a fixed antenna with a small pointing loss. A mobile terminal’s pointing error varies with the platform’s motion, and the loss is not constant. Designers who use a single 0.5 dB pointing loss number are usually off by 2 to 3 dB in the real environment. The fix is to measure or simulate the actual pointing error distribution and budget for the 95th percentile, not the mean.
How much Doppler shift should a Ka-band mobile terminal expect?
For a GEO satellite, the satellite’s own motion is negligible, but the terminal’s motion creates a shift. A vehicle at 30 m/s creates about 2 kHz at 20 GHz. An aircraft at 250 m/s creates about 17 kHz. For LEO satellites, the dominant Doppler comes from the satellite’s orbital velocity, which can exceed ±200 kHz at Ka-band. The terminal must pre-compensate using ephemeris data, or the modem will lose lock.
Can a phased-array antenna replace a parabolic dish for maritime Ka-band?
It depends on the regulatory requirements and the link margin. Phased arrays offer mechanical simplicity but typically have lower aperture efficiency, higher scan loss, and worse cross-pol isolation than a parabolic dish. If the satellite operator requires 25 dB cross-pol isolation and the link budget has less than 3 dB of margin, a phased array may not close the link. For lower-frequency bands or LEO constellations with more margin, phased arrays are increasingly viable.
What is the typical mute-on-loss requirement for mobile terminals?
Most satellite operators require the terminal to mute its transmit carrier within 100 ms to 1 second after losing the receive signal. The exact requirement varies by operator and frequency band. The mute circuit should be a hardware timer independent of the modem, so a software crash cannot leave the transmitter on while the antenna is mispointed.
Next in this series: a detailed look at tracking loop design for stabilized pedestals, including the tradeoff between loop bandwidth and pointing error under sea state 6 conditions.
