Starlink Aero: in-flight outages and the real margin for manoeuvre
Today we focus on Starlink Aero and on why micro-outages or brief drops can occur during flight, even with a "good" signal shown in the app. The LEO architecture requires continuous handovers between satellites and cells; at high speed and during turns, the geometry changes rapidly and any shadow cast by the fuselage or drift cuts the line of sight for an instant. The system's ESA antenna manages pointing and handover autonomously, with no controls to "lock" a satellite or set a cell preference; the user cannot force persistence. The position of the radome is critical: if the stabiliser casts a shadow at certain angles, or the radome becomes wet and introduces additional losses, the probability of outages increases when crossing dense cloud or rain. Power transients on the 28 V DC bus and heater spikes can cause resets or thermal degradation; if the supply margin is tight, the antenna resets and the event appears as a "network drop". During approach or taxiing phases with tall obstacles around the fuselage, the elevation angle worsens and handovers become more sensitive to cell congestion. When the cabin uses VPNs or tunnels, the effective MTU decreases and fragmentation issues emerge; MSS clamping and aggressive QUIC/TCP keepalive reduce broken sessions during handovers. The cabin router matters more than it might seem: double NAT, slow DNS or subnet collisions between crew and passenger segments are mistaken for "satellite failure". Poorly configured cabin Wi-Fi — uneven transmit powers, overlapping channels, roaming without 802.11k/v — adds jitter and losses that are attributed to Starlink without cause. In severe cold or after a prolonged cold soak, the thermal stabilisation time of the radome and antenna extends the start-up period; this is not a bug, it is basic physics. The system prioritises control traffic and may limit throughput if it detects gateway congestion; there is no local setting to bypass that policy. Useful diagnosis separates layers: measuring energy on the bus, checking connectors and grounds, reviewing antenna logs, examining real obstructions by attitude, and recording link losses correlated with bank and pitch. If there are no obstructions or severe weather and the outages coincide with repeatable route segments, the cause is usually cell saturation or poor geometry; relocating the radome or adjusting flight profiles is not practical, but it helps to understand the pattern. When blocking originates from CGNAT and inbound sessions, there is no "port opening" option: the solution is SD-WAN with persistent tunnels or ground-based gateways with a public IP. For operational use, bypass mode and an aeronautical router with per-application QoS policies and per-destination health detection prevents a speed test from degrading the cabin experience. Maintenance events should include an electrical load test of the antenna, verification of radome losses under wet conditions and a cabin Wi-Fi audit, not merely a one-off speed test. On aircraft with high vibration, elastic mounts or bracket flexion degrade electronic pointing; a rigid mount with correct torque and sealing reduces spurious dropouts. None of this allows the user to "lock" the satellite: the user's margin lies in a flawless installation, a well-designed cabin network and resilient tunnels. When the service is critical, the reasonable strategy is to combine Starlink Aero with a traditional low-rate but high-availability backup link (L/Ka GEO or L-band such as Iridium/Inmarsat) and policy-based automatic switchover. This maintains minimum connectivity during the inevitable gaps inherent to the LEO environment and prevents a handover from compromising operations.
NASSAT - Network Satellite Systems