Phased Array Antennas

The principle: constructive interference

An array of small radiating elements emits the same signal slightly offset in time across the array. When all elements are in phase, the wavefronts combine constructively perpendicular to the array's face — the beam goes straight up. When elements are progressively delayed from one side to the other, the wavefront tilts; the beam steers at an angle.

The math:

• Each element radiates with a phase the controller chooses.
• Adjacent elements' wavefronts add (constructive) in the steering direction.
• They cancel (destructive) in other directions.
• The result is a narrow beam pointing where the phases dictate.

Why it matters for LEO satellites

A satellite in low Earth orbit (~500 km altitude) traverses the sky in 5-15 minutes, depending on pass geometry. A fixed dish would lose the satellite quickly; even a motorized dish struggles with the angular velocity. A phased array updates its beam direction continuously to track:

1. Compute current satellite position.
2. Compute the phase delays needed for the beam to point at it.
3. Apply the phase delays.
4. Repeat every few milliseconds.

As one satellite sets and another rises, the array switches its phase pattern to the new satellite — handoff in microseconds at the antenna level (the network protocols take longer to coordinate but the antenna isn't the bottleneck).

Anatomy of a consumer phased array

A Starlink Gen-2 user terminal contains:

• Hundreds of small antenna elements arranged in a flat grid.
• Per-element phase shifters (variable phase delays).
• Per-element amplifiers (each element transmits a small power).
• A central processor that computes phase patterns based on satellite ephemeris.
• A modem that handles the link protocol.

The flat panel form factor is intrinsic to the technology — the elements need to be in a planar array. Round dishes were always going away once mass-manufacturable phased arrays became affordable.

Active vs passive elements

• Passive phased array — single transmitter feeds many elements via a phase-shifter network. Lower cost; less flexible.
• Active phased array (AESA) — each element has its own amplifier and phase shifter. Higher cost; much more flexibility, better noise figure, and ability to form multiple simultaneous beams.

Modern satellite-terminal arrays are typically active (or hybrid).

Multiple simultaneous beams

With per-element control, a phased array can form multiple beams at once — each pointing different directions. Useful for:

• Tracking the current satellite while preparing to track the next (handoff overlap).
• Communicating with multiple satellites simultaneously for redundancy.
• Ground stations communicating with many users in different directions.

Boresight and grating lobes

Phased arrays have geometric limitations:

• Boresight is the direction perpendicular to the array's face. Performance is best there.
• As the beam steers off-boresight, the array's "projected aperture" shrinks (cos of the steering angle), reducing gain.
• At very steep steering angles (e.g., 70° from boresight), the gain drops significantly.
• Grating lobes — unwanted secondary beams that appear if the element spacing is too large compared to wavelength. Element spacing is typically half a wavelength to avoid them.

For a flat-panel terminal aimed straight up, satellites very low on the horizon are harder to use. Hence the Starlink installation guidance to point the dish for the clearest view of the sky overhead, not necessarily true north or south.

Comparison to dish antennas

Why now

Phased arrays have existed in military radar for decades. The consumer breakthrough required:

• Cheap silicon RF integration — per-element amplifiers and phase shifters at consumer price points.
• High-volume manufacturing of dense element arrays.
• Compute cheap enough to run real-time beam steering algorithms.
• A market driver — LEO constellations need them.

All of these arrived together around 2020 with the deployment of Starlink user terminals at consumer scale.

Future directions

• Higher frequency bands. V-band terminals would have smaller elements and finer beam control.
• Conformal arrays. Elements integrated into curved surfaces (vehicle bodies, aircraft fuselages) rather than flat panels.
• Software-defined RF. Element-level digital signal processing for more flexible beam-forming and interference cancellation.

Frequently Asked Questions

What is a phased array antenna?

An antenna made of many small radiating elements arranged in a grid, where the phase of each element's signal is controlled individually. By adjusting the phase relationships, the array can steer its combined beam electronically to point in different directions — no moving parts needed. The principle is constructive interference: elements add coherently in the desired direction and cancel elsewhere.

How does electronic beam steering work?

By controlling the phase delay applied to each element. If all elements transmit in phase, the beam goes straight out from the antenna face. By delaying elements progressively across the array, the wavefront tilts, steering the beam at an angle. The delays are computed by the antenna controller based on the target direction, often updated thousands of times per second to track a moving satellite.

Why use phased arrays for satellite internet?

LEO satellites move fast — completing an orbit in about 90 minutes — so a fixed dish would need constant mechanical re-aiming. Phased arrays steer electronically with no moving parts: faster, more reliable, and able to handle satellite handoffs (transitioning from one satellite to the next as the first sets and another rises) within milliseconds. They also tolerate vibration, motion, and continuous tracking that would wear mechanical mounts.

What are the downsides?

Complexity and cost. A phased array needs hundreds or thousands of individually-controlled elements, each with its own amplifier and phase shifter. Manufacturing at scale is hard; the Starlink dishy is the first mass-market consumer phased array. Power consumption is higher than a passive dish. Performance off-boresight (beam steered far from straight-up) degrades because the projected aperture shrinks.

How fast can a phased array switch beams?

Microseconds. Electronic phase shifters can be updated very rapidly — much faster than required for satellite tracking. The relevant constraint isn't the antenna; it's the network protocols and the satellite handoff coordination. For Starlink's user terminals, beam switching during handoff is essentially instantaneous from a hardware perspective.