Cellular from Space: The RF Challenge Behind Direct-to-Device Connectivity

For decades, cellular communication has relied on one almost unquestioned assumption: the base station is on the ground. Antennas mounted on rooftops, towers and masts were designed to serve nearby devices, operating within relatively short distances, stable links and low latency. Even when satellite communications entered the picture, they remained largely separate, with dedicated terminals, Ku and Ka bands, and architectures that did not speak the language of cellular networks.

That assumption is now being challenged.

A new approach is taking shape: direct cellular connectivity to standard devices, with the base station no longer on Earth but in low Earth orbit, LEO.

The upcoming NG-3 mission of Blue Origin’s New Glenn rocket, set to carry a BlueBird Block 2 satellite from AST SpaceMobile, marks a key step in this transition. Not as a marketing milestone, but as a large-scale system-level test. For the Microwave and RF community, this is a moment where long-standing assumptions are being re-examined.

NG-3 and New Glenn: Why a Heavy Launcher Matters

NG-3 is the third launch of New Glenn, a heavy-lift, two-stage rocket developed by Blue Origin. The mission is scheduled from LC-36 at Cape Canaveral Space Force Station, no earlier than late February 2026.

New Glenn was designed from the outset as a reusable launch system, with a first stage intended for recovery and reuse. The mission follows NG-2, which achieved a successful booster landing. That same booster is expected to be refurbished and reused, reflecting a gradual build-up of operational capability rather than a single leap.

The primary payload is the BlueBird Block 2 satellite, the next generation in AST SpaceMobile’s constellation. The choice of New Glenn, announced in November 2024, was driven by its high payload capacity and its ability to accommodate unusually large structures.

And in this case, large is not an exaggeration.

BlueBird satellites feature a deployable antenna array spanning approximately 223 square meters, one of the largest ever designed for commercial LEO missions. This scale introduces significant constraints in mass, volume and mechanical complexity, while also demanding high orbital precision, directly impacting RF performance after deployment.

A Cellular Base Station in Orbit

AST SpaceMobile is developing a satellite network operating in existing cellular bands, designed to connect directly with standard smartphones and IoT devices without any dedicated satellite terminal.

For RF engineers, this breaks a fundamental assumption: a device designed to communicate with a nearby terrestrial antenna must now maintain a link with one located hundreds of kilometers above.

To enable this, the satellite effectively acts as a massive base station, relying on a large aperture, advanced beamforming and beam steering, and highly sensitive receivers. Still, antenna size is a necessary condition, not a sufficient one.

Why Sub-6 GHz Spectrum Is Essential

The system operates in Sub-6 GHz spectrum, typically in lower cellular bands such as 700 to 900 MHz, and in some cases around 1.9 GHz.

The reasoning follows classic RF principles: lower frequencies offer reduced path loss, better obstacle penetration and higher tolerance to geometric variations.

From a link budget perspective, these are effectively the only frequencies that can, at least theoretically, close a link with a smartphone antenna, which has limited gain and low transmit power. Higher frequencies, and certainly mmWave, would require specialized user equipment, contradicting the Direct-to-Device concept.

However, this choice comes at a cost: congested spectrum, high sensitivity to interference, and the need for extreme beam precision and tight regulatory coordination.

Link Budget: Fighting for Every dB

At the heart of the system lies the link budget challenge.

Free-space losses over hundreds of kilometers are orders of magnitude higher than in terrestrial networks, and cannot be compensated simply by densifying infrastructure or increasing transmit power.

Instead, the system relies on extremely high antenna gain, narrow and focused beams, and strict power management. Even then, the goal is not to deliver full urban capacity, but to serve as a complementary coverage layer, targeting areas where terrestrial networks are unavailable, intermittent, or disrupted.

LEO Latency: Low, But Not Terrestrial

LEO is often described as offering near-terrestrial latency, but reality is more nuanced.

Even at altitudes of around 500 km, round-trip time, RTT, is significantly higher than that of a ground-based base station. While terrestrial networks operate in the range of a few tens of milliseconds, satellite links introduce unavoidable physical distance, along with additional processing and scheduling delays.

The impact is particularly evident in control-plane operations such as handover, signaling, timers and retransmissions in 4G and 5G networks.

The challenge is unique: the satellite system must adapt itself to the expected behavior of existing cellular networks without modifying the end devices.

Doppler Compensation Before the Signal Reaches the Device

A LEO satellite moves at approximately 7 to 8 km per second relative to Earth, creating significant and continuously changing Doppler shifts.

Standard smartphones are not designed to handle Doppler at this scale, meaning the burden falls almost entirely on the satellite system.

To maintain transparent communication, Doppler compensation must be applied dynamically before the signal even reaches the device. From the phone’s perspective, the signal should appear as if it originated from a terrestrial base station.

This is not only an RF challenge, but a system-level one, affecting synchronization, demodulation and overall link stability.

Phased Arrays at Extreme Scale: Power and Thermal Constraints

The BlueBird Block 2 antenna is not simply large, it is a highly complex active phased array, consisting of a vast number of transmit and receive elements.

This enables rapid electronic beam steering, multi-beam operation and dynamic coverage adaptation.

But it introduces another challenge: power and thermal management. Unlike terrestrial systems, a satellite cannot rely on active cooling and must dissipate heat primarily through radiation into space.

Uneven thermal distribution can impact phase stability, distort beams and directly degrade the link budget.

Interference: Lighting the User, Not the Network

Interference management is one of the most critical aspects of Direct-to-Device systems.

These operate in active cellular bands, where dense terrestrial networks are already deployed. Any imprecise illumination can result in co-channel interference.

The solution is inherently systemic: precise beamforming, electronic geofencing, location-aware power control, and continuous coordination with operators and regulators.

In this context, the ability not to transmit is just as important as the ability to transmit.

A New RF Layer, Not a Replacement

Direct-to-Device connectivity from space is not intended to replace dense terrestrial cellular networks. Instead, it introduces a new RF layer, one in which the largest antenna resides in orbit, while the user remains with the same familiar device.

The NG-3 mission represents a shift from isolated demonstrations to a large-scale system validation. For the Microwave and RF community, it is one of the rare cases where link budget, antennas, spectrum, latency and Doppler are all tested together under conditions that have never existed before.

Credits: Based on official publications by Blue Origin and AST SpaceMobile
New-Tech Magazine Group