Direct to Device from Low Earth Orbit

Section 1Regulatory framework

The FCC adopted the Supplemental Coverage from Space (SCS) framework on March 14, 2024. It allows a satellite operator to use a terrestrial carrier's licensed mobile spectrum, under a lease, to serve that carrier's subscribers where the terrestrial network does not reach, by adding a secondary satellite allocation in specific bands and applying Part 25 rules to the space segment, with interim 911 routing requirements and equipment authorization for the handsets [1][2]. The first authorization came on November 26, 2024 for SpaceX and T-Mobile in the PCS G Block, conditioned on out-of-band emission limits and protection of neighboring licensees [3]. On March 7, 2025 the Space Bureau and Wireless Telecommunications Bureau conditionally waived the aggregate out-of-band emission limit in section 25.202(k)(1), after SpaceX explained that as the direct-to-cell constellation approached about 1,500 satellites it would otherwise have to reduce active beams or beam power to stay within the limit [4].

The air interface came from a different body. 3GPP's Release 17 added non-terrestrial network support to 5G NR and to the NB-IoT and LTE-M narrowband standards, with the timing, Doppler, and access adaptations a satellite link needs, following the study items documented in TR 38.821 [5][10]. Releases 18 and 19 continued the work. The first Starlink satellites with direct-to-cell payloads launched on January 2, 2024, and messaging tests with T-Mobile followed within the month [11].

Section 2Link constraints

A conventional satellite terminal has a directional antenna, tens of watts of transmit power, and a receive chain designed for the link. A phone has none of those. A handset transmits at about 23 dBm through an antenna with roughly -3 dBi of gain in the direction of the sky, and its receiver has a noise figure near 7 dB behind that same antenna. Everything the terminal would normally contribute has to come from the satellite: a large phased array for gain, high per-beam EIRP, and a waveform that tolerates the Doppler and delay of a spacecraft moving at 7.7 km/s. The surveys of non-terrestrial networks describe these constraints in general terms [7][8][6]; the numbers below make them concrete for the SCS case.

Path loss at 1.9 GHz is about 24 dB lower than at the 20 GHz of a Ka-band downlink over the same range, which is the only reason the link is possible at all. It is still 150 to 160 dB, and the handset contributes -10 dBW of EIRP on the uplink.

Section 3Uplink budget

With a 23 dBm handset, a -3 dBi antenna, and 3 dB of body and polarization loss, the received carrier-to-noise density at a satellite with a G/T of 5 dB/K is about 61 dB-Hz at 30° elevation and 71 dB-Hz overhead. Figure 3 converts that into carrier-to-noise ratio for three channel bandwidths: the 180 kHz of an NB-IoT carrier, the 1.4 MHz of the narrowest LTE channel, and 5 MHz.

The result explains the service roadmap the operators announced: text first, then voice and data. A 180 kHz channel closes with 8 dB to spare at 30° and nearly 19 dB overhead. A 1.4 MHz channel is marginal below 30°. A 5 MHz channel closes only near zenith unless the satellite has 10 dB/K of G/T or more, which is the reason direct-to-device arrays are the largest antennas on their spacecraft. Doubling aperture area buys 3 dB; there is no equivalent lever on the phone. The narrowband LTE-M and NB-IoT adaptations for satellite that Ericsson's engineers documented exist for exactly this reason [9].

Section 4Downlink budget

On the downlink the handset's receive G/T is about -35 dB/K, so the satellite's per-beam EIRP has to be large. Table 2 gives the EIRP required at the beam center to reach a 0 dB and a 5 dB carrier-to-noise ratio at the handset for the same three bandwidths, at 30° elevation and overhead, including 3 dB of losses.

EIRP is the product of transmitter power and antenna gain, so a 2.7 m array with roughly 32 dBi of gain reaches 30 dBW of EIRP with less than a watt of RF power per beam; the difficulty is not the power but the number of simultaneous beams, each a 26 km spot, needed to cover a service area, and the aggregate emission those beams produce outside the leased channel. That aggregate is what section 25.202(k)(1) limits and what the March 2025 waiver relaxed, on SpaceX's showing that the alternative was fewer or weaker beams as the constellation grew [4]. The emission limit, not the link budget, is the binding constraint on downlink capacity for a mature system.

Section 5Doppler and timing

At 1.9 GHz a satellite at 360 km produces a Doppler shift of up to about 45 kHz at low elevation, sweeping through zero as the satellite passes overhead, with a rate of change of hundreds of hertz per second near zenith. Terrestrial LTE and NR were designed for Doppler of a few hundred hertz. Release 17 solved this on the network side for a GNSS-equipped handset: the satellite broadcasts its ephemeris, the phone estimates its own position, and the phone pre-compensates timing and frequency before it transmits, so the base station on the satellite sees a signal within the terrestrial tolerances [5][10]. Round-trip delay of a few milliseconds and the geometry of a beam moving across the ground at 7 km/s are handled with extended timing advance and scheduling offsets. The consequence for an operator is that the service works only with handsets whose software implements these procedures, or with a satellite that compensates on its own for legacy devices at a cost in capacity.

Section 6Operator obligations under the SCS rules

Three features of the SCS framework shape the business as much as the link. First, the spectrum belongs to the terrestrial partner, not the satellite operator; the satellite operator files a Part 25 modification and the partner files a lease, and the partner must hold all co-channel licenses across a geographically independent area, which limits the arrangement to carriers with contiguous nationwide or regional holdings [2]. Second, the out-of-band emission limit is aggregate across the constellation, so capacity added by launching more satellites is capped unless the operator demonstrates, as SpaceX did, that the limit overprotects adjacent networks [4]. Third, the 911 routing and equipment authorization requirements put the satellite operator inside the public safety and device certification regimes of a terrestrial carrier for the first time [1].

For a new entrant the design questions follow from the link budget. Which service, messaging or data, does the aperture actually support at the elevation angles a real orbit delivers? Which carrier partner holds contiguous spectrum in the target market, and in which band, since path loss and antenna gain both scale with frequency? Will the handsets in that market implement Release 17 NTN procedures, or will the satellite have to compensate for them? And what emission budget is left once the neighbors' protection is accounted for? None of these are answered by a coverage map.

Direct-to-device service is a link budget with the terminal removed. The phone supplies -10 dBW and a -35 dB/K receiver, and every other decibel has to come from the spacecraft.

Section 7Conclusion

Direct-to-device service is a link budget with the terminal removed. The phone supplies -10 dBW and a -35 dB/K receiver, and every other decibel has to come from a spacecraft that also absorbs 45 kHz of Doppler and the emission limits of the terrestrial band it borrows. Narrowband messaging closes with margin across most of a pass; wideband data closes only at high elevation with the largest apertures now flying. The SCS framework makes the service legal in the United States and defines who holds the spectrum, who routes the emergency calls, and how much the constellation may emit in aggregate. An operator planning a direct-to-device payload should size the array to the service it intends to sell, choose the partner by spectrum holdings and handset population, and treat the emission limit as the constraint it will spend the most time negotiating.

References

1. Federal Communications Commission (2024). Single Network Future: Supplemental Coverage from Space, Report and Order and Further Notice of Proposed Rulemaking, FCC 24-28 (adopted March 14, 2024). News release. docs.fcc.gov/public/attachments/DOC-401208A1.pdf
2. Federal Register (2024). Single Network Future: Supplemental Coverage from Space; Space Innovation; final rule, published April 30, 2024, effective May 30, 2024. www.federalregister.gov/documents/2024/04/30/2024-06669/single-network-future-supplemental-coverage-from-space-space-innovation
3. Federal Communications Commission (2024). SpaceX and T-Mobile supplemental coverage from space, Order and Authorization, DA 24-1193 (November 26, 2024). docs.fcc.gov/public/attachments/DA-24-1193A1.pdf
4. Federal Communications Commission (2025). Order granting SpaceX a conditional waiver of section 25.202(k)(1), DA 25-197 (March 7, 2025). docs.fcc.gov/public/attachments/DA-25-197A1.pdf
5. Lin, X., Rommer, S., Euler, S., Yavuz, E.A., Karlsson, R.S. (2021). 5G from space: an overview of 3GPP non-terrestrial networks. IEEE Communications Standards Magazine, 5(4), 147-153. arxiv.org/pdf/2103.09156
6. Giordani, M., Zorzi, M. (2021). Non-terrestrial networks in the 6G era: challenges and opportunities. IEEE Network, 35(2), 244-251. arxiv.org/abs/1912.10226
7. Rinaldi, F., Maattanen, H.-L., Torsner, J., Pizzi, S., Andreev, S., Iera, A., Koucheryavy, Y., Araniti, G. (2020). Non-terrestrial networks in 5G and beyond: a survey. IEEE Access, 8, 165178-165200. doi.org/10.1109/ACCESS.2020.3022981
8. Azari, M.M., et al. (2022). Evolution of non-terrestrial networks from 5G to 6G: a survey. IEEE Communications Surveys and Tutorials, 24(4), 2633-2672. arxiv.org/pdf/2107.06881
9. Khan, T.A., Lin, X., Eriksson Löwenmark, S., Liberg, O., Euler, S., Sedin, J., Yavuz, E.A., Shokri-Razaghi, H., Määttänen, H.-L. (2021). Internet of Things from space: transforming LTE machine type communications for non-terrestrial networks. Ericsson; arXiv:2103.14169. arxiv.org/pdf/2103.14169
10. 3GPP (2020). TR 38.821, Solutions for NR to support non-terrestrial networks (NTN), Release 16. portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=3525
11. TechCrunch (2024). SpaceX launches first batch of direct-to-cell Starlink satellites for testing this year, January 3, 2024. techcrunch.com/2024/01/03/spacex-launches-first-batch-of-direct-to-cell-starlink-satellites-for-testing-this-year