Starlink T-Mobile D2D Test Constellation

This late-night Friday SpaceX is set to launch the first six Starlink satellites with direct-to-device (D2D) capability. This follows the experimental authorization issued by the Federal Communications Commission on Thursday.?

The experimental authorization lets SpaceX test the cellular Starlink system for 180 days. ?Should the first six satellites work as planned, SpaceX will likely step up the launch cadence of D2D-capable sats to enable continuous D2D connectivity in the US.

Test Constellation

As reported by Michael Kan on PC Magazine, testing will eventually involve 840 satellites and 2,000 test devices on the ground. Satellites will communicate with unmodified smartphones in over two dozen US locations including Mountain View, California; Kansas City, Kansas; Redmond, Washington; and Dallas, Texas.

SpaceX plans to launch thousands of D2D satellites but for the T-Mobile test phase, SpaceX will start populating one of the gen2 shells. It is not entirely clear (at least to me) which LEO constellation shell/s SpaceX will use for D2D but, considering that minimizing the orbital altitude is key, I believe it will be the shell with the following characteristics:

• Altitude: 525 km
• Orbital Inclination: 53 degrees
• Number of orbital planes: 28
• Satellites per plane: 30
• Minimum elevation angle for smartphones: 30 degrees
• Minimum elevation angle for gateway antennas: 15 degrees

NCAT4 Simulation

Using the listed assumptions in the non-GEO constellations analysis toolkit (#NCAT4), we can dynamically visualize the test constellation and download datasets (links below) that drive the following calculations:

Round trip delay: 7 to 16 milliseconds for (single hop) smartphone to Starlink gateway. This is propagation-only so it does not consider additional delay by active equipment (e.g. protocol packetization, doppler-shift compensation, etc.)

Smartphone to satellite distance: Ranges between 525 km (90 degrees look angle) to 951 km (30 deg look angle).

Visible satellites: This will depend on the location. For the two sample locations, NCAT4 calculates that:

• In Dallas, TX there will always be between 2 and 5 D2D satellites in line of sight, ~3 on average.
• In Redmond, WA, there will always be between 6 and 10 D2D satellites in line of sight, ~7 on average.

Handoff rate: Another important aspect that the test phase should put focus on is the ability to handle seamless satellite (cell tower) handoffs if goal is to keep continuous connectivity using the best-visible (highest elevation) satellite.

For the 840-sat test phase we should therefore expect better performance results in northern US locations (Alaska excluded). More visible satellites means that smartphones will generally be able to connect with better look angles, minimizing path and atmospheric signal loss.

Download this simulation data (CSV format)

• Starlink satellites
• Timeline of visible satellites
• Sat Tracker

NSR’s Non–GEO Constellations Analysis Toolkit 4.0 (NCAT4) is an assembly of quantitative models that x-ray and benchmark LEO and MEO satellite constellations. The latest edition of NSR’s industry-leading benchmarking toolset has moved online bringing expanded capabilities and processing power. Users can assess the impact of constellations dynamically, across both space and time domains, and with finer resolution and speed than previous spreadsheet versions.