Effective Aperture and the λ² in the Friis Equation: Why Starlink Mobile Adds 800 MHz

Essence: Free space does not absorb high frequencies: a phone antenna of fixed gain collects an area proportional to wavelength squared — about 8 dB more power at 800 MHz than at 2 GHz — while between fixed-size dishes the same equation favors the Ka-through-W bands of Starlink V3 backhaul.

First Principles Foundation

Two facts carry the whole result: one from conservation of energy, one from antenna physics.

Conservation of energy. In a vacuum nothing absorbs a radio wave. A transmitter radiating power P_t spreads it over a sphere of area 4πd², so the power flux density at distance d is

S = P_t G_t / (4π d²)

where G_t is the transmit antenna's gain toward the receiver (G_t = 1 for an isotropic radiator). S falls as 1/d² and contains no frequency at all.

Effective aperture. A receiving antenna delivers to its load the power that crosses an effective area A_e:

P_r = S · A_e

Antenna theory ties that area to the antenna's gain at wavelength λ:

A_e = G λ² / (4π)

For an isotropic antenna (G = 1), A_e = λ² / 4π. The relation follows from reciprocity — an antenna receives with the same pattern it transmits with — and a thermal-equilibrium argument fixes the constant. It is the only place a wavelength enters a free-space link.

Symbols: P_t and P_r are transmitted and received power (W). G_t and G_r are gains relative to isotropic (dimensionless). A_et and A_er are effective areas (m²). d is the separation (m). λ = c / f is the wavelength (m), with c = 299,792,458 m/s.

Step-by-Step Derivation

1. Combine flux and aperture:

   P_r = P_t G_t A_er / (4π d²)

2. Write the receive area through its gain, A_er = G_r λ² / (4π):

   P_r / P_t = G_t G_r (λ / 4πd)²

   This is the Friis transmission equation in gain form. For isotropic antennas its reciprocal, (4πd / λ)², is the free-space basic transmission loss of Recommendation ITU-R P.525. In decibels: 32.4 + 20 log₁₀ f + 20 log₁₀ d, with f in MHz and d in km.

3. Write both antennas as areas instead, using G = 4πA_e / λ²:

   P_r / P_t = A_et A_er / (λ² d²)

   This is the form Harald Friis published in 1946. The same physics now says received power rises with frequency.

4. Both forms are one equation. Which way frequency pushes depends on what stays constant when the band changes:

   • Fixed gain at both ends: P_r ∝ λ² ∝ 1/f². Lower frequency wins.
   • Fixed area at one end, fixed gain at the other: substitute G_t = 4πA_et / λ² into the gain form and the wavelength cancels, leaving P_r / P_t = A_et G_r / (4π d²). A fixed-size transmit array makes a narrower beam at higher frequency, which raises the flux on its axis by exactly the area the fixed-gain receiver loses.
   • Fixed area at both ends: P_r ∝ 1/λ² ∝ f². Higher frequency wins.

5. Apply the first case to 800 MHz and 2 GHz. The wavelengths are 0.375 m and 0.150 m. An isotropic antenna's effective area is 111.8 cm² at 800 MHz and 17.9 cm² at 2 GHz. The ratio is (2000 / 800)² = 6.25, or 7.96 dB. P.525 gives the same gap at any distance — at 10 km, 110.5 dB at 800 MHz against 118.4 dB at 2 GHz — because distance and transmit power cancel out of the ratio.

6. Hold the received power fixed instead. With fixed gains, P_r ∝ (λ / d)², so the reachable distance scales with λ: a 2.5× longer wavelength reaches 2.5× farther in free space for the same signal.

7. Assumptions. Free space, far field, matched polarization and impedance, no absorption. Walls, terrain, foliage, the user's body, rain, and atmospheric gases add their own losses on top of this term, each with its own frequency dependence. None of them is in the Friis term.

Application to SpaceX / xAI

On October 8, 2026, SpaceX said it had agreed to acquire a nationwide portfolio of up to 14 MHz of paired 800 MHz spectrum for Starlink Mobile, pending final FCC approval. It assigned the bands distinct roles: its global 2 GHz mid-band spectrum provides high-bandwidth capacity, and the new low-band spectrum is a coverage layer meant to reach phones through obstacles such as walls. SpaceX said most existing phones already support the band, and that Starlink Mobile will combine its satellite-to-mobile constellation with a terrestrial deployment.

The phone is the first case. An unmodified phone's antenna is not sized to one band. It is built to be roughly omnidirectional across many bands, so to a first approximation its gain is fixed and its effective area scales with λ². At 800 MHz it presents about 6.25 times the collecting area it presents at 2 GHz: about 8 dB more received power from the same flux, before any wall enters the link. The same factor helps the uplink whenever the receiving antenna is also fixed in gain.

A satellite array of fixed size is the second case. The FCC approved the Starlink Mobile Gen2 constellation that same week: 15,000 satellites optimized for 2 GHz spectrum, which SpaceX says will deliver more than 100 times the bandwidth of its current generation to unmodified devices. For a transmit array of fixed physical size, the λ² of the array's gain cancels the λ² of the phone's aperture. What changes with frequency is the beam, which narrows and covers a smaller footprint. That is a capacity lever — more, smaller beams reusing the same spectrum — not a link-budget penalty. SpaceX's update does not say how its satellite arrays are sized or which segment will carry the 800 MHz band, and this lesson makes no claim about either.

The backhaul is the third case. The same FCC action authorized the full range of Ka, V, E, and W backhaul bands used by Starlink V3 broadband satellites, so they can operate with the existing gateway network. Gateways and satellites both use antennas of fixed size, so the aperture form applies. As an illustration, holding both areas fixed and moving from about 20 GHz to about 80 GHz multiplies received power by (80 / 20)² = 16, or 12 dB, and those higher bands also offer far wider allocations. The same equation that favors 800 MHz at a fixed-gain phone favors the high bands between fixed-size antennas.

Implications and Limits

• Coverage is not capacity. A low band buys signal margin, not bandwidth. Shannon's limit, C = B log₂(1 + SNR), caps a narrow channel no matter how strong the signal. SpaceX's 800 MHz acquisition is up to 14 MHz paired, which is why it calls that band coverage and 2 GHz capacity.
• Free space is only the first loss. Building entry loss (modeled in Recommendation ITU-R P.2109), diffraction around obstacles, and body loss sit on top of the Friis term. SpaceX's statement that low band penetrates walls is about those terms, which this lesson does not derive.
• Small antennas at long wavelengths. A phone-sized antenna is electrically smaller at 800 MHz than at 2 GHz. Chu's 1948 limit ties a small antenna's achievable bandwidth and efficiency to its size in wavelengths, so real handsets give back part of the 8 dB aperture advantage at low band. How much is specific to each device.
• High bands need pointing and clear air. The fixed-area gain at E and W band arrives as a narrow beam that must be steered accurately, and rain and atmospheric gases — with oxygen absorption peaking near 60 GHz — take losses that free-space Friis leaves out.
• One link, not a network. The equation says nothing about interference, beam scheduling, or which satellite or tower serves which phone.

NVIDIA stack: Sionna RT, NVIDIA's open-source, GPU-accelerated, differentiable ray tracer for radio propagation, is the tool for carrying this free-space term into real building, street, and terrain geometry. It was not run for this lesson; every number above is closed-form.

Sources & Further Reading

• SpaceX, "Building the World's Most Advanced Mobile Network," spacex.com/updates, October 8, 2026: up to 14 MHz of paired 800 MHz spectrum pending FCC approval; 2 GHz for capacity and low band as a coverage layer; FCC approval of the 15,000-satellite Starlink Mobile Gen2 constellation; Ka/V/E/W backhaul bands for Starlink V3 satellites.
• H. T. Friis, "A Note on a Simple Transmission Formula," Proceedings of the IRE 34(5): 254–256, 1946. The original aperture form of the equation.
• Recommendation ITU-R P.525-5 (11/2024), "Calculation of free-space attenuation."
• Recommendation ITU-R P.2109, "Prediction of building entry loss" — the term this lesson leaves out.
• L. J. Chu, "Physical Limitations of Omni-Directional Antennas," Journal of Applied Physics 19: 1163–1175, 1948.
• C. A. Balanis, Antenna Theory: Analysis and Design (Wiley): effective aperture and A_e = G λ² / 4π.
• J. Hoydis et al., "Sionna RT: Differentiable Ray Tracing for Radio Propagation Modeling," arXiv:2303.11103.
• On this site: Starlink: Connecting the World.

Independent educational fan project. Not affiliated with Space Exploration Technologies Corp. (SpaceX) or xAI Corp. All content is for educational purposes. Sources cited where applicable.