Ka-Band Satellite Communication Link Budget Simulator

17.7-30 GHz Ka-band links pay a heavy rain-fade penalty, but in return they unlock huge bandwidth and high-gain narrow beams. Vary the orbit (LEO/MEO/GEO), satellite EIRP, ground antenna and climate zone, and watch free-space path loss and link margin react live — the same physics that decides whether your Starlink or JCSAT link survives a tropical squall.

Parameters

Orbit

LEO is closer, so FSPL is much smaller

Satellite EIRP

dBW

Effective isotropic radiated power

Ground antenna

Sets diameter and efficiency

Ka-band frequency

GHz

20 GHz = downlink, 30 GHz = uplink

Data rate

Mbps

Rain attenuation

dB/km

Specific attenuation per km of rain

Antenna efficiency

Ground climate

Climate-zone factor on rain loss

Results

—

Free-space loss (dB)

—

Antenna gain (dBi)

—

Received power (dBW)

—

C/N0 (dBHz)

—

Required C/N0 (dBHz)

—

Link margin (dB)

—

Link geometry — satellite, ground antenna, rain

Satellite altitude moves with the orbit, and the RF beam changes from blue to red as rain loss grows.

Link margin vs rain attenuation

Orbit comparison — FSPL and link margin

Theory & Key Formulas

$$\text{FSPL}_{\text{dB}} = 20\,\log_{10}\!\left(\frac{4\pi d}{\lambda}\right), \qquad \lambda = \frac{c}{f}$$

Free-space path loss. d: propagation distance [m], λ: wavelength [m], f: frequency [Hz]. Grows with (d·f)².

$$G_{\text{ant}} = 10\,\log_{10}\!\left(\eta\left(\frac{\pi D f}{c}\right)^{2}\right)$$

Parabolic antenna gain. D: aperture [m], η: aperture efficiency (0.5-0.75). Larger dish and higher frequency mean higher gain.

$$\frac{C}{N_0} = \text{EIRP} - \text{FSPL} + G_{\text{ant}} - L_{\text{rain}} - L_{\text{atm}} - L_{\text{pol}} - k - T$$

C/N0 link equation. k = -228.6 dBW/Hz/K, system noise temperature taken as 20 dB-K. Link margin = C/N0 - required C/N0.

What this Ka-band link-budget simulator does

🙋

How can that tiny pizza-box Starlink antenna deliver fibre-grade 200 Mbps? Aren't satellites supposed to be 36,000 km up?

🎓

Great question. Starlink doesn't use GEO — its satellites fly in low-Earth orbit at just 550 km, 1/65 the GEO distance. Free-space path loss scales with d², so that single change saves about 36 dB. Try switching the orbit selector to LEO-550 with the 60 cm dish; the link margin stays green. Now flip it to GEO-36000 and watch the margin collapse into the red.

🙋

Wow, distance alone makes that much difference! Then why do GEO operators like JCSAT bother with Ka-band at all?

🎓

They buy back the 36 dB of loss with bigger ground antennas — a 3 m VSAT pulls in about 54 dBi of gain, 9 dB more than a 1 m dish. GEO also gives you a single beam covering the whole country, and the satellite doesn't move, so the dish doesn't need active tracking. Starlink's LEO satellites are always moving, so the user terminal needs a phased array to steer the beam electronically — the dish is cheap, but the digital RF front-end is anything but.

🙋

When I bump rain attenuation from 5 to 15 dB/km the margin turns red instantly. Is rain really that brutal?

🎓

That's the Achilles heel of Ka-band. The wavelength is 10-15 mm, the same order as raindrop diameter (1-5 mm), so every drop scatters and absorbs the wave like a tiny antenna. Rain fade is 3-5 times worse than at Ku-band (12 GHz / 25 mm). That's why tropical operators rely on adaptive coding & modulation (ACM) to fall back from 200 Mbps to 20 Mbps when needed, and on site diversity to switch to a ground station that isn't under the same rain cell. Set the climate to tropical and see what happens.

🙋

Why does cranking the data rate from 200 to 1000 Mbps cost 7 dB of margin?

🎓

Shannon's theorem makes this unavoidable. To deliver one bit reliably you need a fixed Eb/N0, so required C/N0 = Eb/N0 + 10·log10(Rb). A 10× rate increase costs exactly 10 dB; a 5× increase costs about 7 dB. You can't engineer your way out of this — that's why real modems run ACM and quietly cut throughput during storms instead of dropping the carrier.

Frequently asked questions

When the carrier wavelength becomes comparable to typical raindrop diameter (1-5 mm), Mie scattering and absorption strip a large amount of energy from the wave. Ka-band wavelengths are 10-15 mm, so each raindrop effectively re-radiates the signal in all directions. Compared with Ku-band (12 GHz / 25 mm wavelength), rain attenuation is 3-5 times larger, and operators accept link outage for 0.1-1% of the year (about 99.9% availability) in tropical regions. LEO systems like Starlink have plenty of margin to absorb this, but GEO operators rely on site diversity or low-SNR modes (DVB-S2X ACM) during heavy rain.

The Starlink LEO satellite sits at ~550 km, only 1/65 of the GEO distance (36,000 km). Free-space path loss scales with d², so this alone saves about 36 dB — the same as adding 36 dB of antenna gain. A 60 cm dish has roughly 4 dB less gain than a 1 m VSAT, but the 32 dB net win from the shorter range still lets it deliver 200 Mbps-class throughput. Switch the orbit in this tool from LEO-550 to GEO-36000 and watch FSPL jump by about 35 dB, turning the link margin negative.

A common industry rule of thumb is 3 dB or more under clear sky, and zero or better during the rain peak. Below 3 dB, you no longer cover implementation losses such as pointing errors, polarisation mismatch and LNB noise-figure drift. This tool flags linkMargin > 6 dB as green (OK), 3-6 dB as orange (caution), and below 3 dB as red (NG). For real projects you should also evaluate the rain-attenuation CCDF (ITU-R P.618) at 99.5% or 99.9% availability.

From Shannon's theorem the required C/N0 grows as 10·log10(Rb) for a fixed Eb/N0 target. Multiplying the data rate by 10 raises the required C/N0 by exactly 10 dB. If EIRP and antenna gain are unchanged, going from 200 Mbps to 2000 Mbps therefore costs you 10 dB of margin. This tool lets you feel the trade-off with one slider. Real systems mitigate this with Adaptive Coding & Modulation (ACM) — the modem drops to a more robust MODCOD during rain to keep the link alive at a lower rate.

Real-world applications

Starlink, OneWeb and other LEO broadband: Starlink at 550 km and OneWeb at 1,200 km enjoy 30-36 dB less FSPL than GEO systems, so a 60 cm home dish can sustain 100-500 Mbps. Because the satellite moves overhead in seconds, the terminal needs a phased-array beam, and inter-satellite optical links can route packets directly from Japan to the US faster than terrestrial fibre. Pick "Starlink Dishy" + "LEO-550" in this tool to see the physics that makes it work.

JCSAT, Inmarsat GX and other GEO services: A single geostationary satellite blankets an entire country, which is ideal for broadcast, maritime and remote-island connectivity. Operators trade 35 dB more FSPL for 1-3 m high-gain VSATs and 500 MHz-class Ka-band transponders to deliver 50-200 Mbps. Typhoon-season service design includes Ku-band fallbacks and inter-beam handover for the inevitable Ka outages.

O3b mPOWER MEO constellation: At 8,000 km altitude, MEO sits between LEO and GEO with ~100 ms latency and a wide equatorial footprint. Cruise lines, remote mines and offshore platforms run on it. Selecting "MEO-8000" in this tool shows a comfortable margin that isn't as generous as LEO but far better than GEO.

Rain-aware service design: Engineers feed weather-radar and MODIS precipitation data into regional margin budgets — Okinawa or Southeast Asia might need 12-15 dB of extra rain margin, while Hokkaido or Scandinavia gets by with 5 dB. By switching climates here you see immediately how many extra dB of EIRP each region would demand to keep the same availability.

Common misconceptions and pitfalls

The first trap is thinking "free-space path loss is energy absorbed by the medium". The name is misleading — no power is lost to free space. FSPL is simply the ratio between the surface area of the d-radius sphere over which the transmitted power has spread, and the effective aperture (λ²/4π) of an isotropic receiver. That is why FSPL grows with frequency on paper, even though it is really the receive aperture shrinking. Use the same physical dish at a higher frequency and the gain actually increases. In this tool, push the carrier from 17.7 to 30 GHz: FSPL goes up, but antenna gain rises faster, and C/N0 ends up better.

Second, do not treat rain attenuation as a simple constant. The "rain dB/km × 5 km" model used here is for didactic purposes; real fade depends on frequency (γ_R = k·R^α with rain rate R), elevation angle (low angles cross longer slabs of rain), and polarisation (vertical is slightly less attenuating than horizontal). Critically, ITU-R P.618 specifies A_0.01 — the rain attenuation exceeded only 0.01% of the year — and that statistic drives 99.99% availability design. For real projects, redo the calculation with the regional ITU-R rain map and its statistical formulas.

Finally, link margin is not synonymous with a single safety factor. Margin must cover at least four buckets: (1) rain attenuation, (2) implementation losses (pointing, polarisation, LNB noise), (3) long-term noise drift, and (4) satellite end-of-life RF degradation. Ten dB of clear-sky margin made up of "rain 8 + impl. 2" will fail in the tropics, while "rain 3 + impl. 7" can still be fine in the Arctic. Subtract another 2-3 dB of implementation loss from the margin you read here to estimate the real operational headroom.

Ka-Band Satellite Communication Link Budget Simulator

What this Ka-band link-budget simulator does

Frequently asked questions

Real-world applications

Common misconceptions and pitfalls

How to Use

Worked Example

Practical Notes