Hacker News Reader

Comment by mg | Hacker News Reader

Do I understand it correctly, that the satellite communicates with a ground station and the ground station communicates with the phones?

It's not that the phone communicates directly with the satellite, right?

View on HN

No, that would be utterly pointless. It's LTE from the sky: https://www.starlink.com/business/direct-to-cell

Not pointless, that's the best use of Starlink with phones. Base station backhaul gets the benefits of not having to run fiber + space for a bigger antenna w/ better signal to noise ratio

SNR is not a function of antenna size. The size is dictated by the wavelength, not how desperate you are to receive a signal.

Directionality does make an antenna larger as it usually involves reflectors and directors, the size and placement of which is also a fraction of the wavelength, but if they are actually able to do direct to phone with decent reception and arbitrary coverage, then that beats having to set up base stations to cover a small circle even if you get 10-15 dB better SNR.

This might give you LTE in, say, all of the Australian outback, all of Alaska, along the Andes mountain range, etc.

I don’t think availability of fiber is the bottleneck here.

I think the problem with fiber in general is not availability, but deployment cost. You need to dig and bury the fiber. That's expensive.

The deployment cost is even lower when you don’t need to put a tower on the ground, which is the point of this. They can cover the entire U.S. to eliminate coverage gaps using the starlink satellites.

You could mount it on poles.

For remote areas that might still be tens to hundreds of kilometers of poles and fiber just to add coverage to one remote area. Then you move on to the next.

No one wants to do that across all of Alaska or the Australian outback.

We have cellular issues in upstate NY that has plenty of infrastructure. Dark fiber is also practically everywhere

Sure, but there's also cellular issues in the Australian outback and most of Alaska, and there you don't have infrastructure.

You need power as well as fiber.

In the sense "you can buy fiber cheaply by the ton", not. But in the sense "laying down fiber in the mountains of Papua-New Guinea, or in the Canadian Arctic, or a simmering warzone full of militias", that is indeed a major bottleneck.

So, location-dependent.

It's absolutely the bottleneck in some remote areas, and satellite backhaul for terrestrial mobile networks has been a thing before Starlink, as far as I know.

It is outside of cities.

A phone has enough transmission power to communicate with a satellite?

I would think the electromagnetic waves the phone sends are going out in a sphere, right? Then the loss on the way to the satellite must be gigantic.

A phone has enough transmission power to communicate with a satellite?

For the last 30 years.

And the Starlink satellites are much closer to Earth than the satellites we used to hit back then.

Not on LTE protocols. Not using a COTS 4G/5G Phone. Satellite phones aren't in the context of this discussion.

You can just about get SMS emergency coverage with the latest generation of modems and aerials designed for the purpose. NTN standards are a long way off being set.

> NTN standards are a long way off being set.

NTN phones have been available since the early 2010s, some even using geostationary satellites (which are 100 times farther away).

If you mean 5G NTN, I believe the newer Pixel models already use that already, and I've been using it using a dedicated device since Summer 2023. They actually use at least some of the same satellites of the 2010s voice service, which I believe is no longer being offered.

I'm talking specifically the ongoing 3GPP work from Rel-16 to Rel-19, and the proposed work by Qualcomm and Thales to validate the use-cases for formalisation of the standard.

Rel-18, the initial release for 5G-Advanced, only brought in the support for the new frequency bands for NTN above 10 GHz - the Ka band where most of our interest lies for large-scale large-bandwidth provisionment - in June of last year for equipment manufacturers.

https://www.lightreading.com/6g/5g-advanced-arrives-with-3gp...

You're not wrong in saying that most of what we hope to be achieved with NTN is, to be fair, NTN-NR - or 5G NTN - as a lot of it depends on NR technologies like MIMO and beamforming in certain spectrums.

Skylo did the first rollout of emergency SMS in Europe as per the 3GPP Release 17 specifications for Direct-to-Handset (D2H) connectivity in Q4 of last year. Note that Rel 17 designated only two usable frequency bands in Release 17 for 5G NTN use in the 5G FR1 range

https://www.broadbandtvnews.com/2024/11/26/deutsche-telekom-...

> Skylo did the first rollout of emergency SMS in Europe as per the 3GPP Release 17 specifications for Direct-to-Handset (D2H) connectivity in Q4 of last year.

I've been already using that exact service in Summer 2023 in both the US and Europe (Bullit, via Sklyo and using Echostar/Terrestar satellites, I believe). It's possible that that was using pre-release specifications, though.

It's curious how this hooks up to existing telecom networks. I always thought of Starlink as eventually bankrupting them.

Phone -> Starlink Sat for the initial connection. That then pings to a starlink groundstation, then it gets routed over their fiber to T-Mobile over the internet (its all just data going over fiber anyway).

> fiber

But with starlink internet as far as I know they bypass traditional ISP's.

Why would that make a difference? You're not connecting to tmobile's network directly, you're connecting to starlinks, they're just routing traffic as part of their deal. It's effectively a glorified proxy service.

If the earth could be covered by only star link ground stations what would be the use of all the satelites?

The stations would not have to be connected via data cables.

Reply on news.ycombinator.com