Airbus to build giant satellite network | Hacker News Reader

Airbus to build giant satellite network

bbc.co.uk

91 points_airh60 comments

Wow, they could complete their objective more quickly by taking a billion dollars in small denomination bills and putting it into a pile and lighting it on fire :-) Only slightly more seriously, it would be very impressive if they could pull it off, but having seen the deployment of Iridium (66), GPS (32), GLONASS (24) , and Galileo (30) it doesn't seem like we have a non-nation state that is up to the challenge of putting 600(!) satellites into orbit all at once.

If we assume the reporters got random factoids mixed up, 600 satellites? in 20 "planes" ? (lets assume they are somehow station keeping in the same orbital plane so 30 satellites each trying to stay 12 degrees away from the next satellite in the same plane, while trying to avoid the other 19 planes? One gets destroyed by space junk and you practically guarantee another 29 join it over the next 120 minutes or so? That is going to be quite the dance.

And what spectrum are they using for uplink/downlink? .5 Thz? Something they can get a license for all over the world presumably. That alone cost a billion dollars for Iridium.

I just can't imagine a scenario where this even works. Perhaps if they start with the old TeleDesic design or something.

    "Only slightly more seriously, it would be very impressive
    if they could pull it off, but having seen the deployment of
    Iridium (66), GPS (32), GLONASS (24) , and Galileo (30) it
    doesn't seem like we have a non-nation state that is up to
    the challenge of putting 600(!) satellites into orbit all at
    once."

It looks like these satellites will be a lot smaller than Iridium ones -- from the BBC article, about 150 kg each, vs. ~689 kg [0] for Iridiums. The total orbiting mass is only twice that of the Iridium constellation.

600 * 150 kg = 90,000 kg

66 * 689 kg = 45,474 kg

Naively, they could lift everything in maybe five Ariane V (ECA) launches, for ~$1 billion at market pricing [1]. That'd mean packing over a hundred microsats in one payload fairing. I'm not sure how close this is to reality: I'd guess there'd need to be a lot of structure overhead too.

[0] https://en.wikipedia.org/wiki/Iridium_satellite_constellatio...

[1] http://www.spaceflight101.com/ariane-5-eca.html

($200 MM/launch, 21,000 kg payload to LEO)

Well for one the satellites wouldn't have enough fuel on board to do orbital plane changes, to at a minimum its going to be 20 launches if they want to hit 20 different orbital planes. I have yet to find out how they expect to do station keeping with that low mass.

Shouldn't station-keeping be nearly independent of mass? Devote the same percentage to the fuel tank and use the tiniest working thruster available.

Unfortunately, small thrusters are less efficient than large ones, overgeneralizing of course. There are a number of components that cannot be scaled down indefinitely.

Although there are other potential options (EDTs, etc, etc) that don't have as much difficulty with scaling.

Not true. You can still get ~300s Isp all the way down to a ~1kg thruster: http://cs.astrium.eads.net/sp/spacecraft-propulsion/bipropel...

That's not what they'll be using, though. They'll be using Hall Effect thrusters.

Except that you'd still only need that same 1kg thruster on a 500kg satellite that you'd need on a 10kg satellite. Which is what I was talking about.

With satellites and chemical thrusters, it's pretty much "the lightest possible". Thrust to mass ratio is almost irrelevant. Almost any thruster will have the thrust required.

And I'm kind of surprised by the ion thrusters. I wouldn't assume that they had the thrust required to stationkeep in LEO - the drag in LEO being as high as it is.

First, this is extremely high for LEO, ~1200km. Drag is negligible, decay will take thousands of years.

At around ~800km, solar sails start to become practical because aerodynamic drag drops below photon pressure.

Ion thrusters and photovoltaics are COTS techs which are vastly more powerful than solar sails, which exist only at the prototype/demo stage.

Ion thrusters are practical to fight drag down to around ~200km given the right design, or are perfectly happy at 300-350 given a typical suboptimal design. GOCE managed it down to ~230km, after performing a multiyear mission at ~260km: http://www.spaceflight101.com/goce-re-entry.html

Good to know, thanks.

And interesting. Always nice to see hard numbers.

>Except that you'd still only need that same 1kg thruster on a 500kg satellite that you'd need on a 10kg satellite.

It might be bad for truly tiny sats, but we're talking about 150kg vs. 700kg. 1kg is perfectly fine for those.

I would like to inform you that the events of _Gravity_ were highly dramatized. Space is not so crowded, and most particles from a collision rapidly drift away.

They chose a very high section of LEO, though, 1200km, where debris lasts millennia. Further populating high orbits like this without a solid plan for deorbit is problematic - not for OneWeb, but for the rest of us who don't want to see a Kessler Syndrome cascade thirty years down the road.

From their site[1]:

"End of Life Disposal When a OneWeb satellite nears the end of its intended service life, it will de-orbit automatically, ensuring that the space around our planet remains free and clear for future generations."

So at least they've thought about that.

[1] http://www.oneweb.world/#technologies

It should be noted that "de-orbit automatically" does not mean orbital decay here, but using thrusters to do a de-orbit burn. If the satellite malfunctions, it will become space junk.

Isn't SpaceX's plan to build a satellite Internet service essentially the same thing?

Yes (although with five or six times as many satellites).

And it will likely do it with reusable rockets.

Some answers here: http://www.oneweb.world/

(It's very promotional though, not very technical)

Thanks, so they are going to use Ku band and dodge and weave around the equator to avoid satellites in geosynchronous orbits. And from the animation on their technology page it really does look like they are going to have 30 satellites all in the same orbital plane following one after the other.

They have set for themselves quite the challenge.

Geosynchronous orbits? You mean, 35,000km away from LEO?

Operating in a single orbital plane is really not as hard as you're imagining it to be. We do it all the time. And there are tens of thousands of large objects in Earth orbit, but we only detect collisions every several years.

Their spectrum license requires that they not create radio interference for geosynchronous satellites.

How they plan to accomplish that is not clear to me.

This link http://www.oneweb.world/#technologies with the 'learn more' button shows how the satellites beam to the left and then switch to beaming to the right when they cross the equator, presumably they briefly turn off at the equator to avoid anyone try to transmit to them and have the back scatter continue on up into GEO where it might confuse a satellite up there.

EDIT: Look for "Progressive Pitch" on the web page.

I interpreted him to mean physical collision risk.

As for interference: They're planning on beaming signal to a wide swath of every plane for full-Earth coverage, not just at a narrow stretch near nadir. That means you'll have users acquiring a fix from a wide swath of sky; Let's say 45 degrees from zenith in various directions. That would make every user below 45 degrees latitude encounter fixes with OneWeb proximate to the geosynchronous belt on a regular basis.

If spectral separation alone was insufficient to prevent interference, it's not something that would affect communications when a satellite is near the equator at all, it's something that would affect all the satellites nearly all the time. A 1200km orbit is 30x closer to Earth than the geosynchronous belt.

The link I posted has some context that makes it clear that the statement is about radio interference. Various press statements about the system say that the satellites will turn, for example, search "Progressive Pitch" here:

http://spacenews.com/virgin-qualcomm-invest-in-global-satell...

(or on my earlier link or Wikipedia)

But it must not be thruster based.

Dodge and weave? They're just avoiding aiming down when they are at the equator.

I really doubt they're going to hit anything, at least in a way that hurts more than one sat.

Upon reading, I'm not sure if "planes" refers to aircraft, or surface, as in strata. I'm not sure how feasible that is, but in general the story itself seems light on technical details overall.

My guess was orbital plane (inclination with respect to the orientation of the earth. So everything in the same orbital plane is going round the same direction at the same speed and distance.

In this context, it refers to orbital plane. In this particular case, you can just substitute "orbit".

I love how in like 2 months we've gone from nobody doing this to three separate entities trying to launch an internet of satellites. Gonna get crowded up there.

Ah, Branson vs Musk.

This is one of the things where investing in a fast moving technology reminds me of deflationary economy - we are putting a lot of satellites up there and spending billions doing it when tech is accelerating at such a fast rate that in 10 years the electronics will be quite dated, so it might be worth a wait. There is no easy way to update electronics on these once they're up.

Suppose billionaires wants to have a legacy in space before their time on earth is done!

Personally can't wait until this is cheap enough for normal individuals/startups. Low orbit nanosats can be done for $20k-ish already - www.economist.com/news/technology-quarterly/21603240-small-satellites-taking-advantage-smartphones-and-other-consumer-technologies

Going to get crowded up there!

I'd like to see a server ON a satellite, beyond the reach of governments and regulation. Sell encryption keys and route your ground-based services through a cache up there.

Gonna be hard to replace a faulty hard drive, though.

And arguably if you put a server on a satellite, governments will be the only entities capable of reaching it.

See, there's a way to motivate another space race: Get a bunch of nanosats up there broadcasting material that's offensive to as many governments as possible.

I think that would eventually motivate people to clean up space junk, once China, India, US and Russia blow up a few pirate satellites and everyone realizes we can't get to orbit for awhile.

Periodically shotgun a bunch of RPIs (or their sensible equivalent) on just big enough satellites. The "servers" are distributed across the little nodes. Assume they'll fail and burn up, or just burn up, and keep on shotgunning them up there as long as subscriber fees are profitable enough.

Maybe they don't even need to be profitable if they also serve some sort of propaganda/free speech function for regions of the world that the developed West wants to poke at.

Edit: or shoot big satellites up there, containing the little satellite/server swarms. Spit some more out when necessary.

Yes - the automated X-37 can get to it no problem.

Anybody with a dish and a key can reach it, right?

Probably. Then again anyone with a transmitting dish can be triangulated by the government body that doesn't like people doing that.

But satellite TV is used in tons of places, with most systems of the past decade also being able to transmit.

People having a satellite dish pointed at a satellite there would be in the range of hundreds of millions alone in europe.

Your TV dish is not transmitting.

But the point is, that if some country decides the new internet via satellite to be illegal, it is _very_ easy for them to find people using it.

Not true. Could be in of a van. Or hidden on a roof. Iraq had once forbidden TV satellite dishes; thousands were in use.

A dish leaks very little energy except in the direction of the satellite. It could be astronomically unlikely anyone could detect its use unless they were line-of-sight between the dish and the satellite. And pointed in the right direction.

And the reason they could hide those TV dishes under roofs is that they are not transmitting.

If you want internet via satellite, they are transmitting, and while they are directional towards the satellite, there's enough signal going off in other direction sides making it not that hard to detect from the ground.

The dish actually supports transmitting, in case we’d want to use it for satellite internet, but we haven’t used it for that, yet.

I meant that in the same sense as my parent comment; "reach" as in "take down" or "tamper with".

We already have satellite-killing technology, so I think your idea of 'beyond the reach of governments' is fanciful.

Indeed! https://www.youtube.com/watch?v=RaKpJ7i-JXA

Ballistic missile defense is a very interesting field. It actually seems uncannily defensive given the rest of the "Defense" Industry, ignoring potentially destabilizing MAD. =P

I'm not sure they would be able to hit something as small as a nano sat with a missile. The X-37 could do it but that's expensive unless you really, really don't want whatever the satellite is doing to continue.

Nano satellites would be cheaper, by orders of magnitude, than the facilities needed to 'shoot them down'. Anyway the idea was, out of reach of regulation.

You better be ready to handle solar storms! :)

a new theorem in the making - CAS: consistency, availability, solar storm tolerance.

It's very interesting that the reported price was ~ 1/2 million $ per 150kg sat. SpaceX's launch price to LEO is ~ $4.6k/kg, or ~ $700k per 150kg. While the price to launch these guys would be a bit higher (they need something to aggregate a bunch of sats into a single payload), normally the cost to build the sat is a lot bigger than SpaceX's launch price.

Volume, volume, volume.

Few unmanned spacecraft flown to date are in principle more complex than a Toyota. If you did 1-unit automotive production runs, and 1-unit production runs for most of the parts rather than using standardized templates, things would be expensive there too.

I thought one world was something new it turns out that it's actually been announced since February this year.

My main question is how fast is it going to be? Can I expect 10Mbps to each terminal / user?

"...OneWeb, a British Channel Islands-registered concern..."

If they're so british why don't they think they should pay tax like the rest of us?

I don't think the phrasing was trying to play up the Britishness of the company, it was merely saying that the Channel Islands are British.

Of course, it's still a tax dodge and the company should pay tax like the rest of most of the world!

No, I know, I was just taking a cheap stab.

Assuming at least 40KM distance from the earth, (40/300) * 2 means a roundtrip time of 267ms. That's without processing time, but that's probably in the ballpark of <5 ms.

This excludes many types of internet use cases.

I assume you mean 40,000 kilometres here.

That'll definitely be lower than that (at that kind of height, you wouldn't need hundreds of satellites for global coverage). Wiki page talks about 800km and 950km orbits.

800km is not so far, it's much closer than Europe to North America, US East Coast to West Coast, etc. A round-trip at speed of light should take about 5.3ms.

The speed of light is 300 Mm/s, not 300 Km/s.

Reply on news.ycombinator.com