News error of the week goes to... (SpaceX telecoms thingy)

[watercooled]

I'm not usually one to point out minor content errors (except where providing feedback might be useful - I get that articles have to be rushed out sometimes) but it's at least mildly irritating that something claiming to be a technology news website drops a clanger of an error in how something technology-related works: https://www.techspot.com/news/73407-...atellites.html

Spotted it?

We've all had issues with GPS connectivity and that has only 24 satellites that don't move in relation to the user.

don't move in relation to the user.

Except they do, don't they... Some of the comments are equally facepalm-worthy, pointing and laughing at the error but getting it completely wrong themselves.

And they're really comparing apples and oranges. Don't get me wrong, I do have my own (fairly substantial) reservations about what this network is claiming to do - e.g. the sheer volume of sustained launches to maintain the constellation is orders of magnitude larger than anything... ever, there's a finite amount of spectrum (and the fact that they'll need licensing in every country they cast spot beams on, perhaps even switching dynamically as they cross different regions?), atmospheric effects of the bands they choose to use, and not least the presumed requirement for receivers to need phased array receivers which don't come cheap, though economies of scale could drive prices down. I think some reports are extrapolating a bit too much on what SpaceX are actually claiming e.g. Techspot referring to its use in handsets which I don't think is what SpaceX are claiming at all?

I mean there must be some sanity check the project has passed to make it even this far, but despite all the media hype and 'all telecom companies are evil' guff floating about, they still have to deal with the same reality as everyone else. As a company they're uniquely well-positioned to get stuff into orbit and for presumably a decent price, but there's so much they've left out about their explanations that I'm not jumping on that bandwagon quite yet.

[g8ina]

I heard a gem on the Discovery channel last week, when a voice over said that an image to do with the astronomy appeard on APOD - "Astrology Picture of the Day"

[TeePee]

For what it's worth, these test satellites weight 400kg each. A Falcon Heavy can lift 63800kg to LEO. That means 150 or so per launch (they are 110x70x70cm in size. while the Falcon has a payload fairing size of 13.1m in height and 5.2m in diameter, so they should all fit.). With a lifespan of optimistically seven years, that means 80 initial launches and a rate of under 12 launches per year to sustain the constellation.

At $150M per launch, this means $12B to launch the constellation and 1.8B/year to maintain it, based on launch costs alone. That's peanuts when you consider the potential revenue.

All this talk about Mars is a sideshow, this is what SpaceX is for.

[Bagnaj97]

All this talk about Mars is a sideshow, this is what SpaceX is for.

No, SpaceX is for Mars. This is going to fund it.

[watercooled]

For what it's worth, these test satellites weight 400kg each. A Falcon Heavy can lift 63800kg to LEO. That means 150 or so per launch (they are 110x70x70cm in size. while the Falcon has a payload fairing size of 13.1m in height and 5.2m in diameter, so they should all fit.). With a lifespan of optimistically seven years, that means 80 initial launches and a rate of under 12 launches per year to sustain the constellation.

At $150M per launch, this means $12B to launch the constellation and 1.8B/year to maintain it, based on launch costs alone. That's peanuts when you consider the potential revenue.

All this talk about Mars is a sideshow, this is what SpaceX is for.

Thanks for the input - some interesting points raised. To get pedantic, it's not as straightforward as just adding satellite masses together to see how many could be theoretically jammed into one fairing (size is another consideration). For this sort of launch and using Iridium as an example, it's typical to launch one plane of satellites per vehicle as the delta-v required to adjust the phase (position in the plane) from there is relatively minor, however given the size of the SpaceX constellation and the number of planes, it's perhaps a bit optimistic to assume they could launch 150 in one go even if that was physically possible to do - the delta-v required to changes planes is non-negligible (read: huge in comparison to phasing) and would add considerable mass to the launch.

Therefore, assuming that 150-per-launch doesn't already take that into account, I suspect the launch costs will be substantially higher. I'm not suggesting it's impossible, it's just on a scale not seen before so going from that to 'routine' is far from straightforward.

WRT cost and I know I'm seriously over-extrapolating, but they suggest 40 million users at $30B per year which works out at, on average, $62.50 per month per user. Again I know it's too much of an approximation to take this with anything but a pinch of salt, but at least in UK terms that's definitely on the high side even in retail (inc taxes etc) prices for broadband. Some people seem to be assuming this will be both faster and cheaper than everything, I think that's a heck of an optimistic assumption to be making this early on.

Then comes the issue with spectrum. Something people who endlessly complain about mobile data caps vs landline don't seem to get, is spectrum is very much finite and the Shannon limit unavoidable. Cellular radio covers far smaller areas, allowing frequency re-use, and the capacity limits are very real even there. Furthermore, different frequency bands have different atmospheric properties making some outright unsuitable for LEO-ground use. Even with entire bands at their disposal (which isn't going to happen) I'm dubious about the bandwidth claims. A while ago I did some back-of-envelope calculations for 5GHz (almost the entire Ku band for reference, and compared to ~20MHz carriers on LTE which is very spectrally efficient and likely far better SNR than this) of spectrum (assuming each spot has that much available) and with a very generous SNR got ballpark 30Gb/s. More recently, I've read claims of 23Gb/s which is within that but would still require a TON of spectrum and good SNR which is unlikely effectively impossible for anything other than line-of-sight and clear weather for V-band. And I'm assuming the spot beams are dynamic as 23Gb/s isn't much for a satellite serving a city or something, promising gigabit connections to end-users. So like I said, I'm really interested in how exactly they're planning to do it.

And then there's transponder+receiver complexity/cost for that sort of bandwidth.

If this was from almost anyone but SpaceX I'd be filing it with the rest of the 'we have this amazing idea that beats all the big players' claims that never materialise. Being from a company like SpaceX though, it does have me curious, but I'd really like to know more about how they intend to actually overcome the substantial technical, financial and legislative challenges rather than just assuming they will do so.

Edit: Reading that back it sounds really pessimistic. It's not indented that way, it's just how I see it at the moment and I'm genuinely interested in how they plan to overcome those challenges - if anyone knows where to read more about it please let me know!

[TeePee]

It is interesting to see technological advancement in such a theatrical way. There are certainly a lot of challenges, and the frequency issues sound like a big one.

In terms of cost, $62.50 is very cheap in the US for Gigabit service. That's a fairly rare level of performance here. A few cities have Google Fiber at $70 a month, but most people are stuck paying Comcast more than that for 60Mbps service.

I'd really like to see this succeed if they can overcome the technical challenges.

[Millennium]

For what it's worth, these test satellites weight 400kg each. A Falcon Heavy can lift 63800kg to LEO. That means 150 or so per launch (they are 110x70x70cm in size. while the Falcon has a payload fairing size of 13.1m in height and 5.2m in diameter, so they should all fit.). With a lifespan of optimistically seven years, that means 80 initial launches and a rate of under 12 launches per year to sustain the constellation.

At $150M per launch, this means $12B to launch the constellation and 1.8B/year to maintain it, based on launch costs alone. That's peanuts when you consider the potential revenue.

All this talk about Mars is a sideshow, this is what SpaceX is for.

Sorry, they weigh 400kg and are 1.1*0.7 m ? What are they made of, cold pressed Latinum?

[TeePee]

Sorry, they weigh 400kg and are 1.1*0.7 m ? What are they made of, cold pressed Latinum?

Probably a lot of Titanium and Silica. I mean, there's a lot of stuff to cram in to a small space. There are a couple of folding solar panels for power, which will deploy in orbit. Probably more expensive than Latinum, even the gold pressed variety...