Neat stuff; thanks. It’s been a few years since I was really paying much attention and obviously some advancements have been made in that time. The CuSP MiPS is a cute little thing; a 0.3U cold gas thruster. That would actually be viable on a 1U.
Know anything about Momentus Space? They claim to have a “water plasma propulsion” system. They kinda ping my baloney meter but I’m not entirely sure what to think. They have a slick site that is almost entirely void of technical information.
Gigabit class. The Air Force has already demonstrated in-flight speeds of 610 Mbps. Netflix will be no problem at all.
It’s sorta ironic, but people in rural areas may end up with better internet capabilities than dense urban environments. The best connection I can get in San Jose for reasonable cost is ~100 megabit for $100/mo. Starlink should offer much better prices than that, but only in low-density areas.
I wonder how many they could pack into a single Starship mission. 400? Maybe that’s why there is a rapid push for Starship development.
Why only in low-density areas? It seems like that would require them to hate profit.
If Starlink’s production costs allow them to offer much better prices than $100/mo for 100 megabit, they can sell all they can in cities for $90/mo and make a superprofit.
Where is there the most money to be made in global satnet-enabled connectivity? Consumer-wise, I think it’s phone data but I might well be wrong. Business-wise, I don’t know but it seems like they’ll be able to outbid rural people who want to watch Netflix. Military-wise, I’d expect the most sought-after uses wouldn’t be communication but observation satellites.
Starlink has a maximum bandwidth density per unit area, because there are only so many satellites over a given area, and everyone that lives there has to share the bandwidth of those satellites. A few hundred gigabits/s of bandwidth isn’t that much for even a mid-size city, and even with a significant oversubscription ratio.
So most likely they’ll use that limited bandwidth for things that need it, like airplanes or cell towers in awkward locations. They’ll make plenty of money; I just doubt there will be much remaining bandwidth for end users.
Rural areas won’t have that problem. There probably won’t be enough high-value users to eat all the bandwidth, so it’ll make sense to sell to consumers.
I think that’s what they’re hoping for, yes. They’ve essentially said that the need Starship to build out the full network affordably (though the network will still be useful in incomplete form).
Keep in mind that the majority of them will be over water at any time, and many of the rest will be over low population land areas.
The number over dense areas at any time is probably not going to be enough to push the demands of an entire city.
Part of the sales pitch that Musk is giving to the astronomers is that he will be providing free or low cost internet service to undeveloped areas and countries.
Anyway, at $250,000 each, and 40,000 of them, that’s $10 billion right there. That’s before launch costs and replacements. What’s the lifetime of these things going to be?
This is why, even though every few years since the nineties, I would hear rumors of someone building a satellite internet system, it wasn’t something that you could even contemplate without launch costs coming down substantially. It may still not be economical. Must has launched pretty much all of the starlinks on test rockets. He would have had to put a dummy weight into them anyway, as it was very unlikely that anyone would risk their multi-million dollar satellite on unproven technology, much less pay for the privilege.
Once he actually has a market for launches, and no longer has to launch rockets simply for tests, then there becomes an opportunity cost to putting up more starlink satellites. We’ll have to see how low he manages to get launch costs before we should really get optimistic about the current starlink fleet growing by two and half orders of magnitude.
Astononmers already have to deal with satelites and planes and other ineteruptions to their view of the sky.
It will require more work to screen out all the new objects, but it will be just more of what they already do, they do not need to learn new techniques or technology to do it. (Though, new techniques or technology may be developed in response to it, and that may be a positive improvement to astronomy.
Also, supposedly, according to Frasier Cain, Musk has also promised to put up some space telescopes.
Just because you live on a farm doesn’t mean that you might not want to play Call of Duty with your friends.)
Anyone stuck on geostationary satellite cannot use most of the real time internet functions that many of us take for granted.
They probably would be used to connect ground based systems as well, but they would not really need that long a line of sight. They can communicate with eachother as well, and light actually goes faster in space than it does in the air. (and much faster than in fiber optic cable, and even far much faster than copper transmission lines.) Less attenuation as well. If they are right above, the signal only passes through a few miles of significant atmosphere. If they are at a 45’ angle ,it will be much greater, and at the horizon even more. More noise, less bandwidth.
The best way to sell it would be to have metered connections. The price of bandwidth could be variable based on demand and supply.
If you are in a major city, you probably have multiple options for broadband, and starlink would just be one of them, and would probably usually not be worth the cost to most consumers.
Stock traders, on the other hand, have systems where fractions of a millisecond count, and have paid enormous amounts of money to connect major cities with high speed fiber for their exclusive use.
Starlink would have lower latency than fiber, and so the stock traders would be willing to pay astronomical fees per kilobyte to use the service.
Outside of major cities, where demand is lower, it will be a more affordable option, and may very well be competitive enough to push the current service providers around.
In rural or undeveloped areas, it would be practically or even actually free.
Having a broadband connection anywhere you are on Earth has its uses. It would certainly help Antarctic or Arctic missions.
About 5 years. Launch costs are comparable with the cost of the satellites: about $15 mil per batch.
Not true. The only recent launch with a (literal) dummy payload was the Falcon Heavy, whereas so far Starlink uses only Falcon 9s; the same ones they use for most other launches.
However, what is probably true is that they’re more willing to use end-of-life boosters; ones that would be too high risk for an external customer. Since Starlink sats are cheap and basically identical, they can get away with losing a batch once in a while.
Insurance costs for the Falcon 9 are some of the lowest in the industry. SpaceX’s rockets are considered reliable.
It’s certainly true that they’ll want Starship running before really filling out the rest of the fleet. The first few hundred satellites are the highest value ones; the next few thousand a bit lower value; and the next few tens of thousand lower value yet. Somewhere along this spectrum they’ll start wanting cheaper launch.
I think you are all thinking too small about this. Starlink has absolutely revolutionary potential, especially for the rest of the world.
Musk isn’t just selling to rural Americans. There are hundreds of millions of people in 3rd world countries who didn’t have a hope in hell of ever getting on the internet with a high-speed, low latenchy connection. With Starlink, a small village could buy one $150 pizza box receiver, and provide Wifi access to the people of the village for a few bucks a month each. This could potentially bring high speed internet to a billion people who can’t get it now.
This also provides a way for people in despotic countries like China, North Korea and Iran to get get on the internet fully uncensored. The little antenna would be trivial to hide. This has the potential to break the control on information in locked down countries. This could be the best thing for the third world to come along in a long time.
As for the astronomy issues… Don’t forget that these things are only really visible for a little while after dark - like Noctolescent clouds, they will be illumated for the period where it’s dark on the ground, but the sun still above the horizon from the satellite’s perspective. And just after dusk or just before dawn are not peak astronomy hours, as there is already residual sunlight coming up.
They may occlude the occasional star in the middle of the night, but these things are not large, and are little more than point sources. The problem for astronomers has been a little bit overblown. I already get the occasional satellite trail in my image stack - the stacking software removes them automatically, and that’s just the amateur stacking software I have. The net result of Starlink might be a slight increase in processing times, or the odd extra failed image in a stack of images.
Starlink also opens up a whole new capability for working remotely from tiny homes in rural areas, RVs, farms, etc. One less thing keeping people from leaving high-density cities. This could accelerate outward migration from city centers - especially in the Covid era. My wife and I were talking about this the other day. We both work from home. With Starlink, we could pack up an RV and hit the road and keep working. That’s pretty cool.
Starlink should also put another nail in the coffin of net neutrality, which was always justified by the fact that many people are locked into their ISP and at their mercy. No longer. Everyone on the planet will have at least one new alternetative to their current choices of ISP.
Starlink and other LEO constellations may be a real black swan that no one talked about or even saw coming, but which suddenly brings maybe a billion new peole to the high-speed internet. This could do more for education in 3rd world countries than all the charitable initiatives in history. And it will allow people in remote places to compete in the internet economy. We should all be cheering this, and progressives especially should be ecstatic over the changes in the quality of life this could bring to the poorest people in the world.
So, $20 billion every 5 years. $4 billion a year to keep the constellation up there.
That’s almost talking real money.
But yeah, they could easily make an order of magnitude more than that in subscriber fees.
That’s what I meant. Every time I heard of Space X using a booster for the nth time for the first time, it always seemed to have a starlink payload.
I’m not sure how to best browse their list of launched payloads, did they launch many starlinks on rockets that were not testing the extremes of their performance?
Don’t they need a certain critical mass before it’s really usable at all?
I did point out exactly that, in that was one of the selling points that Musk is using to mollify the astronomers.
China would very well be able to tell that the starlinks are transmitting into their territory, and may take exception to it. They would also be very capable of finding transmitters in their territory. The antenna may be easy to hide from visual detection, but by it’s very nature, is quite obvious in the radio spectrum.
Most likely starlink will not be operating, or be operating under a different “regime” while flying over developed but despotic countries.
North Korea on the other hand, you could probably use it to get around their controls on information, and underdeveloped nations even more so.
If the satellites are constantly being taken down and replaced since they have a lifespan of 4 years, I assume that means it could be a much better system by the late 2020s or so.
If they eventually end up putting up satellites that can handle 1 Tbps of data and they have 100,000 of them, thats a whole different ballgame than 40k satellites that can do 20 Gbps each.
The cost of launch to SpaceX is likely very low as well. They have extra booster cores, and they use Starlink to keep up their launch cadence. Since the fixed costs of SpaceX are huge, the more they launch, the cheaper each launch gets. The chrge for customers includes a proportion of fixed costs, but for spaceX the extra launches are much cheaper.
Yes, but it’s well under 1000. So they can start making money before the constellation is anywhere close to complete. That will pay for further launches and further development. And once Starship is working, launch costs go down by another order of magnitude (they’ll have to reduce satellite costs by the same factor to reap the full benefit, though).
You can see the list of launches for each booster (labeled B10xx) here.
Of the 7 launches so far, two have been the 3rd use, four have been the 4th use, and one has been the 5th use.
That’s a higher average use count than others, but commercial payloads have been sent on the 3rd use of a booster. As of yet, only Starlink and Crew Dragon tests have been sent with a 4th or 5th use of a booster.
They aren’t really testing the extremes of performance, though. The boosters were developed to launch 10 times without significant refurbishment (and 100x with refurbishment).
Customers are already comfortable with reuse. None of these launches would have been with dummy weights; in fact the very first reuse of a booster was with a commercial customer.
It’s probably helpful for SpaceX to have these flights; maybe insurance providers would charge higher rates for missions on heavily-used boosters without evidence that they’re reliable. So if SpaceX can show real data that they’re fine, they can charge more money. But these aren’t “test rockets” in any way. And none of the boosters so far are really at end of life.
Their marginal cost is $15M per launch. $10M of that is the second stage, which they have to build each time. About $1M is reburb of the booster. The rest is incidentals (range costs, etc.).
Keeping up their launch cadence seems super important to me. The current launch market is somewhat cyclic, so having a known “customer” that can smooth out demand sounds super valuable. The constellation is big enough that they can afford to move launches around demanding on what the external demand looks like.
Seems that the Starlink private beta is getting some results:
Download speeds generally from 30-60 Mbps (some lower numbers but these appear to be glitched), upload from 5-17, pings <100 ms (and mostly 30-40 ms). That is not bad at all for a home connection.
This is not yet close to gigabit speeds of course, but this is still pretty early (they only have 540 satellites flying right now). It’s also not clear whether SpaceX will actually offer that speed of a connection to home users. It’s a damn-sight better than HughesNet, which offers ping times of around 800 ms, and max down/up speeds of 25/3. DSL in rural areas can also be terrible–pings are better but the download speed may only be 1 Mbps.
So, even with the current beta service, I can imagine that millions of people would be interested. It wouldn’t compete with cable modems or fiber-to-the-doorstep, but that was never the plan anyway.
Why is the ping low but speed slow in rural areas? Are there situations where the ping is high and the speed fast?
HughesNet–that is, geostationary satellite internet–is somewhat high bandwidth (25 Mbps download) but high ping. That’s just the speed of light that’s the killer–half a second bare minimum to make the 4 trips up to 35,800 km and down to complete a ping. And slightly more than that if the satellite isn’t directly overhead (i.e.you’re somewhere other than on the equator).
DSL has fast pings (no speed of light problem) but the bandwidth falls off rapidly as you get away from the from the exchange. Even just 2 miles away, you’re down to <10 Mbps. And in a rural area this is very likely due to the low density. As for cable (let alone fiber), it’s likely that it just isn’t wired for it.
Low Earth Orbit satellites fix both these problems. Pings are good because the satellites are at <600 km instead of 35,000 km. And low density is a boon because the bandwidth is shared with everyone for a given radio beam spot.
Starlink works just fine in urban areas too, it’s just that due to the bandwidth sharing they’ll probably want to only target high-value customers, like cell phone providers and vehicles.
The current satellites are at 53 degrees inclination, and due to the way orbital mechanics works, the satellites are closer together at this latitude and so for now, it’s only in that band that customers can get 100% coverage. A few degrees less than that leaves gaps. As they fill out the constellation, lower latitudes will be covered.
Some promising reports from the first field users of Starlink:
“I have never set up any tactical satellite equipment that has been as quick to set up, and anywhere near as reliable” as Starlink, Richard Hall, the emergency telecommunications leader of the Washington State Military Department’s IT division, told CNBC in an interview Monday.
Hall said that, with other traditional services, it typically takes between 30 minutes to an hour to set up a satellite connection, “with a lot less speed and bandwidth and a lot higher latency in a much larger package.”
By comparison, Hall emphasized that it took him between five and 10 minutes to set up and connect a Starlink terminal. And a single person can set up one of the devices: “It doesn’t require a truck and a trailer and a whole lot of other additional equipment,” Hall said.
To my mind, what’s encouraging is that the system does not appear to be finicky in the way that beta products tend to be. It took under 10 minutes to set up and they were pleased with the reliability (not to mention bandwidth and latency). Now, SpaceX just needs to send up a lot more satellites so they can cover more than just the northern latitudes.
I saw a train of them in the sky in the pre-dawn hours when I was in rural Oregon, a couple of weeks ago.
They looked just like this. I didn’t know what they were at first and was (un?)relieved to find they were of earthly origin.
We did have sucky internet coverage at that rural vacation house, but we expected it due to its location. Maybe by the next time we stay there, this will have improved due to hordes of Starlink satellites.
Neat! Despite a few tries, I still haven’t spotted a train of them myself. Too much light pollution, too much clutter near the horizon.
The Starlink beta actually just opened to the public:
As the title says, it’s $500 for the dish and $99/mo. Initial speeds are 50-150 Mbps and latencies are 20-40 ms, but they emphasize that this will get better over time. Also, being a beta, there may be service dropouts occasionally. But still, it’s likely a huge improvement for anyone using geostationary satellite internet.