SpaceX Internet Satellite Network: Starlink

[mongo]

I don't see large calculation errors (didn't check small ones Are you planning 2. stage with larger propellant tanks? I don't think SpaceX will do it, because launch cost of BFR should be smaller than launch cost of F9.

I think they should copy ion drive from Dawn Mission | Mission
Record-breaking use of solar-electric propulsion: 11km/s in space Δv.
Active powered flight 65% of the time, equivalent to 5.5 years of continuous thrust.

BFR would send larger patch of satellites. Satellites would change orbital planes with ion drives. I believe those satellites need lot of power for data link operation. That could be used by ion-drive.

I was going with worst case, so I assumed the 2nd stage would normally be out of fuel after initial 50 sats were in orbit. The recent Elon tweet (Feb 12) indicated lengthening the second stage was easy.

11km/s is 10x of delta v needed, but the actual thrust (90mN) may be too low to change the orbital plane in a timely manner due to making a small push every orbit (.55 years total engine time based on Dawn). I suppose it could do a less efficient (fuel wise) or more complex maneuver to speed things up, but that is beyond my orbital mechanical ability .

Oh sure, BFR will be even better eventually, multiple planes per launch, but you fight with the army you've got.

 

[Kant.Ing]

As long as they will beam and work with Mars' system, that seems alright.

 

[jkn]

I was going with worst case, so I assumed the 2nd stage would normally be out of fuel after initial 50 sats were in orbit. The recent Elon tweet (Feb 12) indicated lengthening the second stage was easy.

11km/s is 10x of delta v needed, but the actual thrust (90mN) may be too low to change the orbital plane in a timely manner due to making a small push every orbit (.55 years total engine time based on Dawn). I suppose it could do a less efficient (fuel wise) or more complex maneuver to speed things up, but that is beyond my orbital mechanical ability .

Oh sure, BFR will be even better eventually, multiple planes per launch, but you fight with the army you've got.

Acceleration (trust) of ion drive is limited by available power. These satellites might have rather large solar panels anyway. Dawn used ion drive about 2.5 AU from the Sun. Earth gets 6 times as much solar energy. LEO satellite is on shadow half of the time, so only 3*. With ion drive satellite would do more than short bursts. But it cannot trust continuously. Perhaps 1/2 of time. This reduces efficiency, but still better than chemical energy.

 

[mongo]

Acceleration (trust) of ion drive is limited by available power. These satellites might have rather large solar panels anyway. Dawn used ion drive about 2.5 AU from the Sun. Earth gets 6 times as much solar energy. LEO satellite is on shadow half of the time, so only 3*. With ion drive satellite would do more than short bursts. But it cannot trust continuously. Perhaps 1/2 of time. This reduces efficiency, but still better than chemical energy.

Good point on available energy, and it has 9x fuel mass advantage compared to a Merlin Vac so total energy needed is less.

The short burst thought was based on not knowing how much you can spread out the ideal instantaneous burn at the initial/ final orbital overlap point. It passes the point every 110 minutes or so, 13 times a day, half the times at night, so 6 manuver opportunities per day at full solar based thrust.
If they were launching two sets from one plane, the initial orbit could be halfway between finals and each set would only need half the adjustment...
Know any good simulators?

 

[Bipo]

After thinking this (and sleeping), I decided my first idea is correct. Change of orbital plane requires change in direction of velocity even if inclination is same. Doing that with v=sqrt(earthmass*G/(1150 km + earthradius)) = 7.281 km / s is difficult. This makes BFR less efficient in launching those.

For polar satellites it's a lot easier: you launch all the satellites on the same orbital plane (let's say 85°), but you don't rise them to the final service altitude. Then, in order to change of plane you just need to wait. Eventually, the faster orbital speed of a lower orbit will translate into a relative polar plane shift. Then you give a little deltaV to rise each one to it altitude at the precise plane.

This is nothing new... It's already used by Iridium to relocate their spare satellites on the proper orbit:

Iridium satellite constellation - Wikipedia

Significan orbital plane changes are normally very fuel-intensive, but orbital perturbations aid the process. The Earth's equatorial bulge causes the orbital right ascension of the ascending node(RAAN) to precess at a rate that depends mainly on the period and inclination. The Iridium satellites have an inclination of 86.4°, which places every satellite in a prograde (inclination < 90°) orbit. This causes their equator crossings to steadily precess westward.

A spare Iridium satellite in the lower storage orbit has a shorter period so its RAAN moves westward more quickly than the satellites in the standard orbit. Iridium simply waits until the desired RAAN (i.e., the desired orbital plane) is reached and then raises the spare satellite to the standard altitude, fixing its orbital plane with respect to the constellation. Although this saves substantial amounts of fuel, it can be a time-consuming process.

 

[mongo]

For polar satellites it's a lot easier: you launch all the satellites on the same orbital plane (let's say 85°), but you don't rise them to the final service altitude. Then, in order to change of plane you just need to wait. Eventually, the faster orbital speed of a lower orbit will translate into a relative polar plane shift.

In the case of Iridium (from your reference), isn't the effect due to increased impact of the Earth equatorial bulge (more iterations and stronger effect) i.e. indirectly due to orbital height/speed? My (limited) understanding is that raising/ lowering an orbit does not change the plane (speaking of absolute plane relative to Aries, not relative position to Earth (geostationary case))

 

[jkn]

In the case of Iridium (from your reference), isn't the effect due to increased impact of the Earth equatorial bulge (more iterations and stronger effect) i.e. indirectly due to orbital height/speed? My (limited) understanding is that raising/ lowering an orbit does not change the plane (speaking of absolute plane relative to Aries, not relative position to Earth (geostationary case))

As I understand this: Precession rate depends on altitude. Change of altitude will change orbital plane relative to satellites on different altitude.

Those satellites must have large batteries for operation on night side. So ion-drives could be used on night side also. I think I could calculate how to change orbital plane with two 1/4 orbit thrusts per orbit. Instead of it we need orbital simulation with mass distribution of Earth included. I don't know if there is one freely available.

SpaceX LEO constellation uses inclinations 53, 53.8, 74, 81, 70 degrees and VLEO 53, 48, 42. Not exactly polar, but I guess close enough for precession. SpaceX can use BFR to launch those. Perhaps initial deployment of 1600 satellites to 53 degree orbit with FH.

 

[Bipo]

That's it: precession is tied to you number of orbits, so if you orbit more times per day (i.e. a lower orbit) your plane will shift westwards faster

 

[mongo]

That's it: precession is tied to you number of orbits, so if you orbit more times per day (i.e. a lower orbit) your plane will shift westwards faster

@jkn
Thanks. Found the right Wikipedia article. Very cool effect.

For the 1,125 km orbit, the sats precess -3.4 degrees per day.
At 900 km (for example, also circular), they process at -3.78.
So a shift of 0.38 per day. To move one 11.25 plane would take 29.6 days. Orbital velocity difference is 7.4 km/s vs 7.289 km/s, delta v of 112 m/s.
Much better than the rocket based change now value.

 

[dhanson865]

7. For good measure, it is not a stretch to see that GPS/Glonass are rendered irrelevant when this is in place.

Realistically, Starlink threatens many, many special interests.

So true for most of your post (voted informative), but can it really replace GPS in my cell phone if the receiver size needs to be laptop/small pizza box sized?

If it could increase GPS accuracy that would be huge. Not so big a deal for the flatlanders in the middle of the US but us that live in or on the edge of a mountain range lose GPS quite easily at times.

 

[bxr140]

So launches, plus satellites, plus operating and maintenance costs could be around 2.5 billion a year.

I think it is very doable.

Its definitely doable, even with a very likely higher per month cost. It is not going to be smooth sailing, however.

Amortizing industrialization of the spacecraft production line digs deep into ROI which, at the risk of sounding like a broken record at this point, is why you see a lot of talk about internet constellations but not a ton of action.

Ground terminals are another step function to overcome for starlink. More traditional RF constellations--even using the 'new' RF technology--are basically adjacent hardware/technology to what we already have on the ground. Startup costs are going to be pretty high for starlink terminals that use essentially a brand new form of communication, and Spacex will probably eat most of that instead of passing on to the early adopters. Remember $1000 VCRs? That won't fly for many starlink users.

The real interesting thing to watch for starlink is how its capacity compares to more traditional RF solutions. Capacity is really the major limiting factor (relative to ground solutions) for all space applications to date. That's driven in a practical sense by MPT constraints of course, but the result is always a much higher cost/performance relative to terrestrial alternatives. Sat phones, sat internet, even sat TV (though less so than the others)...for instance, if you have sat internet now, you have terrible speed and a monthly volume cap. Even services like O3B and Oneweb are/will be capacity constrained to the point where those who live in areas with mature terrestrial solutions won't benefit from switching to the constellation services. How much more capacity will starlink have over RF? I'm not quite sure...

That all dovetails back into the philosophical--and ultimately financial--debate on space vs. terrestrial. If you live in a dense, first world area, its likely the existing and evolving terrestrial infrastructure is almost always going to be your solution...and so that limits the potential constellation customer base to the rest of the world (the "Other Three Billion", as it were) which, at least currently is also the least financially powerful customer base. That debate then leans into fixed vs constellation services, where the later spends a lot of time covering the 80% of the earth that has near zero demand. But...I suppose I'm getting way off topic now...

 

[Mike1080i]

7. For good measure, it is not a stretch to see that GPS/Glonass are rendered irrelevant when this is in place.

but can it really replace GPS in my cell phone if the receiver size needs to be laptop/small pizza box sized?

I don't think the Starlink satellites will have onboard atomic clocks needed to compete with GPS, Glonass or Galileo level precision, even if the antennas are small enough.

 

[bxr140]

I don't think the Starlink satellites will have onboard atomic clocks needed to compete with GPS, Glonass or Galileo level precision, even if the antennas are small enough.

Agree. They will, however, use GPS. No need to reinvent the wheel, especially if someone is giving you a really round wheel for free...

 

[dhanson865]

If you live in a dense, first world area, its likely the existing and evolving terrestrial infrastructure is almost always going to be your solution...and so that limits the potential constellation customer base to the rest of the world (the "Other Three Billion", as it were) which, at least currently is also the least financially powerful customer base. That debate then leans into fixed vs constellation services, where the later spends a lot of time covering the 80% of the earth that has near zero demand. But...I suppose I'm getting way off topic now...

I live in a dense enough area (metro area of 9 counties is just under 1 million full time residents, add a few hotel residents for vacation and business travel and I'm sure we have over 1 million in the area).

Here's what Starlink needs to provide to me in the US to get me off my cable provider.

* net neutrality
* either no data cap or a data cap high enough that I can't max it out watching streaming video non stop 24/7 say something like 5 TB.
* 100 Mb/s down or better**
* 50 MB/s up or better**
* 5 email accounts

all with good ping times and bandwidth to sites like

ping google.com 50 ms or less on average
ping teslamotorsclub.com 50 ms or less on average
ping amazon.com 50 ms or less on average
ping amd.com 50 ms or less on average
ping login-p3.worldoftanks.com 50 ms or less on average no more than 75ms at worst
ping youtube.com 50 ms or less on average no more than 200ms at worst

and similar for netflix, hulu, ps vue, and such.

And if they can do that for under $50 a month I'd switch ASAP, if it is under $75 a month I'd likely switch depending on bandwidth and latency.

** by 2019 or whenever I get the chance I'd expect DL/UL to be much faster and I won't complain if they advertise it as synchronous rates but in reality the vast majority need to DL faster than they UL. I have no problem with Gigabit down and 500 Megabit up or any other high speed asynchronous ratio so long as the bandwidth is faster than my cable company's mid level offering that is sub $100. My local cable provider has upload speeds 1/10th the advertised download rate. I'd like to see that be 1/4 or 1/2 on the cheapest package and save the 1/1 for the most expensive package.

 

[jbcarioca]

I don't think the Starlink satellites will have onboard atomic clocks needed to compete with GPS, Glonass or Galileo level precision, even if the antennas are small enough.

Probably true, but:

Agree. They will, however, use GPS. No need to reinvent the wheel, especially if someone is giving you a really round wheel for free...

and anyway, technologies have advanced so quickly that the solutions of the 1980's are not so relevant anymore. My point about relevancy is not just position but position coupled with vastly more information that is not available from the civil access of the three publicly accessible systems.

 

[ccutrer]

* 5 email accounts

People still use email accounts provided by their ISP? Do you not realize you’ll likely switch ISPs at some point in the future? And then you’ll have to annoy everyone you know with a “my email address changed!” notice.

(My other pet peeve, on a similar note - people that use employer provided email for personal use. Then find they’re locked out of their DirecTV account because they got a new job!)

 

[dhanson865]

People still use email accounts provided by their ISP? Do you not realize you’ll likely switch ISPs at some point in the future? And then you’ll have to annoy everyone you know with a “my email address changed!” notice.

(My other pet peeve, on a similar note - people that use employer provided email for personal use. Then find they’re locked out of their DirecTV account because they got a new job!)

I never give out my ISP email address to people. I use it for websites. I won't have to tell anyone, I'll just go and update my email address on a few sites as I go along. I don't want gmail notifying me every time I get a junk email from the last 10 places I bought something or from each of the monthly bills I pay online (your bill is almost due, your bill is due tomorrow, your bill is due today, you paid your bill, your statement is ready to view, your bill is due in two weeks, your bill is almost due again...).

I also want to have a even lower priority email account for ones I know will spam me but I don't want to have to make disposable email addresses over and over again. Maybe I want to search that account for a password reset or something.

Heck I never give out any email address to people. I give them my cell number if I want them to text me. I sure don't hand out email addresses to humans on any thing close to a regular basis.

I essentially have 3 email accounts for me (primary non ISP, primary ISP, secondary ISP) not counting anyone else in my household. Considering the average household size is ~2.5 people, I think 5 accounts is enough to cover 2 email addresses per average family member or 1 per for larger families up to the max of 5.

If you don't want to use the ISP mailserver you don't have to but it is a cheap service that to me serves a function. I'm sure others might prefer it as well (depending on their relationship with google, microsoft, or yahoo).

Relying on any one email account is a single point of failure. Look at people that have had their entire google account disabled in the blink of an eye with no appeal or recourse of any kind. Look at the way people are getting hacked because someone social engineers Tmobile or ATT or Verizon or your cable provider.

Heck I don't want you to have any idea what email address is being used for any of my 2 factor auths or password resets. I want the ability to have so many email providers and accounts you have no chance of disabling half my life by taking over a single account that has all my mail going to it.

Every email account and mailserver I'm on is another attack vector but it is a trade off in that the worst case scenario is lessened. I'll trade having to manage multiple accounts to have more resilience should one company screw up or should one account be hacked.

 

[wnorris]

A reminder that Spacex has a launch today at ~9-10 EST. The first stage for this mission was previously used will not be recovered. Also, and more importantly, this launch will see SpaceX's two test Starlink satellites deployed.

PAZ Mission | SpaceX

 

[Starno]

It's happening :

 

[ecarfan]

And to try to steer this thread back on topic...

Elon just tweeted: “Today’s Falcon launch carries 2 SpaceX test satellites for global broadband. If successful, Starlink constellation will serve least served.”

I like that phrase, “will serve least served.” The goal is global broadband access. I envision small groups of low income people who have no net access buying a single receiver and a router and sharing the connection.