NASASpaceFlight.com Forum

SpaceX Vehicles and Missions => SpaceX General Section => Topic started by: speedevil on 01/13/2015 01:55 pm

Title: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 01/13/2015 01:55 pm
This is just a summary of Starlink information to get it all in one place...

NSF Threads for Starlink:
SpaceX FCC filing for a 4425 satellite constellation providing Internet service (https://forum.nasaspaceflight.com/index.php?topic=41634.0) (original thread for discussion of the FCC filings)
SpaceX - now a satellite manufacturer (Starlink) (https://forum.nasaspaceflight.com/index.php?topic=36552.0) (general Starlink discussion)
SpaceX New FCC Filings for Starlink (https://forum.nasaspaceflight.com/index.php?topic=46726.0) (use this thread going forward for discussion specifically related to the FCC filings)

SAT-LOA-20161115-00118 (http://licensing.fcc.gov/cgi-bin/ws.exe/prod/ib/forms/reports/swr031b.hts?q_set=V_SITE_ANTENNA_FREQ.file_numberC/File+Number/%3D/SATLOA2016111500118&prepare=&column=V_SITE_ANTENNA_FREQ.file_numberC/File+Number)
SAT-LOA-20170301-00027 (http://licensing.fcc.gov/cgi-bin/ws.exe/prod/ib/forms/reports/swr031b.hts?q_set=V_SITE_ANTENNA_FREQ.file_numberC/File+Number/%3D/SATLOA2017030100027&prepare=&column=V_SITE_ANTENNA_FREQ.file_numberC/File+Number)
These are the Starlink Ku/Ka-band filings.  Approved March 29, 2018 (https://licensing.fcc.gov/myibfs/download.do?attachment_key=1364689)

SAT-LOA-20170726-00110 (http://licensing.fcc.gov/cgi-bin/ws.exe/prod/ib/forms/reports/swr031b.hts?q_set=V_SITE_ANTENNA_FREQ.file_numberC/File+Number/%3D/SATLOA2017072600110&prepare=&column=V_SITE_ANTENNA_FREQ.file_numberC/File+Number)
This is the Starlink V-band filing.  It has 7000 additional satellites in a VLEO constellation and adds V-band frequencies to the original 4425 satellite Ku/Ka-band constellation.  Adopted November 15, 2018 (https://docs.fcc.gov/public/attachments/FCC-18-161A1.pdf) (document released November 19, 2018.)

SAT-MOD-20181108-00083 (http://licensing.fcc.gov/cgi-bin/ws.exe/prod/ib/forms/reports/swr031b.hts?q_set=V_SITE_ANTENNA_FREQ.file_numberC/File+Number/%3D/SATMOD2018110800083&prepare=&column=V_SITE_ANTENNA_FREQ.file_numberC/File+Number)
This modification would allow the initial deployment (~1600 sats) of the Ku/Ka-band constellation to be deployed in a lower orbit.  The filing has not been officially accepted/posted for public notice yet as of Nov. 23, 2018.  Once the notice of the filing is posted it will kick off a (30 day?) time period for comments/oppositions to the filing, followed by more time periods for rounds of responses.  Don't expect any resolution on this before mid-2019.



http://www.businessweek.com/videos/2015-01-13/musk-says-spacex-will-develop-satellites-in-seattle

So this is where the funding for Mars comes from?

Discussed on the above - the ULA competition lawsuit.
Landing failure (not adding anything to the 'more hydraulic fluid').
Cost plus, governmental revolving door, ...
The Boeing corruption issues, ...

17:37 - 'We're creating an engineering centre in Seattle. Big announcement on Friday. We're going to try to do for satellites what we've done for rockets. Several hundred people eventually, starting with 50 or 60, perhaps 3 or 4 years until 1000.'
q: 'People have lost billions on satellite development' a: 'We might join them! But satellite technology is not very advanced, they are designed to be space proven, and if you start the design at proven technology, you're designing with old technology, which means you're launching with 5-10 year old technology'.
'Small satellites that may or may not work, with technologies a decade or two more advanced and the ability to relaunch rapidly may change the paradigm'.
(paraphrased)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: S.Paulissen on 01/13/2015 02:44 pm
Unexpected, but not surprising.  Lots of nuance in those words.  Good luck SpX.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/13/2015 03:09 pm
Potentially a much bigger market than space launch. (Even bigger than shuttling tens of thousands of people to Mars and back every year or so... Though not as awesome :) .)

It also gives them something to fill the potential capacity that a full RLV would have, something that's a real challenge.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 01/13/2015 03:16 pm
http://www.businessweek.com/videos/2015-01-13/musk-says-spacex-will-develop-satellites-in-seattle

17:37 - 'We're creating an engineering centre in Seattle. Big announcement on Friday. We're going to try to do for satellites what we've done for rockets. Several hundred people eventually, starting with 50 or 60, perhaps 3 or 4 years until 1000.'
(paraphrased)


Flexible, automated mass production will be their keys to the kingdom, and 1000 people designing and building satellites is a ton of personnel if they are trying to change the paradigm of satellite construction- unless these assemblers are paid $12 an hour and they want to launch thousands of them....  For comparison, back in 2009 Tesla was churning out around 600 roadsters a year on average with fewer than 900 employees.  Musk's previous comments about this venture were that he envisioned 700 satellites total.

Even some McGuffin technology or manufacturing shift that allows them to replace a few large expensive satellites with lots and lots of inexpensive satellites cheaply launched will have to conform to the ability of a radio (or laser) signal to send and receive information at a distance, so we're not talking iPhones in space here.  There will be definite boundaries on satellite sizes and complexity because of the physical realities of long-distance communications (that we discussed extensively here (http://forum.nasaspaceflight.com/index.php?topic=36047.msg1289923#msg1289923)); this affects build and launch price per unit and the number of units they will launch even if all other factors are ignored for simplicity.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Chris Bergin on 01/13/2015 03:38 pm
Yeah, that is interesting indeed!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/13/2015 03:44 pm
Of course, SpaceX has the advantage of a large RLV, initially a partial RLV (of still quite good payload capability) but by the time this is in full swing, a very large fully reusable launch vehicle which would otherwise have a hard time finding enough customers (currently only 36 or so payloads per year... They'll need 40-100 in order to make full RLV worthwhile).

SpaceX's constellation need not be made of smallsats, they could be as big as typical GSO sats, say 3-5 metric tons. Even with 700 satellites, that's still only 70-100 partly reusable Falcon Heavy (or some future RLV of similar or greater capacity) launches to put that whole thing up, just barely enough to justify a full RLV for a couple years.

So basically, SpaceX will have an enormous advantage over those of the past in that it'll have access to ultra-cheap launch, two orders of magnitude cheaper per kilogram to LEO (by the time their full RLV gets built) than, say, Iridium or ORBCOMM had in the 1990s. (Falcon Heavy is already 1/25th the cost per kilogram that Pegasus XL was in the 1990s.)

Plus they already have some in-house satellite building experience from Dragon.

It makes a lot of sense. Full RLVs don't make sense given the current size of the launch market, so just increase the size of the launch market.

No guarantee of success, of course.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jakusb on 01/13/2015 03:50 pm
I guess they need satellites anyway to get to and survive on Mars. Why not revolutionize the market while doing your own development.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/13/2015 03:51 pm
Personally I think this is awesome, and it would be cool to see if Space X and Elon can do for the satellite industry what they've done for the space launch industry (and Elon and Tesla for the electric car industry, and Elon and Paypal for the payment processing industry, etc, etc).

However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/13/2015 03:55 pm
Personally I think this is awesome, and it would be cool to see if Space X and Elon can do for the satellite industry what they've done for the space launch industry (and Elon and Tesla for the electric car industry, and Elon and Paypal for the payment processing industry, etc, etc).

However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?
Has it opened up Boeing to such lawsuits? Boeing developed Delta IV and builds commercial satellites. As far as I'm aware, that wasn't really a problem (though they did have other problems).

It's not like SpaceX is going to turn away satellite customers, and they certainly haven't so far. (EDIT: They were trying to compete for the Cygnus launch and also were in talks to launch Dreamchaser--SpaceX seemed happy to launch it, from what I heard.) They need all the launches they can get, long-term.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/13/2015 03:56 pm
Personally I think this is awesome, and it would be cool to see if Space X and Elon can do for the satellite industry what they've done for the space launch industry (and Elon and Tesla for the electric car industry, and Elon and Paypal for the payment processing industry, etc, etc).

However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?
Has it opened up Boeing to such lawsuits? Boeing developed Delta IV and builds commercial satellites. As far as I'm aware, that wasn't really a problem (though they did have other problems).

It's not like SpaceX is going to turn away satellite customers, and they certainly haven't so far. They need all the launches they can get, long-term.
Fair point, thank you for your response. This forum is a great place to learn :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RedLineTrain on 01/13/2015 03:57 pm
Looks like their experience with a stake in SSTL was encouraging.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 01/13/2015 04:00 pm
However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?

Nah, as long as there is one other vendor for similar services (doesn't even have to be satellite-based), the FCC doesn't care, and hasn't since Clinton was in office.  This specific argument was settled 15 years ago, though I'm happy Google is cracking it back open.

http://www.extremetech.com/internet/196675-google-calls-on-fcc-to-mandate-line-sharing-pits-itself-directly-against-comcast-and-other-isps
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: SoulWager on 01/13/2015 04:11 pm
Hopefully this is very bad news for current ISPs.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/13/2015 04:12 pm
However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?
I don't think that's a concern initially. Where that tends to become a problem is if you abuse a dominant position, it's not a priori illegal just to have a dominant position, and I don't really think SpaceX has a dominant position right now.

Nor is it a priori illegal to vertically integrate (in the business sense) and use the resulting efficiencies to compete more effectively.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/13/2015 04:13 pm
However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?

Nah, as long as there is one other vendor for similar services (doesn't even have to be satellite-based), the FCC doesn't care, and hasn't since Clinton was in office.  This specific argument was settled 15 years ago, though I'm happy Google is cracking it back open.

http://www.extremetech.com/internet/196675-google-calls-on-fcc-to-mandate-line-sharing-pits-itself-directly-against-comcast-and-other-isps
Makes sense, and the article was interesting reading, thank you :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hotblack Desiato on 01/13/2015 04:16 pm
maybe they can offer a satellite-barebone, and the customer can equip it with his own special hardware.

does anyone know how much radiation-shielding it takes to use standard electronic hardware instead of special space-hardened electronic hardware? as far as I know, the astronauts on ISS use regular notebooks and cameras (of course high end equipment, since the price is irrelevant, only mass matters). it can't be that much.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/13/2015 04:17 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Galactic Penguin SST on 01/13/2015 04:21 pm
I wonder what that means for those already rather "innovative" vendors like SSTL.  ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/13/2015 04:21 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.
They must already need this because of Dragon (with Cygnus, Dreamchaser, and CST-100 as competing spacecraft that may have been launched on F9), correct?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/13/2015 04:43 pm


However, and I'm far from an expert here, wouldn't this open them up to anti-competition lawsuits?

AIUI, part of Orbital Sciences rationale for developing Antares was to provide a Delta II replacement to improve the economics of launching their existing sat business.

Going purely by analogy, ISTM this new announcement shouldn't cause any issues on this front.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 01/13/2015 05:26 pm
Of course, SpaceX has the advantage of a large RLV, initially a partial RLV (of still quite good payload capability) but by the time this is in full swing, a very large fully reusable launch vehicle which would otherwise have a hard time finding enough customers (currently only 36 or so payloads per year... They'll need 40-100 in order to make full RLV worthwhile).

Even without a RLV - they can put experimental payloads into space almost utterly free - as ride-a-longs.

'Ok guys - I want ten designs for the star tracker on my desk by next month', we'll launch in 6 weeks and see if any of them work'.

A minimal communications package and solar array would let you get some use out of testing at almost any orbit.
If you can get to the point where you can point even a small HGA or laser - the game changes rather.
'random' tiny birds in near-GTO orbits might almost make a fun constellation.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 01/13/2015 05:52 pm
Half of this page should be in a "laptops in space" thread. The joke was funny, but the comment chain after that doesn't discuss anything about the given radio interview. Don't get lazy people, make a new thread or topic stick.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Billium on 01/13/2015 05:55 pm
I wonder if this will be a joint venture with MDA Corp (location of new office would make sense). I wonder if this will hurt Spacex's relationship with existing satellite manufacturers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/13/2015 05:59 pm
So back when this was discussed in November on a thread titled "Elon Musk eyeing partnership to launch 700 internet satellites" I wrote about some aspects of this plan that are real game changers:

Quote
1. Instead of today's comms satellite manufacturing we will have a new generation and at least one new manufacturer ready to work for others building much cheaper devices. They are as much designing the manufacturing facility as the satellite in one go. Rather than building a dozen or two devices based on one common 'bus' there will be 700, they will need to be able to make spares in the future, so they will have the facility to cheaply manufacture others of similar design for other customers. I believe that Musk and Wyler will be able to build and launch the last of their birds, and the eventual spares, for < $500,000 per bird or < $1M when you include idea #2

and

Quote
2. My pet idea of having a tender with several spares on each plane that takes the dead ones away. With 700 'birds' these guys will likely be replacing 1-2 per week after a while. Once all the spares in a plane have been replaced, deorbit the duds and tender and send up a new tender.

But see http://forum.nasaspaceflight.com/index.php?topic=36047.msg1286989#msg1286989 (http://forum.nasaspaceflight.com/index.php?topic=36047.msg1286989#msg1286989)  and
for http://forum.nasaspaceflight.com/index.php?topic=36047.msg1285547#msg1285547 (http://forum.nasaspaceflight.com/index.php?topic=36047.msg1285547#msg1285547) for the thoughts on how it makes a perfect early anchor tenant for F9R's to give them an established track record.

but I think the biggest game changer out there is the fact that 5 years from now, this new satellite factory will be churning out 100's of birds a year for other customers.

Elon, if you are listening, make sure the standard satellite buses all accommodate de-orbit disposal burns AND that your constellations have active tenders (equipped with spares of course) on each plane nominally to replace failed units but also to trash collect in case one fails so that it can't deorbit itself.  Lets have a commitment on this venture that sets the bar to prevent Kessler disasters.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 01/13/2015 06:09 pm
Elon, if you are listening, make sure the standard satellite buses all accommodate de-orbit disposal burns AND that your constellations have active tenders (equipped with spares of course) on each plane nominally to replace failed units but also to trash collect in case one fails so that it can't deorbit itself.  Lets have a commitment on this venture that sets the bar to prevent Kessler disasters.
Leetle tugs - perhaps with ion engines - could be really, really useful in mars orbit and general space assembly too, so dual win.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/13/2015 06:10 pm
After PayPal, Tesla, SpaceX, Giga factory for Solar energy, Giga Battery Plant, now Elon is going straight into the Silicon Valley adventure, planting a seed in Seattle.
What is going on with Elon? He is looking like a rocket escaped from the Earth gravity and picking up speed... faster and faster into the Solar System.
Godspeed, Elon...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/13/2015 06:29 pm
ISTM this virtually guarantees involvement in electric propulsion.

And MSNW is in Redmond. Just sayin....
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/13/2015 06:33 pm
Back to Satellites, here's a direct link to the Bloomberg interview....

Satellite comments start  ~17:25

Audio interview....(MP4-21 min) (http://cdn3.videos.bloomberg.com/m/NjI2OTkxNw/hiWcRKWUlvjURRlI0HIcn84qDtAqXbt75pxyssgemxAxNjEw/cfe571da-491a-4550-a52a-f061dba11a6c_240.mp4)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: yg1968 on 01/13/2015 06:51 pm
I am assuming that this relates to the internet satellites news of a couple of months ago:
http://www.theverge.com/2014/11/11/7192173/satellite-elon-musk-spacex
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/13/2015 08:49 pm
The funny part is how this affect the plans of competitors wrt market outlook.

If ULA/ESA have been (behind closed doors of course) evaluating the prospects of doing an RLV of some sort, then they have been analyzing market potential to see how it can possibly support the development cost.

Except whatever projections they've been making are a lot less certain now. New quantities/prices?  New competitor?

Also, how much is Musk cooperating with existing satellite operators, who are his customers?  Is SES (for example) already interested in what he has to offer?  If yes, how long has he been talking to them about it?

His customers are the operators, not the manufacturers - so he's basically staying out of their way.  He can offer them both the satellites and the launch service, and *all* they have to do is concentrate on the service.  Sort of like buying a Boeing satellite and a Boeing launch.    Except without Boeing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hotblack Desiato on 01/13/2015 08:52 pm
Half of this page should be in a "laptops in space" thread. The joke was funny, but the comment chain after that doesn't discuss anything about the given radio interview. Don't get lazy people, make a new thread or topic stick.

http://forum.nasaspaceflight.com/index.php?topic=36557.0

opened....
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 01/13/2015 09:02 pm
However, SpaceX earns a considerable amount of Comnsat money, launching the very things SpaceX has actively announced plans to compete against.

What short-to-mid term ramifications is that going to earn SpaceX? I can imagine possible negative feedback.

Good luck SpaceX, regardless. I'm sure Elon's already outguessed any critiques we could summon on that front, however.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pippin on 01/13/2015 09:10 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.
Hm, is it usually the satellite vendor who's buying the launch or the operator?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/13/2015 09:18 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.
Hm, is it usually the satellite vendor who's buying the launch or the operator?

The operator.  Musk knows better than to compete with his customers.

He's doing what many here predicted - he's going to support large satellite constellations from both ends (platforms and launch), but he's not going to get into the derived businesses.

He's probably had enough talks with customers and VCs that want to do communication, or imaging, or god knows what else with satellite constellations, and he's figured there's enough business there to start a whole new venture.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pippin on 01/13/2015 09:22 pm
Or he's finding that his RLV doesn't have enough payload capacity to compete in the "long duration, very heavy" segment and that his FH is too expensive so he's trying to push other market segments where his launch vehicle works better.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: WindnWar on 01/13/2015 09:23 pm
However, SpaceX earns a considerable amount of Comnsat money, launching the very things SpaceX has actively announced plans to compete against.

What short-to-mid term ramifications is that going to earn SpaceX? I can imagine possible negative feedback.

Good luck SpaceX, regardless. I'm sure Elon's already outguessed any critiques we could summon on that front, however.

Like others have said, he deals with the operators, so it should not cause any negative feedback, and the 700 satellite thing isn't going to make all those other comsats in orbit obsolete. Its simply a different service all together, and years away at that. I think given the costs of buying sats currently, the operators might be very interested in testing his approach. It might not work but for the right price I can see someone out there trying it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: NovaSilisko on 01/13/2015 09:25 pm
25% off launches with included satellites! Limited time offer, this weekend only!  :P
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/13/2015 09:25 pm
Wait, who says SpaceX wouldn't be operating this? It fits with their strategy of climbing the entire value chain ladder, starting with basically raw metal. If SpaceX is going to be building a megaconstellation, then unless Friday brings news of cooperation with Google or something (which is definitely a big possibility), they'll have to operate it themselves.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pippin on 01/13/2015 09:25 pm
I'm a bit sceptic about how many business models there are for large constellations. Imaging is something already well covered by others using very small, very cheap sats in very low orbits. Navigation is a given, too, because it's essentially available for free.
Which leaves global networks based on LEO constellations, certainly a market and probably one that can take a new entrant. Much more cyclic than Elon's traditional government service business.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 01/13/2015 09:28 pm
We're going to try to do for satellites what we've done for rockets.

Maybe Elon should actually do what he promised for rockets before venturing into new territory.

Also I'm not sure what's "old technology" in satellites. They have improved tremendously in the last decades. Power, life expectancy, number of transponders etc.

In contrast to the launch industry there is not shortage of innovative satellite manufacturers out there.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/13/2015 10:03 pm
I'm a bit sceptic about how many business models there are for large constellations. Imaging is something already well covered by others using very small, very cheap sats in very low orbits. Navigation is a given, too, because it's essentially available for free.
Which leaves global networks based on LEO constellations, certainly a market and probably one that can take a new entrant. Much more cyclic than Elon's traditional government service business.

I don't think imaging is even close to being saturated.  Just consider real-time imaging and what it implies.

And we're very far from a high quality global communication network with any appreciable bandwidth.

But I doubt you get people to invest in an entire satellite-building venture if you can't present a convincing case that it will be kept busy.


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pippin on 01/13/2015 10:07 pm
I'm not saying it's saturated but there's at least half a dozen imaging constellations being worked on. And that's not "old tech". You won't see a few hundred constellations, these things have network effects, too.

Now, global communication networks, I agree there's a lot going to come here.

But that's two business models. My main point was that I don't see too many additional ones for whole constellations. Of course, you often don't see business models coming but still, there's usually little _really_ new stuff coming.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/13/2015 10:12 pm
For 1000 people and 700 satellites and the timescales Elon has indicated, it works out at about 5 man-years per satellite, so less than $500,000 in direct labor costs. Remarkably little, component and infrastructure costs will be much more than that - large clean rooms will be required to manufacture 700 satellites in a few years.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/13/2015 10:13 pm
I'm not saying it's saturated but there's at least half a dozen imaging constellations being worked on. And that's not "old tech". You won't see a few hundred constellations, these things have network effects, too.

Now, global communication networks, I agree there's a lot going to come here.

But that's two business models. My main point was that I don't see too many additional ones for whole constellations. Of course, you often don't see business models coming but still, there's usually little _really_ new stuff coming.

Yes, they are worked on, but each is trying to re-invent the wheel.  That market will consolidate, and I don't think Musk is too late for that party.  Besides, he's got as good a view into that industry as any of them.

If he again stays out of the way of the operators, and simply offers a low cost constellation-optimized platform, he may not need to compete with them business wise.

Real time imaging constellations are also a giant communication challenge, which just adds to the potential.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: jak Kennedy on 01/13/2015 10:14 pm
Hopefully Elon will be able to tell the DoD that he doesn't have any free slots for their birds till 2025. "That's a negative AF space command, The pattern is full"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pippin on 01/13/2015 10:15 pm
He won't be able to stay out of the business of the operators when we talk imaging since most of the new entrant operators there are planning (or doing it) to build their own sats, vertically integrating themselves.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/13/2015 10:26 pm
He won't be able to stay out of the business of the operators when we talk imaging since most of the new entrant operators there are planning (or doing it) to build their own sats, vertically integrating themselves.

How much of that is set in stone already?   I hear a lot of talk, and some doves are flying, but are they getting enough bandwidth and longevity out of them?   I'm sure they all came to Musk when looking for launchers, and I wonder if he just saw 10 groups trying to re-invent the same wheel - and maybe made them the offer for their "real" constellation?

A real-time imager moving at 8 km/sec has to snap a 250 m frame 32 times a second, for a resolution of 10 cm at a frame size of 2500 pixels - for no overlap.   That's a hell of a lot of communication bandwidth, and thus a lot of power.  I don't know if the current micro-sats can do all that.

If you're talking real time video surveillance, it gets even worse.

So I don't think the current offerings are the final word by any stretch of the imagination.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/13/2015 11:06 pm
We need to wait until Friday for the official announcement and see what Elon really intends to do: is he going to enter the market in a triple quality of sats producer, launcher and operator or is he stlll into some kind of cooperation with Greg Wyler?

Evidently, he's believable when talking about manufacturing sats, even of yet unknown generation technology, or talking about launching them.

As operator, I wouldn't bet though. Even if price wise the sat could be very competitive, going through multiple tier chain operation, the product could end up dead. At the same time, as he put it so many times, he is already too busy with Tesla, SpaceX, Giga Battery Plant a.s.o., so I wouldn't foresee any interest going into operation field.

He may also have some corp interested to heavily invest into this new venture - such as Panasonic with Battery Plant in Nevada.

Friday could be the day of big news...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Avron on 01/13/2015 11:15 pm
sounds like this with mean a second site for engineers working at Spacex.. and will bring us global wifi that much quicker.. can wait to get rid of my monopolies
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hotblack Desiato on 01/13/2015 11:45 pm
We need to wait until Friday for the official announcement and see what Elon really intends to do: is he going to enter the market in a triple quality of sats producer, launcher and operator or is he stlll into some kind of cooperation with Greg Wyler?

[...]

Friday could be the day of big news...

connected to the fact, that the falcon heavy thread is locked without any obvious reason, there might be something going on (smelling a conspiracy theory and whistling the x-files theme).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/13/2015 11:46 pm
Here's a transcript of the Bloomberg interview.

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

By the end his words-per-minute were making my fingers sting. I think he's excited about the new satellite office. :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/14/2015 12:17 am
Takeaway: he's a big believer in lots of smaller, more frequently replaced and more advanced satellites vs. (as he put it) the Battlestar Galactica model.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 01/14/2015 12:36 am
Here's a transcript of the Bloomberg interview.

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

By the end his words-per-minute were making my fingers sting. I think he's excited about the new satellite office. :)
Enormous thanks to you! Somehow Elon is the person whose english I understand the worst of all people I ever heard speaking.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/14/2015 12:46 am
Here's a transcript of the Bloomberg interview.

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

By the end his words-per-minute were making my fingers sting. I think he's excited about the new satellite office. :)
IMO it's hard not to get excited about his dreams. Thank you for posting the transcript.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mr. mark on 01/14/2015 03:00 am
Wondering if SpaceX will get in the ion/electric propulsion game for satellite boost like Boeing is doing. The technology could be helpful as a stepping stone for deep space.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mme on 01/14/2015 05:34 am
Wondering if SpaceX will get in the ion/electric propulsion game for satellite boost like Boeing is doing. The technology could be helpful as a stepping stone for deep space.
I could be wrong, but I would think that "small" satellites would target lower orbits.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Lars-J on 01/14/2015 05:51 am
It certainly should not be that surprising that SpaceX is entering this field:
1. They already have experience now in building spacecraft (Dragon), even if not for long missions (yet)
2. This is technology they need to master if they are serious about Mars.

So why not?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hotblack Desiato on 01/14/2015 07:04 am
It certainly should not be that surprising that SpaceX is entering this field:
1. They already have experience now in building spacecraft (Dragon), even if not for long missions (yet)
2. This is technology they need to master if they are serious about Mars.

So why not?

and think about it this way. cygnus is a spacecraft derived from a satellite system. spaceX just goes the same way, but the opposite direction.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/14/2015 07:11 am
I could be wrong, but I would think that "small" satellites would target lower orbits.
Well, satellite co-location in geostationary slots is possible.

Something I've picked up watching talks about this. There's trades that can be made with the satellite for things like coverage beams in different areas. You can have a certain amount of fixed beams, then some steerable beams (but those take more mass). For a satellite with a long service life you aren't quite sure where the demand is going to be so you keep the steerable beams in reserve.

Extend that to what Musk seems to be talking about. You'd have several satellites co-located, shorter lives but more frequent launches, able to take over for each other in case of failure. I don't know how small they'd be (potentially not that small), but you'd definitely want to get out of the situation where the payload is an order of magnitude more than the launch. So, a very aggressively cost optimized design, with lots of identical units built and launched to different orbital slots for different customers.

Imagine 20 orbital slots, each requiring a launch per year. That's a decent satellite business.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Galactic Penguin SST on 01/14/2015 12:31 pm
Note: I've moved a few recent posts not quite related to satellites to the General Falcon & Dragon Discussion thread.  ;)

I wonder what kind of satellites is Mr. Musk et al. planning to build - LEO/MEO communication sat constellation? Or Imaging satellites? Commercial navigation fleet that beats GPS? Bear in mind that satellites have had quite a bit of evolution ahead of rockets, so I am not quite sure how much "revolutionary space" is there for Mr. Musk et al. to blast into.  ;)

Also I wonder if they are focusing on the satellite bus (a.k.a. making real "Lego spacecraft") or the specific payloads, or both? Somehow I just can't see them start developing optical sensors, radar antenna systems or atomic clocks, unless they are needed for the Mars Doctrine (TM).....  ::)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 01/14/2015 02:06 pm
Here's a transcript of the Bloomberg interview.

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

By the end his words-per-minute were making my fingers sting. I think he's excited about the new satellite office. :)

Thanks for the digital sacrifice!
Much easier to extract detail, and quotes...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: wannamoonbase on 01/14/2015 02:47 pm
Here's a transcript of the Bloomberg interview.

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

By the end his words-per-minute were making my fingers sting. I think he's excited about the new satellite office. :)
Enormous thanks to you! Somehow Elon is the person whose english I understand the worst of all people I ever heard speaking.

He's not an eloquent speaker.  He is hard to listen too, but it's almost like you can see the ideas working inside his head and the reasoning behind what he is saying.  I find it fascinating. I can only imagine what his teachers and professors thought of the awkward brilliant kid when he was growing up.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 01/14/2015 03:41 pm
He's not an eloquent speaker.  He is hard to listen too, but it's almost like you can see the ideas working inside his head and the reasoning behind what he is saying.  I find it fascinating. I can only imagine what his teachers and professors thought of the awkward brilliant kid when he was growing up.
An eloquent speaker would have said 'There will be an interesting announcement on that at 9AM Friday'.

:)
His excitement really does come through, as does his lack of filter, which does good/bad.
'I think tesla is a bit overpriced at the moment' -> 5% stock loss.

If spaceX is doing lots of satellite busses - that raises all sorts of fun issues.

For coplanar deployments - a moderate sized ion tug possibly has interesting aspects.

They already have some rendevous experience with Dragon - is refuelling and module swapout sane?
I should really not be lazy and look up the cost of a co-planar phase change in GEO over say a month.

Certainly orbit managment of failed sats seems reasonable.




Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/14/2015 03:58 pm
He is not eloquent because he's not delivering rehearsed bullet points, but simply speaking his mind based on personal knowledge and opinion.

That's not common on TV, and so it stands out.  I find it sometimes difficult to pick out some of the words, but no problem understanding exactly what he means.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Callezetter on 01/14/2015 04:05 pm
Its also what makes Mr Musk so darn compelling to root for. He usually cuts right through all the corporate crap we are so used too. It brings out the childhood dreamer in the best of us. And thats a very powerful talent, indeed.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mr. mark on 01/14/2015 04:18 pm
I think Musk's main talent is for disrupting current technologies and seeing new ways forward. The satellite business is just another example of that. I'm still thinking that Tesla and the emerging electric car market had something to do with lowering oil prices by the Saudi's. He's one big disrupter, Elon is.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: macpacheco on 01/14/2015 04:26 pm
Mr Musk's way of speaking wreaks of honesty. Plenty of times he went out of his way to explain risks and downsides to what he's doing (instead of selling only the positives). After his amazing accomplishments, his way of speaking is one reason I'm an avid fan of everything he does.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cosmonautdjp on 01/14/2015 07:11 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/14/2015 07:19 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"
They can't just do something totally custom to streamline for their own payloads to the exclusion of others.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/14/2015 07:24 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/14/2015 07:31 pm
Boeing makes satellites for different companies, as well as a launcher - how is this any different?   Of course there are measures they have to take to compartmentalize the customers
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jdeshetler on 01/14/2015 07:32 pm
There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

We sometime called it the "Chinese Wall".

http://en.wikipedia.org/wiki/Chinese_wall
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Herb Schaltegger on 01/14/2015 07:42 pm
Boeing makes satellites for different companies, as well as a launcher - how is this any different?   Of course there are measures they have to take to compartmentalize the customers

Boeing historically is the reason for the very existence of so many laws and rules against horizontal integration in an industry. Look up the history of United Airlines, the various United Technologies companies, and Boeing Airplane Company sometime. ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sublimemarsupial on 01/14/2015 08:26 pm

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/14/2015 09:07 pm
I think Musk's main talent is for disrupting current technologies and seeing new ways forward. The satellite business is just another example of that. I'm still thinking that Tesla and the emerging electric car market had something to do with lowering oil prices by the Saudi's. He's one big disrupter, Elon is.
If your assumption is correct, and it may very well be, then it is safe to say Elon is going to cause a real industrial revolution, something similar to what the engine meant for the civilization
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 01/14/2015 09:38 pm
I think Musk's main talent is for disrupting current technologies and seeing new ways forward. The satellite business is just another example of that. I'm still thinking that Tesla and the emerging electric car market had something to do with lowering oil prices by the Saudi's. He's one big disrupter, Elon is.
If your assumption is correct, and it may very well be, then it is safe to say Elon is going to cause a real industrial revolution, something similar to what the engine meant for the civilization

His talent is so vast it frequently takes threads off topic in various fora
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/14/2015 09:46 pm
I think Musk's main talent is for disrupting current technologies and seeing new ways forward. The satellite business is just another example of that. I'm still thinking that Tesla and the emerging electric car market had something to do with lowering oil prices by the Saudi's. He's one big disrupter, Elon is.
If your assumption is correct, and it may very well be, then it is safe to say Elon is going to cause a real industrial revolution, something similar to what the engine meant for the civilization
... And still have 20 minutes left before lunchtime!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ronsmytheiii on 01/14/2015 09:58 pm
Going to start off by showing just how expensive satellites are relative to LV's thanks to someone who knows both, Dr Antonio Elias of Orbital Sciences:

Back 10-15 years ago, the cost of launch averaged anywhere between 50% to 100% of the cost of the spacecraft.  Indeed, a classical "Mission Costs Rule of Thumb" used to say there are five nearly-equal pieces of the pie: Spacecraft Bus, Instrument(s), Launch, Ground Equipment and the Systems Engineering/Program Management glue that hold it together.

That is no longer true.

Today, ratios of 10:1 in the cost of spacecraft:cost of launch are not uncommon; two examples: the first SBRS GEO spacecraft will cost over $2B, and ride on a $175M EELV.  The first Minotaur IV, priced around $50M, will launch the SBSS spaceraft that  cost over $500M. (http://www.space.com/spacenews/archive07/sbss_0611.html)

I don't know the cost of the TACSAT-3 spacecraft, but it probably was several times what Orbital got for that launch!!!

Corollary 1: while given two identical rockets with identical reliability and performance customers will prefer the cheaper one ("relative price elasticity" is high) a FREE rocket will not increase the demand for rockets more than 10-20%, given the cost of the rest of the mission ("absolute price elasticity" is very low.)

Corollary 2: customers will gladly pay a significant premium on the launch service if it is PERCEIVED to increase its reliability in any significant amount.

These are two lessons we learned the hard way.

http://forum.nasaspaceflight.com/index.php?topic=16645.60

So now even with a reduction in launch prices, the number of launches wont increase that much.  Plus, with up to 10:1 spacecraft to LV cost ratio, they do look to be VERY profitable (I know Elon says he just wants to go to Mars, but his investors would appreciate a new profitable venture)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ChrisWilson68 on 01/14/2015 10:38 pm

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.

Of course it's possible, they just decide to make sure it's not the same engineers when proprietary data of another company is involved.  Companies do this all the time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/15/2015 12:12 am

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.

No,that is the point, they can't
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: gospacex on 01/15/2015 12:14 am
Musk already needs to have about 50% of sat building expertise in order to make Dragon work. Dragon _is_ a satellite (albeit it is not designed to work autonomously for years).

He probably thought "why not try to leverage that?".
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Lars-J on 01/15/2015 12:26 am

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.

I doubt many (if any) of the launch vehicle engineers will do that. SpaceX now has 4000+ employees. It's no longer garage start up where everyone is all-hands on for every project. People have job responsibilities in specific areas.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 01/15/2015 12:28 am
Going to start off by showing just how expensive satellites are relative to LV's thanks to someone who knows both, Dr Antonio Elias of Orbital Sciences:

Back 10-15 years ago, the cost of launch averaged anywhere between 50% to 100% of the cost of the spacecraft.  Indeed, a classical "Mission Costs Rule of Thumb" used to say there are five nearly-equal pieces of the pie: Spacecraft Bus, Instrument(s), Launch, Ground Equipment and the Systems Engineering/Program Management glue that hold it together.

That is no longer true.

Today, ratios of 10:1 in the cost of spacecraft:cost of launch are not uncommon; two examples: the first SBRS GEO spacecraft will cost over $2B, and ride on a $175M EELV.  The first Minotaur IV, priced around $50M, will launch the SBSS spaceraft that  cost over $500M. (http://www.space.com/spacenews/archive07/sbss_0611.html)

I don't know the cost of the TACSAT-3 spacecraft, but it probably was several times what Orbital got for that launch!!!

Corollary 1: while given two identical rockets with identical reliability and performance customers will prefer the cheaper one ("relative price elasticity" is high) a FREE rocket will not increase the demand for rockets more than 10-20%, given the cost of the rest of the mission ("absolute price elasticity" is very low.)

Corollary 2: customers will gladly pay a significant premium on the launch service if it is PERCEIVED to increase its reliability in any significant amount.

These are two lessons we learned the hard way.

http://forum.nasaspaceflight.com/index.php?topic=16645.60

So now even with a reduction in launch prices, the number of launches wont increase that much.  Plus, with up to 10:1 spacecraft to LV cost ratio, they do look to be VERY profitable (I know Elon says he just wants to go to Mars, but his investors would appreciate a new profitable venture)

By cutting launch prices in half or more, SpaceX nicely grabbed a big chunk of the market.  Intentions appear to do same for sats... Maybe with a similar drop in prices and a push on the technology, the market will finally expand -- or at least provide another revenue stream for bigger things.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/15/2015 03:26 am
So now even with a reduction in launch prices, the number of launches wont increase that much.
I don't think that's necessarily true. It depends on how things are structured. If SpaceX can manufacture satellites inexpensively enough and structure the contracts correctly to protect customers if SpaceX doesn't deliver, that addresses all those reasons.

An example might be SpaceX building a satellite to collocate in an existing slot, to free up some of its beams. The customer wants to use it for the next 5 years until the next "battlestar galactica" is launched. The contract is structured such that ownership will be transferred after the satellite is in orbit and had some time for shakedown - so they don't care if the launch fails. And without the satellite they'll have to turn away some customers for the next 5 years, but that's not the end of the world.

Seems like this is a situation where you could easily have more launches. And if it works, eventually you stop needing the "battlestar galactica" launches, you just have a constant churn of new satellites being launched and old ones going into disposal orbits - meaning upgrades and changes can be done incrementally and there's less exposure to any one satellite failing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sublimemarsupial on 01/15/2015 05:09 am

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.

I doubt many (if any) of the launch vehicle engineers will do that. SpaceX now has 4000+ employees. It's no longer garage start up where everyone is all-hands on for every project. People have job responsibilities in specific areas.

Its exactly because people have jobs in specific areas that will necessite the same engineers doing both. Do you think Elon is going to be gung-ho about have two solar panel designers in house doing very similar work completely isolated from each other? Or two avionics box designers? Or do flight software engineers? SpaceX is all about collaboration and doing more with less, so I have no idea how you expect Elon to be okay with duplicating a huge percentage of his engineering expertise...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ChrisWilson68 on 01/15/2015 05:21 am

There just has to be firewalls between the launch vehicle and spacecraft business, if they are to launch other companies' spacecraft and other companies' are going to launch their spacecraft.

What do you mean by "firewalls?"

They can't share other companies' information internally.

This will be impossible, as many of the same engineers will be working the launch vehicle side and the satellite side.

I doubt many (if any) of the launch vehicle engineers will do that. SpaceX now has 4000+ employees. It's no longer garage start up where everyone is all-hands on for every project. People have job responsibilities in specific areas.

Its exactly because people have jobs in specific areas that will necessite the same engineers doing both. Do you think Elon is going to be gung-ho about have two solar panel designers in house doing very similar work completely isolated from each other? Or two avionics box designers? Or do flight software engineers? SpaceX is all about collaboration and doing more with less, so I have no idea how you expect Elon to be okay with duplicating a huge percentage of his engineering expertise...

SpaceX doesn't just have one avionics designer.  They have a whole team.  Actually, they surely have multiple teams.  Same for any other area of expertise.  They'd have to hire a bunch of new people if they have more work anyway.  It's not as if they have a bunch of people sitting around only working half time right now.  They'll need to hire about the same number of people either way.

Yes, there will be some cases where it will be inconvenient to have a wall and they'll have to duplicate some effort, but it's not likely to be a huge amount.

Remember, the wall is between their satellite construction business and certain parts of the launch vehicle business.  The most overlap is likely to be with Dragon, and there's no need for a wall between satellite construction and Dragon.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 01/15/2015 10:41 am
I've worked in a office where there was a specially build 'office within an office' with secured entry etc. Just to keep that software development separate from the mainline software development, on behalf of the customer who wanted to keep things very secret. So it's not that difficult to achieve. There are clearly some points of contact (interfaces between mainline and project) but they were standardised. I would expect the same to apply here.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sheltonjr on 01/15/2015 01:17 pm
If SpaceX could get rid of the clean room requirement, That could save a lot of money on building satellites.

For those in the know, Could you explain why satellites need to be assembled in a clean room environment. I can understand why some sensors need to be built in a clean room, but once they are closed up is a clean room still required.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/15/2015 01:50 pm
If SpaceX could get rid of the clean room requirement, That could save a lot of money on building satellites.

For those in the know, Could you explain why satellites need to be assembled in a clean room environment. I can understand why some sensors need to be built in a clean room, but once they are closed up is a clean room still required.

"closed up"?  Spacecraft are never really closed up.  There are many items on a spacecraft that are contamination sensitive.  Solar arrays and radiators are subject to particulate and Non Volatile Residue (NVR) contamination.  NVR is material that coats a surface and heat does not get rid it.  These tend to reduce the efficiency of these items.  Particulate and NVR contamination would cause issues for star trackers, sun sensors, and such causing them to make false readings or unable to lock on, etc. And if the spacecraft has optical instruments, contamination would degrade their effectiveness.  And if only these items are kept clean vs the whole spacecraft, once on orbit there would be cross contamination due to the zero g and vacuum.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/15/2015 04:12 pm
This 100% true, but that does not mean it is set in stone. IF you really wanted to remove the cleanroom environment (I don't know that anyone tried, or that it's even a major cost issue) you can try some other approaches.

Take your radiators for example.  Can they be coated with an engineered layer that is designed to sublimate once it hits vacuum?  Let the terrestrial muck accumulate on that, and then once it is in orbit, over a few days, the muck will become loose.  You might then get secondary contamination, but perhaps that's at a level that's already ok.  (Some contamination occurs even in clear rooms, right?)

Perhaps some units can be simply covered while stowed.  (And these of course add cost, etc)

The point is, many of these practices go back decades, and I don't know that anyone re-visited them lately to see if they can be circumvented.  Especially when taking a holistic view of the entire satellite architecture.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 01/15/2015 04:21 pm
Yes, I can't really understand those criticism to Musk read on this thread. He is an Entrepreneur and could as well invest in toys or furnitures companies to make money. But he is investing in a high-tech, (still relatively) futuristic field like satellites.

Even an idiot would understand that more/cheaper satellites <-> more rockets reqd.

So, what?

Let's see if he succeed also in this. If he will not, free room for criticism...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/15/2015 04:26 pm
Yes, I can't really understand those criticism to Musk read on this thread. He is an Entrepreneur and could as well invest in toys or furnitures companies to make money. But he is investing in a high-tech, (still relatively) futuristic field like satellites.

Even an idiot would understand that more/cheaper satellites <-> more rockets reqd.

So, what?

Let's see if he succeed also in this. If he will not, free room for criticism...
More/cheaper satellites only equals more rockets required if there's somebody in need of more/cheaper satellites. You can't just produce infinitely more of a product and reduce the price and expect the market to absorb all of that. If a company started producing a billion cars a year and priced them at $1,000 I would still only need two cars for my family.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/15/2015 04:33 pm

1.  Take you radiators for example.  Can they be coated with an engineered layer that is designed to sublimate once it hits vacuum?  Let the terrestrial muck accumulate on that, and then once it is in orbit, over a few days, the muck will become loose.  You might then get secondary contamination, but perhaps that's at a level that's already ok.  (Some contamination occurs even in clear rooms, right?)

2.  Perhaps some units can be simply covered while stowed.  (And these of course add cost, etc)

3. The point is, many of these practices go back decades, and I don't know that anyone re-visited them lately to see if they can be circumvented.  Especially when taking a holistic view of the entire satellite architecture.

1.  No.  The muck would just move to a different part of the spacecraft and the "engineered layer " would then coat cooler parts of the spacecraft.  It would go in mucking up joints, optics, MLI, and other important items.

2.  Some are, some can't, some cost too much to cover.

3.  The holistic view of the entire satellite architecture is what has driven to these practices. 

Gloves are worn to prevent body oil from being transferred to items.  It off gasses in the vacuum and discolors from UV radiation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/15/2015 04:55 pm

1.  Take you radiators for example.  Can they be coated with an engineered layer that is designed to sublimate once it hits vacuum?  Let the terrestrial muck accumulate on that, and then once it is in orbit, over a few days, the muck will become loose.  You might then get secondary contamination, but perhaps that's at a level that's already ok.  (Some contamination occurs even in clear rooms, right?)

2.  Perhaps some units can be simply covered while stowed.  (And these of course add cost, etc)

3. The point is, many of these practices go back decades, and I don't know that anyone re-visited them lately to see if they can be circumvented.  Especially when taking a holistic view of the entire satellite architecture.

1.  No.  The muck would just move to a different part of the spacecraft and the "engineered layer " would then coat cooler parts of the spacecraft.  It would go in mucking up joints, optics, MLI, and other important items.

2.  Some are, some can't, some cost too much to cover.

3.  The holistic view of the entire satellite architecture is what has driven to these practices. 

Gloves are worn to prevent body oil from being transferred to items.  It off gasses in the vacuum and discolors from UV radiation.

Yes, I know that, I've designed both clean room mechanisms and UHV mechanisms.

And true, these practices are not a result of people being masochistic...  But it's a common phenomena in many industries that "best practices" are established for some reason, and 40 years later things have changes and the reason can be eliminated, but the best practices are part of the global know-how database.

As for the sublimating layer, if it is thick enough to last a while, the muck will have several cycles of being unable to stick to anything that matters, and every time it is knocked loose, most of it will escape, less will remain.  Anyhow, it was just an example. I don't think clean rooms are that big of a deal.
 
Joints in vacuum are a big deal of course, and you really want to eliminate them.  There are several new phased array concepts that are maturing nowawdays that might be able to help.  Star trackers can probably take advantage of the insane progress there has been in imaging devices, and become 100% solid state (if they aren't already)

---

From an article referenced on the other satellite thread:  "According to the ITU, Wyler only has until 2019 to get the network operating, or else the spectrum rights could be granted to another company. "

Assuming the Wyler is Musk's "anchor customer", they probably have some concept already pretty mature, from the Goog days.  If they're starting completely from scratch, 4 years is not a lot of time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 01/15/2015 05:03 pm
More/cheaper satellites only equals more rockets required if there's somebody in need of more/cheaper satellites. You can't just produce infinitely more of a product and reduce the price and expect the market to absorb all of that. If a company started producing a billion cars a year and priced them at $1,000 I would still only need two cars for my family.

Never heard about the law of supply and demand?

Of course there are also inelastic products, but those are typically related to very basic needs, like food or water. Certainly not satellites, that could compete with a pretty large number of "ground products" as their cost decrease.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mme on 01/15/2015 05:30 pm
Yes, I can't really understand those criticism to Musk read on this thread. He is an Entrepreneur and could as well invest in toys or furnitures companies to make money. But he is investing in a high-tech, (still relatively) futuristic field like satellites.

Even an idiot would understand that more/cheaper satellites <-> more rockets reqd.

So, what?

Let's see if he succeed also in this. If he will not, free room for criticism...
More/cheaper satellites only equals more rockets required if there's somebody in need of more/cheaper satellites. You can't just produce infinitely more of a product and reduce the price and expect the market to absorb all of that. If a company started producing a billion cars a year and priced them at $1,000 I would still only need two cars for my family.
Have we forgotten that Greg Wyler wants a 700 satellite constellation (http://www.cnet.com/news/next-up-for-elon-musk-launching-satellites-into-space/)?
Quote
Musk is working with former Google executive Greg Wyler on figuring out a way to manufacture and launch a fleet of as many as 700 satellites that would weigh less than 250 pounds each, according to a report from The Wall Street Journal Friday. SpaceX, which Musk founded in 2002 and is now a leading private spaceflight company based in California, would likely launch the satellites into space, the report added.

So maybe there is already some demand.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/15/2015 05:34 pm

1.  Joints in vacuum are a big deal of course, and you really want to eliminate them.  There are several new phased array concepts that are maturing nowawdays that might be able to help. 

2. Star trackers can probably take advantage of the insane progress there has been in imaging devices, and become 100% solid state (if they aren't already)


1.  there are other mechanisms than antenna.

2.  They have and are just as susceptible to debris and NVR.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 01/15/2015 05:43 pm
So maybe there is already some demand.

It is difficult to imagine (I mean without some expertise and a dedicated study) what this market could become if costs would decrease by a factor - we say - 10. If this would go to 100, I guess it would open up entire new market sectors like drugs, new chemicals, manufacturing (ultra-high precision manufacturing?).

At present, there is no one in the world more than Elon Musk to appreciate this. Others will follow. It's part of the practical and flexible approach to business typical of the more recent generation of entrepreneurs to fully appreciate one cannot just sit down when something is starting to work and make profits. As a "bayesian", I like this, exactly as I dislike hedgehogs (http://www.chforum.org/library/choice12.shtml) in any field.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ericspittle on 01/15/2015 06:04 pm
More/cheaper satellites only equals more rockets required if there's somebody in need of more/cheaper satellites. You can't just produce infinitely more of a product and reduce the price and expect the market to absorb all of that. If a company started producing a billion cars a year and priced them at $1,000 I would still only need two cars for my family.

Never heard about the law of supply and demand?

Of course there are also inelastic products, but those are typically related to very basic needs, like food or water. Certainly not satellites, that could compete with a pretty large number of "ground products" as their cost decrease.
Supply has virtually no effect on demand (other than deciding which competitor gets more demand in a competitive market) in a saturated market, and I don't think it can be confirmed that the satellite market is or isn't saturated at this point.

Obviously there's a point to be made that using cheaper processes, or inventing markets that don't previously exist (ie: worldwide satellite internet coverage), can lead to either taking market share from competitors if the market is, in fact, saturated or just creating a demand that didn't previously exist. However it's difficult to argue that either thing is an absolute. In the recent interview that brought up this entire topic he was asked "Why are you optimistic about satellites? People have lost billions." He responded: "Yeah, we might join them" (source: http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14). If he isn't confident that they can make money doing this neither am I. His idea for small replaceable satellites may or may not work as well as he hopes and if it does there may or may not end up being customers, but at some point the money has to come from somewhere to fund continued R&D.

Personally I think he can do it, he isn't exactly known for having his lofty dreams fail, but it's still far from a sure thing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 01/15/2015 06:17 pm
Personally I think he can do it, he isn't exactly known for having his lofty dreams fail, but it's still far from a sure thing.

Would you agree that is at least more likely that the satellite market is not saturated than the opposite, and therefore elastic? Note that who lost money on this market did it in an almost cost-constant condition. I can't really believe a satellite market with substantial decrease in costs, would not see a demand increase.

It's funny but I share a different point of view on his capabilities. I have a feeling he is an incredibly visionary person, but I got the impression he likes to start things much more than to complete them, if there is nothing qualitatively new challenging him.

Just a feeling, I may be wrong.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mr. mark on 01/15/2015 06:23 pm
Looks like Branson wants to go head to head with Musk. http://www.virgin.com/richard-branson/creating-the-worlds-largest-ever-satellite-constellation-0 Different business model but possibly same result. Interesting that Virgin released this one day before SpaceX.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 01/15/2015 06:32 pm
Not sure if the figures add up, but make satellite and launch cheap enough, and the time spent in orbit per satellite can be decreased, meaning a lot of the points about cross contamination mentioned above MAY not be so relevant. Negating or reducing the need for clean rooms, hence making the satellites cheap enough to be replaced more easily.

In addition, continual reductions in the size of devices means satellite and the clean rooms they may require can be a lot smaller.

Not relevant to optical stuff, still needs to be very clean.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 01/15/2015 09:06 pm
I think ultimately it boils down to whether a satellite with an expected lifetime of, say, 5 years is significantly cheaper to build than one with an expected lifetime of 15 years. I'm sceptical.

Its good to see though that SpaceX is trying to build a market for its upcoming RLV, because that's what is really missing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 01/15/2015 09:11 pm
Not sure if the figures add up, but make satellite and launch cheap enough, and the time spent in orbit per satellite can be decreased, meaning a lot of the points about cross contamination mentioned above MAY not be so relevant. Negating or reducing the need for clean rooms, hence making the satellites cheap enough to be replaced more easily.

In addition, continual reductions in the size of devices means satellite and the clean rooms they may require can be a lot smaller.

Not relevant to optical stuff, still needs to be very clean.

I concur. At the end of the day the need to use increasingly expensive tools and procedures to produce something is related to the tread-off between

(1) cost of failure
(2) cost of the tool.

Of course there are many other factors included. However, the satellites market is very peculiar as the cost of a single failure is overwhelmingly high, also because maintenance and repair is virtually impossible.

However, there must be a minimum satellite production cost such that not using clean room could become convenient, although I guess that threshold is still very far even respect to the most optimistic future situation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 01/15/2015 11:10 pm
I think ultimately it boils down to whether a satellite with an expected lifetime of, say, 5 years is significantly cheaper to build than one with an expected lifetime of 15 years. I'm sceptical.

Its good to see though that SpaceX is trying to build a market for its upcoming RLV, because that's what is really missing.

Though I agree that this action builds the launch market, the Internet global market dwarfs launch plus sat market combined. Seems that is some huge low-hanging fruit if you can launch at will for lowest cost on the planet... And have the bulk of sat technology already in hand. Revenue, for Mars or whatever, is a powerful incentive.

This business strategy may fund EM vision even if launch market is slow to respond.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/15/2015 11:49 pm
Russian/Soviet satellites were/are no doubt very inexpensive but had around a 3-5 year lifetime. They weren't built with space rated components, instead relying on a pressurized container. Russia is starting to modernize it's space hardware, so this isn't as true as it once was. But there is something to be said that lower lifetime spacecraft can be cheaper.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 01/16/2015 02:01 pm
This idea of global WiFi-like coverage just blows my mind...

So people in Afrika will be able to get Internet by paying in bitcoins or alternative internet currencies making internet "cheap for all" and "free for all".

But what about govs of Russia and China, which have antisatellite weapons? I see the queues of lobbists from local cellular companies holding suitcases full of cash :) trying to persuade govs how important it is to shut down/strike down those constellation for national security (of someone's deep pockets).

Now the fun part of it: if Elon can launch satellite 10x cheaper than hostile gov can launch anti-satellite weapon, this adds new dimension to "economic wars" - "Commander, they are launching faster than we can strike them down". So according to Musk's wishes global demands for rocket launches increases twofold! :)

Is it cheap to jam the satellite WiFi in the whole country? Or, may be just jamming in major cities (90% population of Russia, 60% of China) would be considered enough to protect the wealth of cellular operators?

Wish Elon success to stir that hornet nest.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/16/2015 07:32 pm
So are we still expecting a formal announcement today? 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Saltvann on 01/16/2015 07:38 pm
So are we still expecting a formal announcement today?

Yep, but I think it's on the Pacific Time schedule. Still a while yet to go as it's only 12:38 PT.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/16/2015 07:55 pm
So are we still expecting a formal announcement today?

Yep, but I think it's on the Pacific Time schedule. Still a while yet to go as it's only 12:38 PT.

Geekwire http://www.geekwire.com/2015/photos-spacex-dragon-capsule-delivered-seattle-center-private-event-elon-musk/ (http://www.geekwire.com/2015/photos-spacex-dragon-capsule-delivered-seattle-center-private-event-elon-musk/) says "this evening"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ronsmytheiii on 01/17/2015 01:34 am
Quote
@seattlelibby: #ElonMusk has landed @SpaceX Seattle event! http://t.co/AJIFhYMYx6

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 01/17/2015 01:44 am
SpaceX is competing with Wyler/Oneweb, VG on an identical system. But EM wants a more ambitious system. OneWeb has spectrum already.

10 billion dollars. Ouch.

http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2 (http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2)

Might be a bridge too far, IMO.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 01:51 am
SpaceX is competing with Wyler/Oneweb, VG on an identical system. But EM wants a more ambitious system. OneWeb has spectrum already.

10 billion dollars. Ouch.

http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2 (http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2)

Might be bridge too far, IMO.
...a sufficiently large constellation of sufficient capability (not talking about mild adaptation of existing tech but orders of magnitude greater capability) with sufficient spectrum would have annual revenue in the hundreds of billions of dollars.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: funkyjive on 01/17/2015 02:26 am
Dragon in Seattle. Couldn't sneak in :(
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 02:30 am
...
Might be a bridge too far, IMO.
A Mars city of a million people and transport there and back for half a million is just fine, but Internet from a LEO constellation is a bridge too far... ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 02:38 am
Talking about 4025 satellites in the constellation, each a few hundred kilograms (but capability of much larger satellites) at around 750km altitude. (Robotbeat: Sounds about right to me... roughly 50 reusable FH flights or so) Using internally produced Hall thrusters. Won't be until about 5 years for the first version, 12-15 years for full capability. Produced more like cars (than current spacecraft), can afford to lose a few in a big constellation (compares mindset of a bunch of PCs in a room versus a couple mainframes), have to assume that competitors will improve as time goes on (unlike Teledesic). Will enable building Mars city (which will be expensive, though hopes to get transport down to a lot less than half a million dollars).

Paraphrased from: https://twitter.com/malderi/status/556256655428435968
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 03:03 am
From: http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p1
Quote
While there are major similarities between the two ventures [Greg's OneWeb constellation and SpaceX's constellation], Musk says he’ll have an edge through SpaceX’s smarts and manufacturing techniques. “Greg and I have a fundamental disagreement about the architecture,” Musk says. “We want a satellite that is an order of magnitude more sophisticated than what Greg wants. I think there should be two competing systems.”

I think Musk is on the right track, here. If you want to build a Mars city, you need the revenue to do that. I don't really see anything that could credibly bring in sufficient revenue other than an enormous constellation like Musk is describing, 4000 satellites instead of 700 plus the order of magnitude greater capability per satellite. Remember, Musk is partially looking for this as a way to monetize the technology advance of a large RLV based on Raptor. A 700 satellite constellation with 350kg per satellite (and OneWeb could easily be thinking much smaller than that!) weighs only 250 tons, it'd be launched in about 6-8 Falcon Heavies. Not really leveraging the capability that a large RLV gives you if you're only launching a few times. The larger 4000 satellite constellation, however, would require 50 or so reusable Falcon Heavies. And if you're talking about refreshing the constellation every 3-5 years, you're talking about a dozen extra reusable Falcon Heavy launches per year (or a mini BFR), which is starting to seriously take advantage of having a large RLV at your disposal. (That's on top of other launches.)

...Not only is this building the revenue needed for a Mars colony, it's also building the high-volume technology needed for truly getting the cost of MCTs down enough to be viable. And the high launch rate needed is proving out your RLV technology, too.


...in some ways, Musk doesn't have much choice. There are NOT that many ways of making a lot of cash in space (talking about tens of billions of dollars of revenue), and there won't be for at least another decade or so. Telecomms is pretty much the only way to do that, and a big enough constellation could bring in plenty to pay for a Mars colony.


(there are, of course, many ways this could fail)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Coastal Ron on 01/17/2015 03:14 am
SpaceX is competing with Wyler/Oneweb, VG on an identical system. But EM wants a more ambitious system. OneWeb has spectrum already.

10 billion dollars. Ouch.

http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2 (http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2)

Might be a bridge too far, IMO.

Qualcomm is also an investor with Virgin on OneWeb.  Qualcomm has experience in the satellite communications business from Globalstar, which they were partners with when it started up in the 90's.  Globalstar didn't work out, as cell phone networks grew too fast, so it will be interesting to see what they see in OneWeb.

I wonder if they got the spectrum for OneWeb from Qualcomm?  And where is Musk going to get spectrum from?

Funny how they announced so close together, but it would be odd for Musk to not have known about possible spectrum issues, so maybe he's not targeting the same end user OneWeb is?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 03:22 am
I don't think Musk is being outrageous. Iridium's satellites cost (in real terms, today's money) about $7-8 million each with a mass of 700kg, and less than 100 have been produced. $10 billion for 4000 satellites works out to $2 million (only $500,000 for launch, assuming near today's Falcon Heavy prices or some improvement), which is not an unreasonable cost reduction if you assume they have about half the mass and you're making at very least 40 times as many of them (allowing you to invest in a more automated production line), leveraging in-house existing experience. Iridium also took about 5 years to build from when Motorola started work in 1993 to 1998 when the constellation was operational. And given that SpaceX will have the world's first (at least partial) RLV family (and a large one at that) by the time the constellation is filled out, they will have access to much, much lower launch costs than Iridium did.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 03:34 am
SpaceX also has the advantage of being privately held (and, as a growing company, probably has access to plenty of capital if they need it) whereas Iridium had to put on a whole bunch of debt in order to build out their network. It took a really long time for Iridium to get enough subscribers (they needed 600,000 to pay for everything including debt but had only about 55,000 at time of bankruptcy), although now Iridium has I think over 750,000.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 03:36 am
SpaceX is competing with Wyler/Oneweb, VG on an identical system. But EM wants a more ambitious system. OneWeb has spectrum already.

10 billion dollars. Ouch.

http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2 (http://www.businessweek.com/articles/2015-01-17/elon-musk-and-spacex-plan-a-space-internet#p2)

Might be a bridge too far, IMO.

Qualcomm is also an investor with Virgin on OneWeb.  Qualcomm has experience in the satellite communications business from Globalstar, which they were partners with when it started up in the 90's.  Globalstar didn't work out, as cell phone networks grew too fast, so it will be interesting to see what they see in OneWeb.

I wonder if they got the spectrum for OneWeb from Qualcomm?  And where is Musk going to get spectrum from?

Funny how they announced so close together, but it would be odd for Musk to not have known about possible spectrum issues, so maybe he's not targeting the same end user OneWeb is?
Globalstar went bankrupt, but the constellation is still up and a new company formed in 2006 which runs it. In that sense, it did indeed work out, just took longer than investors and creditors may have hoped.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 01/17/2015 03:41 am
Will these be launched from Vandenberg?. There was talk a year ago about a lot of launches from here.

Wonder where Google stand in all this?. They've not with Oneweb anymore. If they partner with Elon then funding shouldn't be a problem.

As for launching OneWeb satellites, Spacex is still cheapest LV provider and likely to get cheaper.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AJW on 01/17/2015 04:08 am
Could Musk be counting on picking up WorldVu's Ku spectrum rights that expire in 2019 if they can't get a system operational?   Even if WorldVu raises $3 Billion, maybe they can move pretty far on design and engineering, but launching 650 sats will eat through that cash after just a handful of launches.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/17/2015 05:13 am
SpaceX also has the advantage of being privately held (and, as a growing company, probably has access to plenty of capital if they need it) whereas Iridium had to put on a whole bunch of debt in order to build out their network. It took a really long time for Iridium to get enough subscribers (they needed 600,000 to pay for everything including debt but had only about 55,000 at time of bankruptcy), although now Iridium has I think over 750,000.
Also the way it would be structured would likely be with much of the spending happening late in the project. They'd want to keep the cash burn rate low while they're designing the basic pieces like the hall thrusters.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 05:24 am
How is worldVu going to launch competitively?  Will SpaceX sell them launches?

The demand for global internet coverage is obvious, and is only increasing.  You can't supply that with a handful of satellites, and you don't build an entire factory just to make 700 of them.  But with 4000, even if they live 5 years, you need to make 800 satellites each and every year.  Now we're talking.

Some folks whine about how "there isn't a market" or "the market is not elastic", and then others go and make it happen.

Oh, and a company I like might do good business with these:
http://www.tethers.com/tt.html



Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ggr on 01/17/2015 05:42 am
I don't think Musk is being outrageous. Iridium's satellites cost (in real terms, today's money) about $7-8 million each with a mass of 700kg, and less than 100 have been produced. $10 billion for 4000 satellites works out to $2 million (only $500,000 for launch, assuming near today's Falcon Heavy prices or some improvement), which is not an unreasonable cost reduction if you assume they have about half the mass and you're making at very least 40 times as many of them (allowing you to invest in a more automated production line), leveraging in-house existing experience. Iridium also took about 5 years to build from when Motorola started work in 1993 to 1998 when the constellation was operational. And given that SpaceX will have the world's first (at least partial) RLV family (and a large one at that) by the time the constellation is filled out, they will have access to much, much lower launch costs than Iridium did.

I don't have figures to hand, and it's too late in the evening for me to search for them, but Iridium satellites are HUGE! I'd believe 700kg for Globalstar satellites though.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 05:47 am
I already looked up the figures.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 05:53 am
From Twitter @malderi
Quote
Normal satellites, big, huge testing, very important. But with large constellation, can afford to have some not work.

We had long arguments about that in this forum in the past.  (How rapid launch of a large number of satellites is "impossible" since you need a checkout campaign before you launch each one.)   Glad to see they've put this issue front and center.  Duds in orbit are allowed.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Port on 01/17/2015 06:23 am
just wanted to note, that with a distance related latency of just [email protected] orbit this could really replace optical lines - even for latency-dependant things like onlinegames and such (which are okay even at about 75ms or so imho)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/17/2015 06:51 am
Also twitter from @malderi

"Musk: Have to pay attention to security. Bad if hacked, either by AI or people."
    - very important IMO, this would get round state censorship, so attack by states possible.

"Musk: biggest concern about success: important to assume that competitors get better, too. Teledesic didn't."
 - it will be interesting to see how they build a system that can be expanded to more bandwidth and increased capabilities. This is one of the biggest system design problems IMO, get it wrong and the system is doomed in the long term.

"Musk: Hopeful that we can structure this to work with different countries. Don't want China shooting out our satellites."
 - setting up a global sales network will take time, need financing and other options that are sensitive to local circumstances.

"Musk: Plan on using Hall thrusters. Easy to make, no real production difficulty. Doesn't make sense to outsource."
- good choice.

"Musk: Cheaper to have a bunch of PCs on racks than a few mainframes, this is the same idea."
- distributed aspect more similar to PCs on desks

"Musk: Don't see bandwidth as difficult issue. Space to ground has plenty of usable spectrum."
- doesn't explain how he expects to get permission to use that spectrum

"Musk: Going to cost a lot to build. Ten or fifteen billion dollars, or more. But revenues fund city on Mars."
- SpaceX does not have the resources to fund this internally
- Such a high investment, with significant chance of failure risks taking down SpaceX as well.

"Musk: Smaller satellites, few hundred kg, but capability of much larger satellites."
- my guess is optimum size is somewhere in 300-600 kg range, driven mainly by antenna size.

"Musk: (timeline?) In the past, I've been optimistic on schedule. Recalibrating. Envision version 1 in about 5 years."
- my translation - I'm aiming for 3 years, but I've been optimistic
- planning on multiple versions from the start

"Musk: Talking about around 4000 satellites. 4025 exactly in current design."
- that is probably 100s of launches, maybe 50/year.

"Musk: Talking mostly around 1100km level. Space debris not much of a problem there."
- with that number of satellites, they will have to take great care to dispose of them at end of life.

"Musk: 20-30ms latency everywhere on earth. Expand tech to Mars, not much fiber there yet."
- a unique selling point, lower latency for long distances than earth bound networks

"Musk: Developing world, but also options for people stuck with comcast."
- big problem in developing world is that it needs to be very cheap both in cost of equipment and charges. Larger antennae on satellites mean smaller cheaper ground stations.
- significant numbers in developed world have poor internet speeds and few prospects of significant improvement, not enough of a market on its own but still maybe $1B/year.

"Musk: More than half of global long haul comm over this system."
- this is a biggie, global long haul market is massive. Competes directly with GEO telecom sats.

"Musk: Long term, rebuilding Internet in space. 10% of local business traffic."
- another biggie, business traffic is concentrated in a far smaller number of customers than consumer, but generally needs higher quality of service and grade of service guarantees.

"Musk: Start by developing our own constellation. Comm, earth science, space science. Focus is global comm system."
- looks like long term they plan for more than just this initial 4025 satellite global internet system.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Halidon on 01/17/2015 07:10 am
Dragon in Seattle. Couldn't sneak in :(
Hey I was there too, hanging over the railing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 01/17/2015 07:55 am
How is worldVu going to launch competitively?  Will SpaceX sell them launches?

The demand for global internet coverage is obvious, and is only increasing.  You can't supply that with a handful of satellites, and you don't build an entire factory just to make 700 of them.  But with 4000, even if they live 5 years, you need to make 800 satellites each and every year.  Now we're talking.

Some folks whine about how "there isn't a market" or "the market is not elastic", and then others go and make it happen.

Oh, and a company I like might do good business with these:
http://www.tethers.com/tt.html
Still need something to deorbit dead satellites ie don' t deploy this tape. I can see a market for a fleet of cubesats that can attach themselves to dead satellites and deploy this deorbit tape.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Port on 01/17/2015 08:36 am

"Musk: Talking mostly around 1100km level. Space debris not much of a problem there."
- with that number of satellites, they will have to take great care to dispose of them at end of life.

"Musk: 20-30ms latency everywhere on earth. Expand tech to Mars, not much fiber there yet."
- a unique selling point, lower latency for long distances than earth bound networks


i dont see how this add's up, 1100km/c is not 20-30ms ONE-WAY
(it's 36,69ms for up and again for down, add transmission between the satalites..)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Barrie on 01/17/2015 09:18 am

"Musk: Talking mostly around 1100km level. Space debris not much of a problem there."
- with that number of satellites, they will have to take great care to dispose of them at end of life.

"Musk: 20-30ms latency everywhere on earth. Expand tech to Mars, not much fiber there yet."
- a unique selling point, lower latency for long distances than earth bound networks


i dont see how this add's up, 1100km/c is not 20-30ms ONE-WAY
(it's 36,69ms for up and again for down, add transmission between the satalites..)
I make it 3.37ms, so one of us has got the decimal point in the wrong place
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 01/17/2015 09:20 am

"Musk: Talking mostly around 1100km level. Space debris not much of a problem there."
- with that number of satellites, they will have to take great care to dispose of them at end of life.

"Musk: 20-30ms latency everywhere on earth. Expand tech to Mars, not much fiber there yet."
- a unique selling point, lower latency for long distances than earth bound networks


i dont see how this add's up, 1100km/c is not 20-30ms ONE-WAY
(it's 36,69ms for up and again for down, add transmission between the satalites..)
I calculated that 1100km altitude, so speed of light up and down is just 7ms, but around-the globe path of 20000+km is indeed ~80ms, but that cannot be better with a cable (cable also goes around the globe at about the same altitude, 6400 vs 7500).

Satellites may offer more direct channel than existing ocean cables, so could be even faster, than cable (unless you wire your cable right through the center of the Earth :)).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: rklaehn on 01/17/2015 09:39 am

"Musk: Talking mostly around 1100km level. Space debris not much of a problem there."
- with that number of satellites, they will have to take great care to dispose of them at end of life.

"Musk: 20-30ms latency everywhere on earth. Expand tech to Mars, not much fiber there yet."
- a unique selling point, lower latency for long distances than earth bound networks


i dont see how this add's up, 1100km/c is not 20-30ms ONE-WAY
(it's 36,69ms for up and again for down, add transmission between the satalites..)
I calculated that 1100km altitude, so speed of light up and down is just 7ms, but around-the globe path of 20000+km is indeed ~80ms, but that cannot be better with a cable (cable also goes around the globe at about the same altitude, 6400 vs 7500).

Satellites may offer more direct channel than existing ocean cables, so could be even faster, than cable (unless you wire your cable right through the center of the Earth :)).

For long distances, this will definitely be better than fiber. The speed of light in fiber optic cables is just 2/3 of the vacuum value, and as you say fiber optic cables often don't go the direct route.

There are also large bottlenecks in the earth fiber network. E.g.: a lot of traffic from europe to asia goes through egypt.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Port on 01/17/2015 10:33 am
yeah i got the decimal point wrong, so its 3,669ms or 3,7ms (used the correct value vor c not approximated :P)

also the longest distance possible is half-way around the globe or was that implied by the term "arround the globe"?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 01/17/2015 10:58 am
It does smell like a Google moonshot. I wonder if Project Loon is only an initial R&D phase for this.

 Relocation of the backbone to LEO is a naughty idea which potentially threatens to deprive governments of any control of information their population consumes or produces

Scary shit, ladies and gentlemen.

Remembering the recent involvement ($10M) of Musk in Beneficial AI research and Google involvement in AGI... Are they preparing the stage for the arrival of our Saviour the Digital Overlord?

"Don't be evil"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pospa on 01/17/2015 12:32 pm
Pictures and few comments from the event in Seattle center on Friday.
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

I like this one:
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pospa on 01/17/2015 12:36 pm
SpX new office is located in Redmond a few minutes down the highway from Microsoft’s Xbox campus and the Nintendo of America headquarters, and in the midst of numerous video-game startups. They are seeking permits to make extensive renovations of that building.
http://www.geekwire.com/2015/spacexs-new-seattle-area-office-redmond-elon-musk-visit-region-week/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 01:48 pm
So the two biggest non-tech points are;

- He does want to operate this. Will he partner with any of his telecom customers?  He's not in direct competition with what they do today, but still...
- As stated above, being the backbone is getting yourself into the  politics of data surveillance, on an international level.  Every government will demand back doors, some will demand filters, etc. I don't know that this has a solution. I also don't think any government can shoot down 4000 satellites at 1000 km.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Llian Rhydderch on 01/17/2015 03:04 pm
This idea of global WiFi-like coverage just blows my mind...

So people in Afrika will be able to get Internet by paying in bitcoins or alternative internet currencies making internet "cheap for all" and "free for all".

But what about govs of Russia and China, which have antisatellite weapons? I see the queues of lobbists from local cellular companies holding suitcases full of cash :) trying to persuade govs how important it is to shut down/strike down those constellation for national security (of someone's deep pockets).

Now the fun part of it: if Elon can launch satellite 10x cheaper than hostile gov can launch anti-satellite weapon, this adds new dimension to "economic wars" - "Commander, they are launching faster than we can strike them down". So according to Musk's wishes global demands for rocket launches increases twofold! :)

Is it cheap to jam the satellite WiFi in the whole country? Or, may be just jamming in major cities (90% population of Russia, 60% of China) would be considered enough to protect the wealth of cellular operators?

Wish Elon success to stir that hornet nest.

I think you make an important point here, hrissan.  When the economics of commsat launch become just a fraction of the cost of traditional anti-satellite technology, then something important has happened.  Just as low-cost integrated-circuit-based copiers, printers and computers made the traditional means of the suppressing the open flow of information in the 1970s-1990s, truly low-cost sats could have an aspect of doing the same thing.

Musk's new venture, and that of OneWeb/Wyler/Qualcomm/Branson, extending internet comm more widely, and potentially making global comm harder to stop/limit at national borders, could be a game changer in more ways than just opening up the high-volume/low-cost satellite market.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: llanitedave on 01/17/2015 03:14 pm
So the two biggest non-tech points are;

- He does want to operate this. Will he partner with any of his telecom customers?  He's not in direct competition with what they do today, but still...
- As stated above, being the backbone is getting yourself into the  politics of data surveillance, on an international level.  Every government will demand back doors, some will demand filters, etc. I don't know that this has a solution. I also don't think any government can shoot down 4000 satellites at 1000 km.

Musk should hire Edward Snowden as his back-door point man.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 03:20 pm
So the two biggest non-tech points are;

- He does want to operate this. Will he partner with any of his telecom customers?  He's not in direct competition with what they do today, but still...
- As stated above, being the backbone is getting yourself into the  politics of data surveillance, on an international level.  Every government will demand back doors, some will demand filters, etc. I don't know that this has a solution. I also don't think any government can shoot down 4000 satellites at 1000 km.

Musk should hire Edward Snowden as his back-door point man.
Just think of the list of demands governments will make:

Backdoors when the satellites are over their territory, inbound or outbound data.

Backdoors for connections that "belong" to their citizens

Same for "people of interest"

This stuff already exists if you're in the internet equipment business, except in this case Musk will be the single global go-to guy.

That's just nuts.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Semmel on 01/17/2015 03:25 pm
Dont get your hopes up that Elon Musk plays the hero on this one. If he wants global internet, he needs to have a global business infrastructure as well to sell bandwidth and hardware. Therefore, any business needs to cooperate with the local government and obey to the local law. I am not aware of any communications company that denies government backdoors or denies internet filters. It is equally unlikely that anyone tries to shoot down the satellites.

A much more interesting and relevant question is: where does the InterSatX (or what ever name it will have) company pays taxes? At the headquaters? At the location where the customer buys the service? No tax paying like Apple et al. ?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 01/17/2015 03:38 pm
So the two biggest non-tech points are;

- He does want to operate this. Will he partner with any of his telecom customers?  He's not in direct competition with what they do today, but still...
- As stated above, being the backbone is getting yourself into the  politics of data surveillance, on an international level.  Every government will demand back doors, some will demand filters, etc. I don't know that this has a solution. I also don't think any government can shoot down 4000 satellites at 1000 km.

Musk should hire Edward Snowden as his back-door point man.
Just think of the list of demands governments will make:

Backdoors when the satellites are over their territory, inbound or outbound data.

Backdoors for connections that "belong" to their citizens

Same for "people of interest"

This stuff already exists if you're in the internet equipment business, except in this case Musk will be the single global go-to guy.

That's just nuts.
Seriously, most likely govs will fight this on the ground - punishing the unlicensed users who dare to connect to the "cloud" directly. Local companies may be allowed to have a connection up, and share it on the ground if they agree to have a gov-mandated filter/traffic archiver, like it is now with cellular operators. Licenses will be issued as well.

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: starsilk on 01/17/2015 03:52 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/17/2015 03:52 pm

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.

That is nuts.  You are making assumptions that are not simply not based on reality.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/17/2015 03:53 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...

And that is also why duds are bad.  They can't be controlled.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 03:55 pm
So the two biggest non-tech points are;

- He does want to operate this. Will he partner with any of his telecom customers?  He's not in direct competition with what they do today, but still...
- As stated above, being the backbone is getting yourself into the  politics of data surveillance, on an international level.  Every government will demand back doors, some will demand filters, etc. I don't know that this has a solution. I also don't think any government can shoot down 4000 satellites at 1000 km.

Musk should hire Edward Snowden as his back-door point man.
Just think of the list of demands governments will make:

Backdoors when the satellites are over their territory, inbound or outbound data.

Backdoors for connections that "belong" to their citizens

Same for "people of interest"

This stuff already exists if you're in the internet equipment business, except in this case Musk will be the single global go-to guy.

That's just nuts.
Seriously, most likely govs will fight this on the ground - punishing the unlicensed users who dare to connect to the "cloud" directly. Local companies may be allowed to have a connection up, and share it on the ground if they agree to have a gov-mandated filter/traffic archiver, like it is now with cellular operators. Licenses will be issued as well.

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.
Yes, they'll want to push the first level filter to the sats.

And yes, the world is crazy.

And yes, there 0 chance they'll come out of this as the good guys. 

No guts, no glory, I guess.

Maybe that's why competition is good here.  Having more than one "bad guy" really helps.

On a holistic level, the time for an uber sat constellation is here from a tech and market point of view, so politics can only delay it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 03:57 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...

And that is also why duds are bad.  They can't be controlled.
Duds can be dead in terms of main functionality, but still be able to deploy a drag tether (even by default) if the main control computer is dead.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:07 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...
I know, but the debris cloud won't stay at 1000 km.  It is a MAD move, and MAD generally works.  It gives the sat operator some significant level of immunity, so they can negotiate. 

A long time before any ASAT shots, there will be agreements over access restrictions. 

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/17/2015 04:09 pm

Duds can be dead in terms of main functionality, but still be able to deploy a drag tether (even by default) if the main control computer is dead.

The main functionality isn't tested at the launch site, just the bus, which knocks holes in your premise about testing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:11 pm

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.

That is nuts.  You are making assumptions that are not simply not based on reality.
Jim, major telecom vendors have to build wiretapping backdoors into their equipment.

Law enforcement overreaches when using them.

Those are facts.

Whether preventive surveillance amr collection by governments is a good idea is a separate debate, not for this thread.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 04:12 pm
You could have a totally separate system that could be commanded to deploy a drag device, perhaps two per bird for redundancy. Or a tug.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:14 pm

Duds can be dead in terms of main functionality, but still be able to deploy a drag tether (even by default) if the main control computer is dead.

The main functionality isn't tested at the launch site, just the bus, which knocks holes in your premise about testing.
How so?   I always argued that you can reduce satellite cost by accepting a certain % of faults in orbit.  Musk apparently thinks the same.

I showed you how the orbital junk issue is mitigated.

So what holes?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/17/2015 04:21 pm

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.

That is nuts.  You are making assumptions that are not simply not based on reality.
Jim, major telecom vendors have to build wiretapping backdoors into their equipment.

Law enforcement overreaches when using them.

Those are facts.


It is not a fact that "demand to be able to analyse ALL satellite traffic"

1.  The demand is not a fact.  It doesn't ask for it now on all existing systems.
2.  The ability doesn't exist now for all existing systems
3,  And regardless of # 1 & 2, the capability to analyse existing satellite traffic doesn't exist.

It is also not a fact that NSA servers are directly connected to the infrastructure.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/17/2015 04:27 pm

1.  How so?   I always argued that you can reduce satellite cost by accepting a certain % of faults in orbit.  Musk apparently thinks the same.

2. I showed you how the orbital junk issue is mitigated.


1.  Your argument also included a lower orbit, which is not applicable to concept.   Anyways, it doesn't mean it is right. 

2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:28 pm

BTW I bet USA gov will be the first to demand to be able to analyse ALL satellite traffic, making the network much less optimal (instead of sending from the satellite above Mongolia to satellite above China, traffic will have to go through USA ground infrastructure). May be Elon will agree to orbit NSA servers on satellites directly? :)

The World is nuts.

That is nuts.  You are making assumptions that are not simply not based on reality.
Jim, major telecom vendors have to build wiretapping backdoors into their equipment.

Law enforcement overreaches when using them.

Those are facts.


It is not a fact that "demand to be able to analyse ALL satellite traffic"

1.  The demand is not a fact.  It doesn't ask for it now on all existing systems.
2.  The ability doesn't exist now for all existing systems
3,  And regardless of # 1 & 2, the capability to analyse existing satellite traffic doesn't exist.

It is also not a fact that NSA servers are directly connected to the infrastructure.
I don't want to derail the thread.  I'd just point out that hrissan's argument does not change even if it's 20% of traffic, and that "directly connected" is mostly meaningless in a digital world.

As an example, think about the ability to install an SQL query vs. a stored procedure.  Different levels of capability, but just two shades of gray between "no access" and "totally own it"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:34 pm

1.  How so?   I always argued that you can reduce satellite cost by accepting a certain % of faults in orbit.  Musk apparently thinks the same.

2. I showed you how the orbital junk issue is mitigated.


1.  Your argument also included a lower orbit, which is not applicable to concept.   Anyways, it doesn't mean it is right. 

2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
Actually, if you're going to say what my argument was, let me correct you.

Myself and others have said that:
- Musk is probably planning a constellation of thousands of satellites
- LV reusability will be part of the plan
- Much smaller and cheaper mass-produced satellites are possible
- No need for pre-launch checkout campaigns if you're willing to accept some % of failure in orbit
- Shorter lifespans will be part of the plan
- Very low orbit is possible, but has both upsides and downsides (The very words I used) and COLA is not one of the down sides because math says so.

For the record.

I think we did pretty well.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Lars-J on 01/17/2015 04:38 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...

Really? C'mon. Watched 'Gravity' a bit too much lately? People who throw around the Kessler effect/syndrome rarely seem to understand how ...spaceous... ;) space is.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:39 pm
2. I showed you how the orbital junk issue is mitigated.
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.

Drag tethers can be installed as a fail-safe mechanisms.  Every single satellite is built to drag down unless it comes up healthy.

You can even have satellites auto-deploy the drag device if they've been out-of-contact for 24 hours, for any reason.

You can also launch all satellites to an orbit that's just shy of the target orbit, and only have them boost up a few km if they're healthy.

---

Finally, even if there is a situation where the above steps somehow fail, you're now dealing with a percent-of-a-percent issue, and a single non responsive satellite can just be avoided.  Happens all the time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 04:49 pm
I also don't think any government can shoot down 4000 satellites at 1000 km.

Only four or five need to be shot down, the rest will take care of themselves with the chain reaction from the debris...

Really? C'mon. Watched 'Gravity' a bit too much lately? People who throw around the Kessler effect/syndrome rarely seem to understand how ...spaceous... ;) space is.
I never bothered with Gravity, but saw the idiotic space junk sequences.  The Kessler effect will not cause a 20 minute chain reaction that brings down all satellites.   But the fact that Hollywood needs to dramatize everything doesn't means there isn't a problem.

However - if you have a situation where satellite-debris collisions occur, say, once a year, generate more debris, increase the collision frequency, etc, then you're seeing a Kessler effect, and it is a possible scenario.

Space is spacious, but it's a constant.  An exponential, even with a very low coefficient, will always catch up with it.

And on second thought, since collisions are an NxN problem, it's an exponential in x^2.   I wouldn't dismiss it without disproving Kessler's math first.

(EDIT: For a single given target, the risk of collision is proportional to the density of debris.   But if there are more pieces out there, then there are more "single given targets", and so the risk of a collision occurring is proportional to the square of the density of debris.  This is somewhat mitigated by the fact that small debris hitting small debris will generate very small debris, which will decay faster - a little bit like the infamous nObjects limitation in Asteroids....    But once you have a quantity that increases in some proportion to its magnitude, you get an exponential, and that's what Kessler is talking about.)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Mongo62 on 01/17/2015 07:09 pm
I think that this statement from Elon is very interesting:

"Musk: Going to cost a lot to build. Ten or fifteen billion dollars, or more. But revenues fund city on Mars."

So Elon is now planning to actually build the Mars colony itself?

I had understood his business strategy was to provide large-scale, inexpensive transportation services to and from Mars, but allow other organizations (private, governmental, or both) to do the heavy lifting of actually constructing (and paying for) 'Mars City". This statement makes it sound like Elon has changed his mind, and now plans to at least partly pay for the Mars colony himself.

It is certainly true that if this "second Internet backbone" idea pays off, he will have the revenue stream to be able to do this himself. And he always struck me as somebody who is interested in money as a means to accomplish his goal of establishing a branch of human civilization on another planet, rather than simple greed.

But what changed? Did he discover that few if any other organizations with access to the amount of money needed to colonize Mars, are interested in spending it (throwing it away, from their point of view) on Mars colonization? Or maybe he decided that the money would go much further if it was spent "the SpaceX way", instead of through traditional "Oldspace" methods? SLS would be a huge counterexample of how NOT to spend money.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/17/2015 07:11 pm
But what changed?

Nothing changed. He's a Californian. They feel guilty about making money.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Mongo62 on 01/17/2015 07:25 pm
But what changed?

Nothing changed. He's a Californian. They feel guilty about making money.

Don't forget that he is also Canadian (Canadian mother, hence Canadian citizenship, and he also spent 7 years living in Canada from ages 17 to 24), so he has a predisposition toward socially beneficial efforts.

He actually has three concurrent citizenships (South Afican, Canadian and American), which I guess is rather unusual.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 07:28 pm
But what changed?

Nothing changed. He's a Californian. They feel guilty about making money.
Or he's, you know, an actual human being who realizes money is a means (to build a Mars city, hard to think of anything more awesome than that) not an end to itself.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/17/2015 07:42 pm
I think that this statement from Elon is very interesting:

"Musk: Going to cost a lot to build. Ten or fifteen billion dollars, or more. But revenues fund city on Mars."

So Elon is now planning to actually build the Mars colony itself?

I had understood his business strategy was to provide large-scale, inexpensive transportation services to and from Mars, but allow other organizations (private, governmental, or both) to do the heavy lifting of actually constructing (and paying for) 'Mars City". This statement makes it sound like Elon has changed his mind, and now plans to at least partly pay for the Mars colony himself.

It is certainly true that if this "second Internet backbone" idea pays off, he will have the revenue stream to be able to do this himself. And he always struck me as somebody who is interested in money as a means to accomplish his goal of establishing a branch of human civilization on another planet, rather than simple greed.

But what changed? Did he discover that few if any other organizations with access to the amount of money needed to colonize Mars, are interested in spending it (throwing it away, from their point of view) on Mars colonization? Or maybe he decided that the money would go much further if it was spent "the SpaceX way", instead of through traditional "Oldspace" methods? SLS would be a huge counterexample of how NOT to spend money.
Who would pay for it? SpaceX will only ever access a fraction of NASA's limited budget. NASA's budget is not likely to expand, and SpaceX is not likely to get anywhere near half of it even in crazy-world-optimism. So you're looking at maybe $2-3 billion if NASA does the unthinkable and pays SpaceX to do their Mars program. That's clearly not enough, especially if done to NASA's specifications for a Mars base.

I think Musk more and more realizes he has little choice but to expand the market himself. There is no one currently who is interested in spending the money to do a Mars city (Mars One is pretty much a joke, unfortunately... certainly not going to be funded enough to make a dent in a Mars city).

It's all well and good to propose that a big market will pop up from nowhere once space launch becomes cheap (thus paying for the cost of developing that big reusable rocket), but that's an article of faith and somewhat a miracle /unless/ you create it yourself. Same logic extends to a Mars city. There's no good market reason that would justify such an investment right now, but if you have the funds to build it, you can make it happen regardless.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Malderi on 01/17/2015 09:36 pm
I think that this statement from Elon is very interesting:

"Musk: Going to cost a lot to build. Ten or fifteen billion dollars, or more. But revenues fund city on Mars."

So Elon is now planning to actually build the Mars colony itself?

I had understood his business strategy was to provide large-scale, inexpensive transportation services to and from Mars, but allow other organizations (private, governmental, or both) to do the heavy lifting of actually constructing (and paying for) 'Mars City". This statement makes it sound like Elon has changed his mind, and now plans to at least partly pay for the Mars colony himself.

It is certainly true that if this "second Internet backbone" idea pays off, he will have the revenue stream to be able to do this himself. And he always struck me as somebody who is interested in money as a means to accomplish his goal of establishing a branch of human civilization on another planet, rather than simple greed.

But what changed? Did he discover that few if any other organizations with access to the amount of money needed to colonize Mars, are interested in spending it (throwing it away, from their point of view) on Mars colonization? Or maybe he decided that the money would go much further if it was spent "the SpaceX way", instead of through traditional "Oldspace" methods? SLS would be a huge counterexample of how NOT to spend money.

Just to clarify - and I'm the Malderi everyone is quoting from Twitter, I was there - this is absolutely what Elon is saying, but maybe I didn't get completely through on the twitter stream. Elon said a Mars city will be very expensive no matter what, but that the revenues from this global space-based Internet will be the thing that pays for it. After last night, no question about his plans - he's going to fund the Mars city "himself", at least a significant portion of it, through profits from building this space-based Internet.

When you think about it, it makes sense in a Elon-ish sort of way. He needs tens of billions of dollars, if not more. What's big enough to make that kind of money, that he already has a competitive advantage in? If you think that way, it's obvious that this is the only way to do it. Launching a hundred F9's every year won't do it, but selling Internet to half the world will.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 09:51 pm
Thanks for the quotes!

About Mars, I never bought this "I'll just do the rocketry work and let someone else have all the fun" bit.  I don't know which quotes lie at the basis of that, but it never made any sense to me.  It runs 100% opposite of anything Musk ever did.

He has the means. He has the will. He has the opportunity.  Plus, there's nobody else with his kind of drive to do it if he doesn't.

(And the "X" in "SpaceX" does not stand for "Transportation".)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RonM on 01/17/2015 10:08 pm
When you think about it, it makes sense in a Elon-ish sort of way. He needs tens of billions of dollars, if not more. What's big enough to make that kind of money, that he already has a competitive advantage in? If you think that way, it's obvious that this is the only way to do it. Launching a hundred F9's every year won't do it, but selling Internet to half the world will.

The big issue has always been money. Even though he is a billionaire, Elon doesn't have enough money. Who's going to pay for a colony on Mars? If SpaceX can provide 100+ F9 flights per year, that's great, but who's going to buy them?

Now we have the answer. Elon is going to create his own internet satellite business!

It's going to be tough, but if anyone can do it, it will be Mr. Musk.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: DanielW on 01/17/2015 10:47 pm
I certainly hope they succeed, but it will not be anything like easy. They won't be competing directly with their launch customers, but such a constellation would make many of them obsolete in the same way that broadband internet is killing cable and satellite tv and cellular is killing land lines.

This has the opportunity to kill all of the above plus m2m communications. While I suspect launch price will sway customers a good deal it may make some a bit shy about supporting a technology that may one day make you obsolete.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 01/17/2015 10:56 pm
Let's just hope he doesn't die of physical degradation. He works himself more than his body is designed for and probably burns calories like a furnace. It's a potential risk and a heavy point of alarm for me.

Another concern is legal troubles and the huge amount of obstruction he will face from the SatCom operators who are currently his customers. He's already dealing with significant amount of paper obstruction in launch services; what happens when he tries to make this satellite venture which is equally as risky, if not more so?

Mind you, his past record looks pretty impeccable and he seems to learn from his mistakes. He's even curtailed some of his estimate optimism.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Barrie on 01/17/2015 11:11 pm
Another concern is legal troubles and the huge amount of obstruction he will face from the SatCom operators who are currently his customers. He's already dealing with significant amount of paper obstruction in launch services; what happens when he tries to make this satellite venture which is equally as risky, if not more so?

Ideally those customers will be persuaded that what EM is doing is the inevitable future of their business, and want to be part of it.  Could they be persuaded that this is the next generation of their service, and invest in it?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/17/2015 11:12 pm
I hope that what he's doing is getting his customers on board.

They realize the age of broadcast GEO satellites will end, and are looking for a way to join the transition. The thing is, I don't think they know how. Internet constellations are very different - both technically and business wise.

However, selling in each particular market is always a tough job. Musk can easily partner with each of these operators in the specific markets they are strong in, and it's a win-win.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: inonepiece on 01/17/2015 11:24 pm
So how are the people at ORBCOMM feeling now, do we think?  Somebody once said capitalists will sell you the rope to hang them with; have ORBCOMM funded the rope to hang themselves with?

What does OG2, which presumably had a long life on paper, provide that this will not (if it is successful in providing service)?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/18/2015 12:31 am
Elon said it would fund a city on Mars, he has said in the past that he expects there to be several cities on Mars. Now whether this idea or some other one, or any at all will generate enough revenue for Elon to start a Mars colony, if he can do it and does start one, others will be built. As well Elon's will be the most expensive because it will be the first, he will happily sell capacity on his ships (and build more) for others to set up their own cities.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 01:03 am
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
What about a de-orbit tether? The spacecraft has to reset a timer on the tether module or it triggers automatically and de-orbits the satellite.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CJ on 01/18/2015 01:39 am
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
What about a de-orbit tether? The spacecraft has to reset a timer on the tether module or it triggers automatically and de-orbits the satellite.

What's the proposed orbital altitude of this sat constellation? They're supposed to use hall effect thrusters for station keeping... and be in low orbit. So, what about placing them a hair under ISS altitude? The drag wouldn't be anything an occasional thruster fire wouldn't handle, but any dead or dud sat would self-dispose in a few months.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 01:44 am
What's the proposed orbital altitude of this sat constellation? They're supposed to use hall effect thrusters for station keeping... and be in low orbit. So, what about placing them a hair under ISS altitude? The drag wouldn't be anything an occasional thruster fire wouldn't handle, but any dead or dud sat would self-dispose in a few months.
Musk mentioned >1000 km so that won't be enough...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 01:49 am
So how are the people at ORBCOMM feeling now, do we think?  Somebody once said capitalists will sell you the rope to hang them with; have ORBCOMM funded the rope to hang themselves with?

What does OG2, which presumably had a long life on paper, provide that this will not (if it is successful in providing service)?
Well, they got a good deal on their launch, they can still get a good deal on future launches, and they have competition that would have been there with or without them.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 01:54 am
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
What about a de-orbit tether? The spacecraft has to reset a timer on the tether module or it triggers automatically and de-orbits the satellite.
Yes, that's very similar to what I was proposing upthread.

A drag tether, a dead man's switch, and for extra measure deployment to a slightly lower orbit so that those that are stillborn never make it into the active orbits.

With any of these things, you can either try to make it happen, or come up with every reason why it can't possibly work. I think SpaceX is taking the first approach.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: su27k on 01/18/2015 01:59 am
So how are the people at ORBCOMM feeling now, do we think?  Somebody once said capitalists will sell you the rope to hang them with; have ORBCOMM funded the rope to hang themselves with?

What does OG2, which presumably had a long life on paper, provide that this will not (if it is successful in providing service)?

I doubt they're worried, version 1 of this thing is NET 5 years, we all know how a NET date is likely to slip and version 1 is rarely a big hit. It would take a few iterations for the internet constellation to become a serious threat, probably after OG2's life has run out. When designing their next generation service, Orbcomm could just join SpaceX's effort instead of rolling their own hardware (they don't build it anyway).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 01/18/2015 02:32 am
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
What about a de-orbit tether? The spacecraft has to reset a timer on the tether module or it triggers automatically and de-orbits the satellite.
Yes, that's very similar to what I was proposing upthread.

A drag tether, a dead man's switch, and for extra measure deployment to a slightly lower orbit so that those that are stillborn never make it into the active orbits.

With any of these things, you can either try to make it happen, or come up with every reason why it can't possibly work. I think SpaceX is taking the first approach.


I think regulators will mandate that such satellites in lowish Earth orbits each have a deorbit mechanism to remove them at end-of-life. It's best practice. This will still work for some malfunctioning satellites, but there'll be some where it doesn't. You could have a back-up and even a backed-up back-up system for deorbiting, but this only reduces the number of satellites that won't deorbit, not eliminate them.

Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 02:38 am
2.  No, you did not show a method that:
a. works on a dud
b. works without causing other issues to other spacecraft.
c.  if the main computer is dead, the whole spacecraft is dead.
What about a de-orbit tether? The spacecraft has to reset a timer on the tether module or it triggers automatically and de-orbits the satellite.
Yes, that's very similar to what I was proposing upthread.

A drag tether, a dead man's switch, and for extra measure deployment to a slightly lower orbit so that those that are stillborn never make it into the active orbits.

With any of these things, you can either try to make it happen, or come up with every reason why it can't possibly work. I think SpaceX is taking the first approach.


I think regulators will mandate that such satellites in lowish Earth orbits each have a deorbit mechanism to remove them at end-of-life. It's best practice. This will still work for some malfunctioning satellites, but there'll be some where it doesn't. You could have a back-up and even a backed-up back-up system for deorbiting, but this only reduces the number of satellites that won't deorbit, not eliminate them.

Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
The dV for satellite removal are IMO daunting.

But if you treat de-orbit like an FTS - complete separate system, self contained, etc - then I think you can have a good enough reliability rate.

From what I've read, the drag tether is fairly passive.

Another option is to have the drag tether be a backup to a hall thruster deorbiter, and then it is really hard to have dual failures.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 02:55 am
Yes, that's very similar to what I was proposing upthread.

A drag tether, a dead man's switch, and for extra measure deployment to a slightly lower orbit so that those that are stillborn never make it into the active orbits.

With any of these things, you can either try to make it happen, or come up with every reason why it can't possibly work. I think SpaceX is taking the first approach.
I assume the engineering options aren't a mystery to the spacecraft folks, it's just a question of how low the failure rate can go for de-orbit failsafe. A tether wouldn't be perfect either. Can potentially have two but then what if they both fire? Maybe a second tether system could trigger on a low speed omni antenna if the ground confirms it's needed.

I don't think it could achieve perfection but a reasonable goal would be a lower likelihood of stranded failed satellites than existing smaller constellations.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 02:59 am
Yes, that's very similar to what I was proposing upthread.

A drag tether, a dead man's switch, and for extra measure deployment to a slightly lower orbit so that those that are stillborn never make it into the active orbits.

With any of these things, you can either try to make it happen, or come up with every reason why it can't possibly work. I think SpaceX is taking the first approach.
I assume the engineering options aren't a mystery to the spacecraft folks, it's just a question of how low the failure rate can go for de-orbit failsafe. A tether wouldn't be perfect either. Can potentially have two but then what if they both fire? Maybe a second tether system could trigger on a low speed omni antenna if the ground confirms it's needed.

I don't think it could achieve perfection but a reasonable goal would be a lower likelihood of stranded failed satellites than existing smaller constellations.
Completely agree, since they consequence of a single or a few dead satellites are minimal - this situation already already exists.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 03:19 am
Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
I think you're really overestimating the overhead from including tethers on everything, and underestimating the effort of creating a totally separate de-orbit design.

Another option is to have the drag tether be a backup to a hall thruster deorbiter, and then it is really hard to have dual failures.
Presumably the hall thrusters already there serve as a primary, but require way too much of the spacecraft to be working to be safe to assume. The tether is a backup. That might cut expected debris satellites from dozens to hundreds to a handful. An extra tether with a separate trigger might cut that from a handful to <1.

The separate trigger would be maybe a low speed omni antenna or something like that. You probably don't want both deployed, but operators can confirm the primary tether failed before triggering the secondary.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 03:26 am
I think a much more interesting question is what technologies they have in mind in order to make a < half ton satellite that can do the job.

Did they find a way to reduce power consumption for example?

These satellites are on an almost 50% power cycle. So you need to produce 2-3 times your average power consumption. Any idea what this level is?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: butters on 01/18/2015 03:56 am
Orbcomm and Iridium market turnkey network appliances for specialty applications.

SpaceX will be a wholesale provider of a general-purpose network based on open standards for third-party devices and (probably) a franchise model for network service retailers. This is a very different business model which is unlikely to penetrate those niche markets until it becomes ubiquitous in the commercial mass market.

Besides retail marketing and billing, these carrier franchises would act as certificate authorities and keep all the records mapping cryptographic keys to real-world identities subject to the legal jurisdiction in which they operate. This model would allow governments to spy but not censor, which is a compromise worth noting.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/18/2015 04:00 am
I really do look forward to Elon talking about this issues in interviews.  ;D
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 01/18/2015 04:17 am
$10 or $15 billion dollars isn't chump change. 
But there's no specific mention of external funding and presumably (hopefully) Elon wants to keep his controlling position in SpaceX. 

Personal proceeds from selling equity in other ventures in 3 or 4 years?
Creating a satellite subsidiary that brings in market booty or partners with existing sat companies?

My guess is drag it along for 2 or 3 years on SpaceX dime (designing sat offerings and architecture chasing several paths, design the kilofactory and processes, etc.) then spin the subsidiary to investors, with SpaceX maintaining a controlling position and the market/partners coughing up to $15 billion for almost half of the subsidiary's equity.  That distances the big financial gamble from the rest of SpaceX. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 01/18/2015 04:22 am
Also, if I were to sell millions of autopilot vehicles, what benefits could 4000+ satellites bring to that? 
For safe discussion, let's frame that question in the context of a very large, well connected Mars colony with good roads (since Mars SpaceX satellites have finally been mentioned), but understanding in the back of our minds that such benefits could accrue, and perhaps be initially pilot tested on other planets such as Earth (without naming any forbidden brand names or famous inventors of AC power or common units of magnetic flux density).   
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 04:48 am
Also, if I were to sell millions of autopilot vehicles, what benefits could 4000+ satellites bring to that? 
For safe discussion, let's frame that question in the context of a very large, well connected Mars colony with good roads (since Mars SpaceX satellites have finally been mentioned), but understanding in the back of our minds that such benefits could accrue, and perhaps be initially pilot tested on other planets such as Earth (without naming any forbidden brand names or famous inventors of AC power or common units of magnetic flux density).   

Westinghouse?!   :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/18/2015 10:37 am
For deorbit, I think it's possible they will have little tugs that attach to a latch point on a dead sat, and deorbit it.

If sats are at similar inclinations, I believe that could switch across to other planes if they're prepared to wait a while for precession (?) to do the job.

Possibly some of the CommSats (the spares?) might be launched with a tug in place at the attach point. If another sat needs to be de-orbited, it would detach, transit, attach, then do its job.

Actually, the spare that replaces the dud has to transit to the gap in coverage. The tug may push it there to maximise on-orbit life of the spare.

If the dead sat still has propellant, perhaps the tug could drain it for the deorbit burn (or control the sat's own thrusters), and maybe have enough left over to get back amongst the constellation afterwards.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Llian Rhydderch on 01/18/2015 12:56 pm
For deorbit, I think it's possible they will have little tugs that attach to a latch point on a dead sat, and deorbit it.

If sats are at similar inclinations, I believe that could switch across to other planes if they're prepared to wait a while for precession (?) to do the job.

Possibly some of the CommSats (the spares?) might be launched with a tug in place at the attach point. If another sat needs to be de-orbited, it would detach, transit, attach, then do its job.

Actually, the spare that replaces the dud has to transit to the gap in coverage. The tug may push it there to maximise on-orbit life of the spare.

If the dead sat still has propellant, perhaps the tug could drain it for the deorbit burn (or control the sat's own thrusters), and maybe have enough left over to get back amongst the constellation afterwards.

Cheers, Martin

Alternatively, in lieu of full deorbit by a tug, the tug might accumulate n of the defective/dead little buggers, then move them up to a higher graveyard orbit, analagous to what is done in the Clarke belt. 

With this option, the aluminum, silicon, etc. could subsequently be "mined" someday by Planetary Resources, Deep Space Industries and such.   ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/18/2015 03:59 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/18/2015 04:46 pm
There's a team of bright engineers behind Virgin Galactic. I don't like the hate that's spewed towards the company just because people don't like Branson's personality.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Zed_Noir on 01/18/2015 05:00 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Think Musk will see what each state will be offering. IMO it is more likely to be in either South California, Seattle or Denver. There are existing facilities and human resources concentrated in those areas. There is about zero infrastructure at Boca Chica. Even water have to be trucked in.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/18/2015 05:17 pm
If I were him I'd build the satellite plant in Brownsville proper not at Boca Chica. It has water, power, a shipping channel to any of their launch sites, UT-RGV and, very importantly, a very friendly tax and regulatory structure.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/18/2015 05:37 pm
There's a team of bright engineers behind Virgin Galactic. I don't like the hate that's spewed towards the company just because people don't like Branson's personality.
I did not mean either hate, or I any disrespect towards Virgin Galactic, by the contrary I was quite enthused about their project with the White Knight and its Spaceship.
It is the fact that, in spite of the good feeling that many shared, the Project did not come to fruition even after prolonged delays. It could be just the result of inadequate budget...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/18/2015 05:41 pm
It hasn't disappeared, either. And frankly, something like Virgin Galactic or XCOR is the best chance we have at visiting space.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/18/2015 05:50 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Think Musk will see what each state will be offering. IMO it is more likely to be in either South California, Seattle or Denver. There are existing facilities and human resources concentrated in those areas. There is about zero infrastructure at Boca Chica. Even water have to be trucked in.
Of course Musk will do his own judging. It just made sense to me to put the Plant right by the launching pad. It could also be Brownsville, as @docmordrid stated, and very important the friendly policy Texas state is showing to SpaceX and all other investors of this magnitude.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/18/2015 06:02 pm
Can Brownsville be used to launch into the satellites' orbits? To cover most of the earth they would need a high inclination.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/18/2015 06:04 pm
It hasn't disappeared, either. And frankly, something like Virgin Galactic or XCOR is the best chance we have at visiting space.
I agree. Unfortunately, it is not enough to have a great idea and brilliant engineers, you should also have good management, too. Or, in our case, to launch a large fleet of satellites in a very short period of time they would need a lot more powerful vehicle, hence more complex problem to deal with.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Callezetter on 01/18/2015 06:26 pm
Just posting Elons Seattle video from yesterday if anyone missed it. Mostly about Satellites etc

https://www.youtube.com/watch?v=AHeZHyOnsm4
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 01/18/2015 06:35 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Think Musk will see what each state will be offering. IMO it is more likely to be in either South California, Seattle or Denver. There are existing facilities and human resources concentrated in those areas. There is about zero infrastructure at Boca Chica. Even water have to be trucked in.
Of course Musk will do his own judging. It just made sense to me to put the Plant right by the launching pad. It could also be Brownsville, as @docmordrid stated, and very important the friendly policy Texas state is showing to SpaceX and all other investors of this magnitude.
The problem with Brownsville is attracting talented staff. Satellite factory can be anywhere as transportation of small satellites is not a problem.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/18/2015 06:42 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Think Musk will see what each state will be offering. IMO it is more likely to be in either South California, Seattle or Denver. There are existing facilities and human resources concentrated in those areas. There is about zero infrastructure at Boca Chica. Even water have to be trucked in.
Of course Musk will do his own judging. It just made sense to me to put the Plant right by the launching pad. It could also be Brownsville, as @docmordrid stated, and very important the friendly policy Texas state is showing to SpaceX and all other investors of this magnitude.
The problem with Brownsville is attracting talented staff. Satellite factory can be anywhere as transportation of small satellites is not a problem.
Agreed! Was going to say the same thing. Lots of people wouldn't want to move down there, and no reason you need to put the satellite plant there.

Anyway, you're not going to be launching the constellation from Texas, the plane change required would be way too big. Probably would launch from California.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JH on 01/18/2015 06:47 pm
I don't see any reason to place the satellite factory near the launch site. First of all, they can only do a dozen launches per year from Boca Chica, meaning that they'd probably also be launching their in house satellites from some of the 3 other pads they currently have access to (even assuming that Boca Chica is at a favorable latitude for the launches). Secondly, satellites aren't so large that the only way to move them around the country is over water.

Then there is the staffing issue. I can't picture people who are used to Seattle happily moving to the far end of Texas.

-typo
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/18/2015 06:48 pm
No need to move down there. U. Texas  Rio Grande Valley has an engineering school there,  and Texas A&M is planning on building another. That would be 3 engineering schools within 60 miles of Brownsville. Not to mention that if the design work is done in Redmond the factory shouldn't need as  many.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 07:36 pm
My guess is drag it along for 2 or 3 years on SpaceX dime (designing sat offerings and architecture chasing several paths, design the kilofactory and processes, etc.) then spin the subsidiary to investors, with SpaceX maintaining a controlling position and the market/partners coughing up to $15 billion for almost half of the subsidiary's equity.  That distances the big financial gamble from the rest of SpaceX.
This broadly makes sense. Some potential synergies (secondary payload on CRS launches will allow recurring opportunities to validate satellites in orbit).

Also even watching the video it's totally non-obvious what stages this involves and how much investment over what time period. It definitely sounds like they're starting small. This sounds like engineering of the basic spacecraft bus, propulsion, etc for the first few years. Then scale up would get expensive. It's entirely possible they'll go multiple years before they spend the first $100 million, then start getting spendy when they start scaling up. At that point the traditional investment path you describe makes a lot of sense. Then they can go to investors with "here's the satellite bus, we've flown it 5 times and tested it, here's the manufacturing line, here's the launch vehicle".
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/18/2015 08:07 pm
For deorbit, I think it's possible they will have little tugs that attach to a latch point on a dead sat, and deorbit it.
Just not seeing it. You don't know what condition it's in, whether it's spinning or whatever. Doing this economically seems orders of magnitude harder than just including passive failsafe, harder even than the main payload.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 01/18/2015 09:00 pm
My guess is drag it along for 2 or 3 years on SpaceX dime (designing sat offerings and architecture chasing several paths, design the kilofactory and processes, etc.) then spin the subsidiary to investors, with SpaceX maintaining a controlling position and the market/partners coughing up to $15 billion for almost half of the subsidiary's equity.  That distances the big financial gamble from the rest of SpaceX.
This broadly makes sense. Some potential synergies (secondary payload on CRS launches will allow recurring opportunities to validate satellites in orbit).

Also even watching the video it's totally non-obvious what stages this involves and how much investment over what time period. It definitely sounds like they're starting small. This sounds like engineering of the basic spacecraft bus, propulsion, etc for the first few years. Then scale up would get expensive. It's entirely possible they'll go multiple years before they spend the first $100 million, then start getting spendy when they start scaling up. At that point the traditional investment path you describe makes a lot of sense. Then they can go to investors with "here's the satellite bus, we've flown it 5 times and tested it, here's the manufacturing line, here's the launch vehicle".
There's also a substantial amount of work that needs to be done on network architecture which will probably run in parallel with the bus and payload work.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nydoc on 01/18/2015 09:28 pm
I think a much more interesting question is what technologies they have in mind in order to make a < half ton satellite that can do the job.

Did they find a way to reduce power consumption for example?

These satellites are on an almost 50% power cycle. So you need to produce 2-3 times your average power consumption. Any idea what this level is?
Assuming we're talking about sat-to-sat laser connections and sat-to-ground RF connections at an altitude of 1100-1200km, what is the power requirement per sat?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RedLineTrain on 01/18/2015 09:36 pm
Thanks to Callezetter for posting the video.  It sounds as if SpaceX has already filed with the ITU for this system (at 17:20).  If anybody has access to the ITU filing database, it would be great to see SpaceX's filings or a list of its filings.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AJW on 01/18/2015 10:06 pm
An element I haven't seen mentioned is that SpX may be able to launch a number of satellite prototypes as secondary payloads on unrelated missions at minimal cost.  They could try different iterations during development while refining capabilities such as de-orbit. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/18/2015 10:07 pm
A question for those who know - by there being a global Spectrum allocation, I gather that those frequencies are expected to be used, well, globally.

But satellites at 1000 km can only see spots at a radius of about 1000 km, which is sub-continent sized.  From a regulatory point of view, would it be possible for the system to use one set of frequencies, for example, in N. America, a different set in Europe, and a third in Africa?  Would that allow for more options?

From a technical point of view, what would it take to run at multiple frequencies?  I know a dish antenna reflector is good for a very large frequency band, but a multi-element antenna is pretty finely tuned. Phased array receivers work at a certain frequency, but I think that can be soft-controlled, right?


EDIT:  and btw, if you do approximate area math, it looks like they can run a tighter beam and still have very good coverage.  So perhaps a 30-degree high-quality zone, and a 45-degree (half angle) contingency service zone
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 01/18/2015 10:13 pm
Very satisfactory to me that he mentionend deorbiting satellites when out of service or defective. Important enough to him to actually mentioning it twice.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 01/18/2015 10:37 pm
There's a SpaceX Genereral section, Reusability Section, Missions section...  I suspect it's time for a SpaceX Satellite section.  This thread seems to have a wide variety of jumbled, but all interesting sub-topics related to the SpaceX satellite business, which might benefit from targeted incisions.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/18/2015 10:58 pm
Not nearly time for a whole new section. Come on.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/18/2015 11:00 pm
Not nearly time for a whole new section. Come on.

Agreed. Wait until there's actually something to see and talk about..
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/19/2015 12:13 am
Thanks to Callezetter for posting the video.  It sounds as if SpaceX has already filed with the ITU for this system (at 17:20).  If anybody has access to the ITU filing database, it would be great to see SpaceX's filings or a list of its filings.

Apparently their SpaceQry software has to be installed on a Windows system and I'm on my mobile. Not certain if it covers applications,

https://www.itu.int/ITU-R/go/space-software-space-query-and-extract-system/en
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 12:44 am
I listened to the event video. He said the plan is to have a global capability of gigabit-level connectivity in the 20-30ms latency range. He was very clear that it was going to use a phased array satellite terminal that has a view of the sky (in the $100-$300 range) and that wouldn't attempt to target the cellphone market (it wouldn't use wall-penetrating frequencies, unlike some other attempts). He also said it'd target low-density areas, but would still be able to serve roughly 10% of cities.

Overall, this seems feasible technically, though very challenging. Big question is all the electronics on the satellite. ViaSat-1 had 140Gbps of throughput, and there should be roughly 10 of Musk's satellites in view at any one time. Given a 1:100 over-subscribed ratio for 1 Gigabit (which is typical, I think) and each satellite having 1000 times the capacity of Viasat (which isn't totally unrealistic for 15 years in the future so 10 Moore's Law doublings and the far lower altitude), then SpaceX's constellation should be able to serve 140 million people in a region. For a region like Eastern China, that works out to roughly 10% of the population. Most areas of the world won't be that highly populated.

...it can be done, but it'll be very difficult. Not pursuing cell spectrum makes it a lot easier. It also means they aren't directly competing with Iridium, which focuses more on voice (although Iridium does intend to do more data with their new constellation).

One inherent advantage here over a new fiber or cellular internet start-up is that the constellation can be its own backhaul. And they can offer lower long-distance latency than any terrestrial service (other than those super-expensive microwave links that high frequency traders use between like Chicago and New York, but that's not affordable for mortals, of course).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 12:55 am
...what's interesting to me is that eventually such a constellation will start caching content (given how much bandwidth Netflix uses, you'd want to start with that very soon), some companies would want some processing to happen actually on the satellite itself for lower latency. That opens up data centers in orbit. Very interesting possibility for future growth (although somewhat hampered initially by the space environment and heat rejection). It allows you to harness space-based solar power without having to somehow transmit the energy to Earth (just the information). That's an economic forcing-function that'd allow totally new ways of increasing the space economy. But that's decades in the future.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 01:09 am
It's worth noting that SpaceX would definitely not be the first company to do satellite broadband. ViaSat has over 600,000 subscribers in the US, offers 10Mb/s (or better?) connections largely through a single geosynchronous satellite, ViaSat-1, which has 140Gb/s throughput capability. ViaSat has about $1.4 billion in annual revenue.

To put that in perspective, as of summer 2014, SpaceX has had a total of just $4 billion in revenue, so ViaSat probably has greater revenue than SpaceX, currently (it's estimated SpaceX had $800 million in revenues for 2014: http://www.fool.com/investing/general/2014/09/21/how-much-is-spacex-stock-worth.aspx ). ..with basically just a single satellite! (edit: the 1 satellite is for North America. they also have service in Europe, so they have at least one satellite for that) If anyone needed convincing that the space industry is WAY bigger than just the launch industry, this should provide it.


SpaceX's constellation would easily be 1000x more ambitious with several million times more throughput (EVENTUALLY, not at first). It's not unreasonable to suggest that SpaceX may get tens of billion in revenue through this constellation... perhaps even pushing into the hundreds of billions, long-term as the world economy grows.

The upside to a project like this is so big that even if it has a 50% chance of failure, it's still a pretty good investment idea (especially when you consider that corporate loans have less than half the interest rate of when Iridium, Globalstar, Teledesic, etc were being built and went bankrupt). Especially since it allows SpaceX to build up spacecraft expertise and fill out its launch manifest anyways if it ever wants to enable Mars colonization. It's kind of a no-lose situation for SpaceX with an ENORMOUS upside if it works.


(and I know Musk said it wasn't going to go after cell spectrum, but if the network is successful enough at providing broadband internet, then they may well start buying up more cell spectrum as they gain confidence on whether or not they have much hope technically of competing with cell companies)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Port on 01/19/2015 01:20 am
by the way:

in the video he said around 3:30 the speed of light is about 40-50% faster in vac than in space

optical satellite comm anyone?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/19/2015 01:42 am
I'm very happy with the phased array terrestrial antenna approach.  Rooftops, cars, airplanes, boats.   Last 1000' can be done with WiFi in almost any scenario.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 01:51 am
by the way:

in the video he said around 3:30 the speed of light is about 40-50% faster in vac than in space

optical satellite comm anyone?
Radio is light. Radio is the same exact thing as visible light, just different wavelength. Nobody says "speed of radio" when speaking of propagation speed.

...which doesn't mean they couldn't use optical for satellite-to-satellite communications, of course. But you wouldn't really want to rely on it for space to ground because of clouds. And he specifically mentions phased array antenna.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Port on 01/19/2015 01:55 am
I'm well aware that any electromagnetic radiation can be seen as light :)

just saying that he's using the analog of optical lines, also laser-com in space should be pretty advantageous for reasons planetary resources can tell you, obviously not for ground-space because of obvious reasons :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/19/2015 02:02 am
Still finishing this transcript. http://shitelonsays.com/transcript/spacex-seattle-2015-2015-01-15

Okay, it's done now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 02:10 am
I'm well aware that any electromagnetic radiation can be seen as light :)

just saying that he's using the analog of optical lines, also laser-com in space should be pretty advantageous for reasons planetary resources can tell you, obviously not for ground-space because of obvious reasons :)
Well, it's likely to be really high frequency, whether that means millimeter or terahertz or optical or whathaveyou (although the higher frequency is needed, since we're talking on the order of 10-100Tb/s or greater per satellite... eventually)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 02:21 am
Radio link speed (or actually, from some perspective, optical as well) is a combination of antenna gain on receiver (proportional to area of antenna, roughly) and on transmitter, how distant the two are from each other, how much attenuation due to atmosphere and obstacles, signal power, bandwidth, and noise (including extraneous signals and thermal noise in the amplifier itself). Also, you can use multiple links to different satellites to improve throughput as well (and sometimes even take advantage of signals bouncing off the ground).

So, you can use 100GHz of bandwidth to produce a 10Tb/s link if you have enough gain on each end and use enough power.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 02:31 am
I'm well aware that any electromagnetic radiation can be seen as light :)

just saying that he's using the analog of optical lines, also laser-com in space should be pretty advantageous for reasons planetary resources can tell you, obviously not for ground-space because of obvious reasons :)
Keep in mind Planetary Resources is more restricted in antenna size than SpaceX is likely to be, due to the fact it's already going to need large phased arrays for space-to-ground (and back). Additionally, they will be operating at greater distances and won't need as fast of a slew rate as SpaceX will (this last point tilts heavily toward a lower frequency option since phased arrays generally don't operate in the optical range).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 02:50 am
What will be critical here for getting the necessary throughput will be very high performance (and cheap!) phased arrays. I'm talking millions of elements per satellite, or at least a hundred thousand or so. That will be a huge challenge to do cheaply.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/19/2015 03:17 am
I got the impression that Elon is talking about roof mounted antennas. Perhaps Solar City could install them :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 03:22 am
I got the impression that Elon is talking about roof mounted antennas. Perhaps Solar City could install them :)
Yes, roof-mounted phased array antennae. Need to be phased array so you can steer them electronically and quickly without mechanical wear and tear (and alignment issues with mechanical antennae).

It'd be a great backhaul for cellphones, actually. Especially if combined with solar power and battery. Just install a cell site on a roof somewhere as a pod. Doesn't need any wires or anything, just access to the sky. Or, put it on a high-altitude solar drone (now we're getting ahead of ourselves).

Note that phased array satellite receivers are already in widespread use to provide in-flight internet on ocean-crossing flights. Also, they used to be used for getting satellite TV on RVs. But the cost will need to come down (this is partly a Moore's Law sort of thing).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/19/2015 03:40 am
I got the impression that Elon is talking about roof mounted antennas. Perhaps Solar City could install them :)
Yes, roof-mounted phased array antennae. Need to be phased array so you can steer them electronically and quickly without mechanical wear and tear (and alignment issues with mechanical antennae).

It'd be a great backhaul for cellphones, actually. Especially if combined with solar power and battery. Just install a cell site on a roof somewhere as a pod. Doesn't need any wires or anything, just access to the sky. Or, put it on a high-altitude solar drone (now we're getting ahead of ourselves).

Note that phased array satellite receivers are already in widespread use to provide in-flight internet on ocean-crossing flights. Also, they used to be used for getting satellite TV on RVs. But the cost will need to come down (this is partly a Moore's Law sort of thing).

http://governor.state.tx.us/news/press-release/20148/  This is a link in case you missed this piece of info re phased array antenna. It is going to be developed built and distributed in collaboration with University of Texas at Brownsville, under the Stargate Project. I think it makes a lot of sense now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 03:46 am
I got the impression that Elon is talking about roof mounted antennas. Perhaps Solar City could install them :)
Yes, roof-mounted phased array antennae. Need to be phased array so you can steer them electronically and quickly without mechanical wear and tear (and alignment issues with mechanical antennae).

It'd be a great backhaul for cellphones, actually. Especially if combined with solar power and battery. Just install a cell site on a roof somewhere as a pod. Doesn't need any wires or anything, just access to the sky. Or, put it on a high-altitude solar drone (now we're getting ahead of ourselves).

Note that phased array satellite receivers are already in widespread use to provide in-flight internet on ocean-crossing flights. Also, they used to be used for getting satellite TV on RVs. But the cost will need to come down (this is partly a Moore's Law sort of thing).

http://governor.state.tx.us/news/press-release/20148/  This is a link in case you missed this piece of info re phased array antenna. It is going to be developed built and distributed in collaboration with University of Texas at Brownsville, under the Stargate Project. I think it makes a lot of sense now.
Good call!

There's a lot of potential here for both cost and performance (mostly just more elements) improvements in phased arrays. Chip-level phased arrays are a thing, meaning you can take advantage of Moore's Law-like advancements (however long that will last).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 04:03 am
Interesting, looks like STARGATE started in 2012. There isn't a mention of phased array, but that may be assumed when talking about advancing the technology for satellite and spacecraft tracking. I wonder if this is a sign SpaceX was seriously considering this sort of project since at least 2012? http://www.brownsvilleherald.com/news/local/article_d8af741e-3841-11e2-9dd5-001a4bcf6878.html
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/19/2015 04:10 am
Interesting, looks like STARGATE started in 2012. There isn't a mention of phased array, but that may be assumed when talking about advancing the technology for satellite and spacecraft tracking. I wonder if this is a sign SpaceX was seriously considering this sort of project since at least 2012? http://www.brownsvilleherald.com/news/local/article_d8af741e-3841-11e2-9dd5-001a4bcf6878.html
I do not know, but this guy is moving like playing a high level chase game
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/19/2015 04:20 am
 "I wonder if this is a sign SpaceX was seriously considering this sort of project since at least 2012?"

He said he established SpaceX to serve this purpose - colonizing Mars - and then he said Sats network is meant to finance Mars operation... He might've thought of such a plan for long time ago now
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/19/2015 04:54 am

At the stage they are today - where there's already a baseline design - it must be at least a year since they started the effort, and two years is perfectly reasonable.

They had to get key concepts together, get key people on the core team, file whatever regulatory paperwork was necessary, oh, raise some serious funds (probably not $10B, but at least they had to feel the water about getting that much money if development is successful)

This announcement only happened because they got to the point where it could not remain a secret anymore - they needed to ramp up hiring, perhaps some of the filings become public soon, etc.   Also, the saga with Wyler probably would have made it public anyway.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 01/19/2015 10:32 am
Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
I think you're really overestimating the overhead from including tethers on everything, and underestimating the effort of creating a totally separate de-orbit design.
Very possibly! :) But there may be a cross-over point financially. And you need to consider reputational and regulatory risk from leaving any dud satellite up there. And of course such a dud removal satellite needn't be limited to SpaceX's duds!

Musk is talking about doubling the number of satellites. Concern about space junk is only going to grow with possible consequential regulatory action. Musk is very good at spotting such potential problems in advance, so I wouldn't be surprised if SpaceX hadn't given at least some thought to this.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/19/2015 12:23 pm
Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
I think you're really overestimating the overhead from including tethers on everything, and underestimating the effort of creating a totally separate de-orbit design.
Very possibly! :) But there may be a cross-over point financially. And you need to consider reputational and regulatory risk from leaving any dud satellite up there. And of course such a dud removal satellite needn't be limited to SpaceX's duds!

Musk is talking about doubling the number of satellites. Concern about space junk is only going to grow with possible consequential regulatory action. Musk is very good at spotting such potential problems in advance, so I wouldn't be surprised if SpaceX hadn't given at least some thought to this.

They can use the Hall thrusters for normal disposal. An automatically released tether as a backup mechanism. For those few satellites which are not deorbited by either of these (which may be in the order of 1/year) then a dud removal satellite can be launched on the next generation satellite launch to that plane.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/19/2015 12:35 pm
It won't have escaped Elon's notice that

profits from applications > profits from operation > satellite manufacturing profits > launch profit

The profits from applications are defuse spread out among thousands of companies, so they are not a good target to go after.

He seems to be going after the profits from operating a large network, the profits from satellite manufacturing and launch for that network will follow.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 03:22 pm
You're missing the fact that the taxpayer wouldn't be paying for this. Also, satellite internet is incredibly profitable and competes just fine with "wires." ViaSat had more revenue last year than SpaceX, and ViaSat only has a couple satellites.

Also, satellite TV is a one-way signal, there's no opportunity for graceful degradation like there would be here. It's not likely to be a problem, could easily have better service availability than cable Internet does.

Also, Musk mentioned Gigabit/s speeds, which is just as good as Google fiber.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: symbios on 01/19/2015 03:23 pm
Do not count any revenue from NASA on this. Maybe if they manage to snag something for the Mars part that NASA has a request out for.

Everyone should remember that the Investment is spread over a long timer period. We are not talking about a 10 b in one time payment.

Regarding sales. In Sweden we have local "city-nets" where the city puts in the investment for the backbone and putting in the lines to individual buildings. Then "operators" do the sales and administration for the individual customers and rent from the city for the usage. This way the backbone operator has very little administration and sales. They only handles the infrastructure and the operators.

I hope they don't try to make all the sales themselves to the end customer.

They also have a guarantied worldwide customer base. Remember that Elon Musk is hoping to build 500.000 new cars per year in 2020 that all need internet connection. :D
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/19/2015 03:36 pm
Bringing this over from the comparison thread:

Will these be launched on F9 or FH?
How many per shot?
What is the cost per launch?

For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.

I'm also thinking it might be useful to have two or more layers, for enhanced coverage when there's a failure.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 03:41 pm
Bringing this over from the comparison thread:

Will these be launched on F9 or FH?
How many per shot?
What is the cost per launch?

For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.
Not unreasonable, although maybe on the high side for satellite mass (though not if you include the per-satellite deployed mass). But anyway, launch is likely to be a minority of total costs even at current prices.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/19/2015 03:47 pm
Bringing this over from the comparison thread:

Will these be launched on F9 or FH?
How many per shot?
What is the cost per launch?

For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.
Not unreasonable, although maybe on the high side for satellite mass (though not if you include the per-satellite deployed mass). But anyway, launch is likely to be a minority of total costs even at current prices.
If we assume single layer, if looks like there's a very large number of sats visible from the ground at any given time.

Still, what multi-layer does, assuming each layer is still dense enough to provide continuous service, is reduce the significance of the "hole" created by a single dead satellite, making it larger and softer.  (The absolute ideal would have been a 1/4025 degradation spread evenly, right?)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RAN on 01/19/2015 03:58 pm
Google fiber may no longer be a competitor!

https://www.theinformation.com/Google-Nears-Major-Investment-in-SpaceX-to-Bolster-Satellites
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 01/19/2015 04:09 pm
Google fiber may no longer be a competitor!

https://www.theinformation.com/Google-Nears-Major-Investment-in-SpaceX-to-Bolster-Satellites
So the X in Google's SolveForX, which was started in Feb '12 iirc, is probably the same X as in SpaceX
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 04:42 pm
$10 billion is much smaller than the amount of money needed to put fiber all over the world.

Anyway. It's not intended to completely replace fiber. There's huge potential for a project like this, even if it doesn't end up being the primary Internet provider for the world. I mean, if they hit even 1% of their target market, they will be profitable enough to pay the interest on their corporate loans.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/19/2015 04:51 pm
To me, this is like the Elon trying to 'Rick Roll' US taxpayers to pay for a substantial portion of the world's commercialized internet services (or perhaps the planet Mars).
Why would taxpayers pay for it? I'd expect it to be privately financed. E: aaaaaaaand that already looks to be the case :)

Just as a consumer, I'd rather pay a little more for the super-fastGoogle Fiber when it becomes available versus a choppy signal from space.
Agreed, but connectivity this fast has very limited reach. Even densely populated areas have pretty limited connectivity for the most part. That buildout will also take billions of dollars, repeated for everywhere else in the world.

Everytime a huge thunderstorm rolls thru, satellite TV stops working.
A few advantages here for the SpaceX scheme...

-BotE ElonSats will be able to provide >100x the power for their footprint, 20 dBm
-It can boost power over the storm with battery power well beyond steady state. It's moving at orbital velocity so it only has to do that for a few seconds.
-Satellite TV is broadcast. Two way communications can use adaptive coding and step down to more robust codings when attenuation is higher.

Still not perfect, but outages could be limited quite a bit most places. And rain knocks wired service out all the time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Blackjax on 01/19/2015 04:53 pm
Elon probably has multiple reasons for a plan like this and likely has multiple angles he is trying to get synergies from.  I've see a lot of good angles being discussed in this thread, but an angle I don't see being discussed much is the impact on launch volumes of going from a satellite constellation model which involves relatively few long lived expensive satellites to frequent launches of large numbers of inexpensive sats. 

RobotBeat touched on it in his post here: http://forum.nasaspaceflight.com/index.php?topic=36552.msg1317745#msg1317745

However I don't think that went far enough.  SpaceX bumping up flight rate from their own constellation might add a fair amount to their launch rate and help keep the reusables stocked with payloads, but is that the extent of the ambition here?

Consider a moment what is happening in the launch industry itself.  Elon announced he was going to change the launch market and drive the costs way down and the flight rate up.  At first established players like Arianespace and ULA ignored him, then they sneered at him, then they fought him, now they are starting to change their business practices in response to him and are both designing rockets to optimize for cost to a greater extent than they would have without his competitive pressure.

Now stop and consider what would happen to the global launch market if SpaceX acts as the first mover and proves to existing  (or potential new entrant) constellation operators (and consequently satellite manufacturers) that this sort of paradigm is a good idea.  Others will follow, possibly enough to prompt a sea change in how a substantial part of the market does business.

In this scenario, what might happen to the demand for launches and the total amount of payload the world launches? 

10 years down the road, if SpaceX makes this look like a good approach to operating a constellation, what will be the impact on business practices *beyond* SpaceX?


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/19/2015 04:58 pm
1) So is SpaceX going to pay for their own launch facility to launch this stuff?  Is this what Brownsville TX is supposed to be all about?
Unlikely. Can't reach the needed inclination. From Brownsville it couldn't even reach most of the US.

2!!!)  Well.... how do I put this.  if Google Fiber is failing, why all of the sudden would this idea work?
Google Fiber works great, it's just extremely slow to build out, and $10B is nothing compared to what it would take to reach even half the US with fiber.

There are more affordable (as well as faster) ways to put up a 'Google Fiber level bit rate' infrastructure around the world.
Google's working on that too, millimeter wave experiments to avoid having to run fiber to every house, but that would require an extremely dense deployment, much tighter than cell towers, so it still basically only works in densely populated areas.

A powerful new satellite constellation might bring billions of new internet users online all over the world in addition to nudging the US towards better connectivity, that's well worth it for Google.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 01/19/2015 05:00 pm
The vast majority of the US (outside of the urban/major markets) will NEVER get gigabit speeds, using fiber to the home/business. Wireless tech is the only hope. Not to mention the developing world is even in worse shape.

GoogleFiber will only be in a few US markets. And the US cable companies, that roll out DOCSIS 3.1 over the next few years, will only give 1-3Gbps downstream. There is no plan in the US for massive rural downstream, and massive urban and rural upstream bandwidth. Musk sees a problem, finds a solution. Now I just have to see it work, for which I have my doubts. Not really related to tech, nor even the capital.

Remember Gwynne shotwell has been complaining about the broadband divide for years.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ronsmytheiii on 01/19/2015 05:02 pm
1) So is SpaceX going to pay for their own launch facility to launch this stuff?  Is this what Brownsville TX is supposed to be all about? 
2!!!)  Well.... how do I put this.  if Google Fiber is failing, why all of the sudden would this idea work?

There are more affordable (as well as faster) ways to put up a 'Google Fiber level bit rate' infrastructure around the world. 

1.  The new satellites  wont be GEO which is pretty much what Brownsville is optimized for, so probably not. Either CCAFS or Vandenberg (or even both) depending on how SpaceX wants to do the orbital planes.

2. Google is looking all around to improve internet connectivity to get more Google users, as the internet search providers in developed nations is becoming saturated they need to expand access to get more customers.  They invested in o3b as well, more internet access means more customers using google, and therefore more profits.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ronsmytheiii on 01/19/2015 05:05 pm
In addition to Google's capital investment, wonder if they will help SpaceX in hardware development.  They are in fact the worlds largest hardware producer (though its for their own use)

http://www.wired.com/2012/07/google-server-manufacturing/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: butters on 01/19/2015 05:14 pm
So let me see if I have this straight: Wyler quits O3b and takes the spectrum rights (expiring in 2019) with him to Google. He fails to win Google's approval to put up a constellation in that timeframe, quits, and forms his own company OneWeb. Wyler possibly tries to pitch SpaceX, but they aren't interested in rushing into operation by 2019 either, so Google finds their kindred spirit and invests in SpaceX. Wyler takes his spectrum and his chosen corners of the triangle ("cheap" and "soon") to Branson, who is better suited to attracting investors seeking a quick exit.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 01/19/2015 05:23 pm
Google actually putting in some cash and knowledge would definitely be a huge plus. But those are just rumors, right now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/19/2015 05:34 pm
So let me see if I have this straight: Wyler quits O3b and takes the spectrum rights (expiring in 2019) with him to Google. He fails to win Google's approval to put up a constellation in that timeframe, quits, and forms his own company OneWeb. Wyler possibly tries to pitch SpaceX, but they aren't interested in rushing into operation by 2019 either, so Google finds their kindred spirit and invests in SpaceX. Wyler takes his spectrum and his chosen corners of the triangle ("cheap" and "soon") to Branson, who is better suited to attracting investors seeking a quick exit.

It makes zero sense.  Wyler is taking his launch business to someone who is

A) Still does not have a launch vehicle or motor, and has never put anything in orbit (and for all practical purposes, not even in sub-orbit)
B) Even in his most optimistic paper-rocket projections is stating a cost level that is not viable ($10M per each?  That's $6.5B launch costs for a much weaker constellation)
C) Has a bad track record in anything launch-related.

It's not even a Hail Mary... It's more like throwing the ball up in the air just to defer the buzzer, and hoping a plan comes together before the ball comes back down.

....  but this is OT, we have a separate thread for the Wyler constellation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 05:44 pm
One more comment on Wiley.

This sort of thing isn't that uncommon. Someone has an idea which they think is the best ever (whether or not it is... many times they weren't the first to think of it, as in Teledesic) which they're over-protective of, they develop it some, then when the project doesn't go exactly how they want, they pack up and try to make it work in their own, unadulterated way (which nearly always fails). Musk is right. It's the ability to execute that counts. Interesting ideas are dime a dozen (just have a look at http://www.sti.nasa.gov/ if you think you're the first person to think of some aerospace idea).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/19/2015 05:52 pm
For all of those people who suggest that copper or fiber is cheaper than a $10B satellite constellation, I want to point out that for a certain percentage of North Amercian residents (arguably between 5 and 10%) getting wired internet services faster than old dial up speeds would cost in excess of $5k (being the mininum price to add one more utility pole by a utility company).   Where you put copper or fiber, it can generally service a large number of clients at a cost effective price, but you are left with a significant number of rural and suburban residences with no good choices for higher speed connectivity.

As for the recent rumor in the media about Google working with SpaceX on a satellite constellation.  I don't see why it couldn't be true, but we don't have anything official confirming it. The one thing that does jump out at me is that there must be some player, other than Wyler, Page, and Musk trying to get Google into this.  Maybe this other player was motivated by Wyler's involvement, or maybe it is someone who had courted Wyler for Google or Musk, but it seems to me that the comments that are attributed to "someone close to the deal" were probably far more calculated that either Musk or Page is when they speak publicly.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 01/19/2015 07:00 pm
This potential deal with Google remind me of the time that Larry Page said he'd prefer to give his money to Elon Musk (in case Page dies or whatever) instead of straight up charity since Musk is doing a lot to change the world (and Page sort of considers what Musk wants to do with Mars as a sort of philanthropy).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mmeijeri on 01/19/2015 07:10 pm
If they're going to do this, I don't understand why Musk isn't interested in a fully reusable FH, it would be perfect for this sort of thing. Even a fully reusable F9 might be enough.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hauerg on 01/19/2015 07:14 pm
If they're going to do this, I don't understand why Musk isn't interested in a fully reusable FH, it would be perfect for this sort of thing. Even a fully reusable F9 might be enough.
Elon said doing the reusable upper stage would be a "distraction" for his engineers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Hauerg on 01/19/2015 07:24 pm
4 ?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 01/19/2015 07:28 pm
Google fiber may no longer be a competitor!

https://www.theinformation.com/Google-Nears-Major-Investment-in-SpaceX-to-Bolster-Satellites
I was hoping/almost expecting Google to be involved in this project. With Google backing it funding will not be an issue.

The question is will this satellite company be a new entity/company that uses SpaceX as launch provider or just another division of SpaceX?. A separate company means Elon doesn't need to dilute his shareholdings of SpaceX, plus if it fails it will not take SpaceX with it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: dante2308 on 01/19/2015 09:38 pm
Assuming $15 billion for the constellation and 14 Gbps/satellite (the 'less sophisticated' number for WorldVu/Oneweb) with 4000 satellites you get 60,000 Gbps.

If you assume a 20% usage factor due to the presence of oceans, 100kbps/customer average, and the satellites lasts 5 years you get $0.40/month/customer cost. With $10 billion you get $0.26/month/customer. This requires about 629 million customers to sign up though....

100kbps per customer average is about 250 Gb per month.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/19/2015 09:49 pm
The Moog dispensers for the ORBCOMM  OG2's can hold 4 satellites per segment ,  so the upcoming launch will need to use 3 segmets to launch the remaining 11 OG2 satellites. Could 4 or more Moog dispenser rings fit?

This is using the existing 5.2x13m fairing. IIRC there was talk of a larger FH fairing down the road?

Edit: yup - here

Link.... (http://forum.nasaspaceflight.com/index.php?topic=30850.msg1192430#msg1192430)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 01/19/2015 10:13 pm
Could the larger fairing be used on Falcon 9? I suspect the payload could be limited by fairing size rather than weight.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 01/19/2015 10:17 pm
If they're going to do this, I don't understand why Musk isn't interested in a fully reusable FH, it would be perfect for this sort of thing. Even a fully reusable F9 might be enough.
Elon said doing the reusable upper stage would be a "distraction" for his engineers.

...because of the energy associated with GTO launches.  A 4025 satellite constellation in LEO/MEO seems to be closer to a reason to build it... development cost is a rounding error on this size of opportunity.  Maybe the "distraction" issue will resolve itself in say 5 years.

By the way of predictions:
Quote
In the next decade, low earth orbit satellite networks will make global mobile service a reality and costs will plummet as competition emerges.  Current mobile comms revenue of $1.2T will shift to space-based, quickly outstripping the $200B TV and radio revenue.  Low cost launches, probably using reusable vehicles, will propel the following shift of WIFI coverage to global. Today's 3B Internet users will grow to 6B, and increasingly expect uninterrupted coverage like cell phone users will have by 2020. Space satellite revenue will grow to surpass $2T by 2025.

and...
Quote
Space debris will become a major issue...

http://forum.nasaspaceflight.com/index.php?topic=35872.msg1276623#msg1276623
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mlindner on 01/19/2015 10:28 pm
I fixed a few things in the transcript of the Seattle SpaceX event.

A few of the things I changed:
He said that there'd be new versions of the system every two to three years instead of one to two years.
He said that the satellite system can't be free for use.
He pointed out that China could shoot down the satellites specifically.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pathfinder_01 on 01/19/2015 11:48 pm
In response to Robotbeat's earlier post (now deleted)...

K-lamba-T has a value of 165.301 if :
1) SpaceX can launch every 60 days (at the current rate of "6" per year) ,
2) with a satellite that has a failure rate of 1 every 4 years (0.029 per 1000 hrs ... current Orbcomm satellite service life)
3) and assuming 100% delivery to orbit success.
4) there is are a constellation of 4025 satellites

If there is no less than 95% probability of having a replacement available to keep the network active... then the number of satellites required to be launched is >> 100 (value is off the chart).  A higher probability of a spare on hand will drive a higher launch rate or more satellites per launch.

The weight of the satellite per launch is essentially irrelevant.  The problem is the production rate of the FH engines... it is a bottle neck (its a race to get a F9 with 9 engines up 6 times a year).  An engine with that level of production capability has not been demonstrated.... thus there has to be a new agenda at SpaceX.

Graphic taken from the maintainability NAVSHIPS 94324 nomograph (circa 1964):

1. This guy is aiming for re-usability of the first stage at least. That alone could cut the number of engines needed.

2. An rocket can launch more than one satellite at an time.

3. This is an 10-15 year plan and those are long term goals. He could launch more or less.
 
4. The satellites in question are to be only a few hundred kg.  The current Falcon 9 could lift 13,000 kg to LEO and 4,800 to GE0.  If each satellite only masses an few hundred kg he could put up an lot in a single launch.

5. He could also take advantage of any left over lift capacity of an sold flight to put some up lowering his costs.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: tvillars on 01/19/2015 11:52 pm
The Wall Street Journal

Google Inc. is close to investing roughly $1 billion in Space Exploration Technologies Corp. to support its nascent efforts to deliver Internet access via satellites, according to a person familiar with the matter.

http://www.wsj.com/articles/google-nears-1-billion-investment-in-spacex-1421706642

Musk said his constellation could potentially handle 90% of the long haul needs of the internet with the remaining 10% handling direct connections to end users.  Would it make sense to combine Google's drones / airship concept for the last mile connection and SpaceX for everything else?  If this what they are really planning then it is the end of terrestrial ISPs.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mlindner on 01/19/2015 11:54 pm
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Why would they build the satellites near the launch facility? They're not size restricted like rockets. They can build the factory anywhere. Several hundred kilogram satellites are also quite small. They could easily build them right in Seattle adjacent to their Seattle location (likely even in the same building).

He even stated that half of the employees at the Seattle office (which apparently will number in the many hundreds) will be hardware people.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/19/2015 11:56 pm
What he said was:

Quote from: Elon Musk
The focus is going to be on creating a global communications system. This is quite an ambitious effort. We're really talking about something which is, in the long term, like rebuilding the Internet in space. The goal will be to have the majority of long distance Internet traffic go over this network and about 10% of local consumer and business traffic. So that's - probably 90% of people's local access will still come from fiber but we'll do about 10% business to consumer direct and more than half of the long distance traffic. - source (http://shitelonsays.com/transcript/spacex-seattle-2015-2015-01-15)

i.e., he hopes SpaceX can win a majority of long distance traffic, but only a very small amount of local traffic - as, ya know, going to space and back doesn't make a lot of sense for that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/20/2015 12:10 am
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Why would they build the satellites near the launch facility? They're not size restricted like rockets. They can build the factory anywhere. Several hundred kilogram satellites are also quite small. They could easily build them right in Seattle adjacent to their Seattle location (likely even in the same building).
Well, that was my opinion. I came to it considering this project has to be very competitive from the start, so rationalize your investment would make sense.
So, land is already there, available, the launching pad is there, they have a cooperation with University of Texas at Brownsville to build up the engineering force, there is a STARGATE project with Brownsville University aiming at Phased Array Antennas, they have the launching pad, the very friendly tax policies of the State of Taxes, designing mostly to be done in Seattle but the Plant, the hardware, could very well be stationed at Brownsville.
Not connected with this (?) , but also important in the whole picture, Elon has decided to start building a 5 mile hyperloop to proove the concept of supersonic train. Where do you think? In Texas.
So, this was my reasoning, it may not be so, we will see.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: tvillars on 01/20/2015 12:21 am
What he also said was:

Quote from: Elon Musk
If you've got a large mass of land where they're relatively low density of users, space is actually ideal for that. It would also be able to serve as, like I said, probably about 10% of people in relatively dense urban/suburban environments - cases where people have been stuck with Time Warner or Comcast or something this would be a ripe opportunity. It's something that would both provide optionality living in advanced countries/economies as well as people living in poor countries that don't even have electricity or fiber or anything like that. So it's a real enabler for people in far reaches of the world and it gives optionality for people in wealthier countries. It's something that I think definitely needs to be done, and it's a really difficult technical problem to solve. So that's why we need the smartest engineering talent in the world to solve the problem.
source (http://shitelonsays.com/transcript/spacex-seattle-2015-2015-01-15)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mlindner on 01/20/2015 01:15 am

The weight of the satellite per launch is essentially irrelevant.  The problem is the production rate of the FH engines... it is a bottle neck (its a race to get a F9 with 9 engines up 6 times a year).  An engine with that level of production capability has not been demonstrated.... thus there has to be a new agenda at SpaceX.

You'll have to cite that they have a production bottleneck or it looks like you're talking out of your nether regions. They don't have a production bottleneck they have a launch processing bottleneck.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/20/2015 01:19 am
With more payloads available I think they could have gotten to eight launches last year.

That said, I think Elon is looking towards launching this constellation on the fully reusable BFR. It should be ready in five years. If so, that launch rate is all they'll need.


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/20/2015 01:56 am
human rated

You do understand this is a nonsense term, right?

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Lar on 01/20/2015 03:14 am

 ;D


So in actual seriousness, wondering what SpaceX customer Orbcomm is thinking with this news?  Have they gone elsewhere or are they part of this idea?  They are already providing the best business case for delivering communications in remote locations of the Earth.  Looks like this news event makes them almost disappear.  Is this a merger or a tech grab from a customer?



Your posts are starting to come off as a bit concern-trollish. You might want to work on that. Think hard before you post again, please.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 01/20/2015 03:17 am
I'm sure they're even more sceptical than you are (but not as much as me, no-one is.)

Most of the world takes Elon's ramblings with a pinch of salt - assuming he hasn't outright insulted them, which he seems to be aiming to do to someone in almost every interview this month. Compared to, say, everyone who worked on Teledesic, I expect Orbcomm is feeling pretty good about themselves right now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 01/20/2015 03:23 am
http://www.geekwire.com/2015/photos-spacex-founder-elon-musk-unveils-new-10b-space-internet-plan-private-seattle-event/

It seems that Elon has decided to go alone in this new venture covering its triple main components: (1) designing and manufacturing satellites, (2) launching and (3) operating sat network. If the first two were obvious the third one appeared plausible after his friend Greg Wyler went the separate way along with Qualcomm and Virgin Group. It seems that Virgin Group will be in charge of launching the sats which, based on the recent long-effort-no-results of Sir Richard Branson's Space Tourism, I personally doubt very much will be a success. It remains tbs if any truth in "OneWeb backers questioning whether Musk will be able to get rights to the spectrum required to build the SpaceX network."

"Musk has not yet determined where to put the satellite production factory — an unclaimed prize for which Washington and other states will no doubt compete fiercely."
Imo, that place will be in Texas on the same land Elon builds the new Launching Pad. It looks like Texas is gonna be the winner again.

Why would they build the satellites near the launch facility? They're not size restricted like rockets. They can build the factory anywhere. Several hundred kilogram satellites are also quite small. They could easily build them right in Seattle adjacent to their Seattle location (likely even in the same building).
Well, that was my opinion. I came to it considering this project has to be very competitive from the start, so rationalize your investment would make sense.
So, land is already there, available, the launching pad is there, they have a cooperation with University of Texas at Brownsville to build up the engineering force, there is a STARGATE project with Brownsville University aiming at Phased Array Antennas, they have the launching pad, the very friendly tax policies of the State of Taxes, designing mostly to be done in Seattle but the Plant, the hardware, could very well be stationed at Brownsville.
Not connected with this (?) , but also important in the whole picture, Elon has decided to start building a 5 mile hyperloop to proove the concept of supersonic train. Where do you think? In Texas.
So, this was my reasoning, it may not be so, we will see.

It's advisable for SpaceX to spread the goodies around; build political support in Congress etc. It's probably also not escaped Elon's notice that goodwill for any of his three companies will probably rub off on the other two; if so, you'd expect the companies to avoid as far as possible setting up major new facilities in the same state!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bocephus419 on 01/20/2015 04:36 am
I'm seeing a group position themselves in an already established market that has been collapsing due to costs.

Well, I bet SpaceX agrees with you and wants to develop technologies to lower those costs. I take it you think they'll fail, but if they succeed it could open up huge new markets for themselves and their competitors/customers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ludus on 01/20/2015 04:48 am

It's advisable for SpaceX to spread the goodies around; build political support in Congress etc. It's probably also not escaped Elon's notice that goodwill for any of his three companies will probably rub off on the other two; if so, you'd expect the companies to avoid as far as possible setting up major new facilities in the same state!

For example if it was to be built in the great state of Alabama NASA might find it coincidentally easier to get full funding on completely unrelated commercial space activities  :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Razvan on 01/20/2015 05:16 am
It's advisable for SpaceX to spread the goodies around; build political support in Congress etc. It's probably also not escaped Elon's notice that goodwill for any of his three companies will probably rub off on the other two; if so, you'd expect the companies to avoid as far as possible setting up major new facilities in the same state!
I agree, building support in Congress is critical, if only thinking of AF/ULA story.
OTOH ISTM, Elon is seeing Texas as a major base in his operation plan. Remember him mentioning first human colonist will go to Mars probably from Boca Chicca, Texas

edit/Lar: fix quotes
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/20/2015 06:23 am
And in his Texas legislature testimony he also mentioned a possible new factory at Brownsville. And it's a deep water  seaport.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 01/20/2015 09:28 am
Bringing this over from the comparison thread:

Will these be launched on F9 or FH?
How many per shot?
What is the cost per launch?

For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.

I'm also thinking it might be useful to have two or more layers, for enhanced coverage when there's a failure.
64x64? The Earth is flat and square? :)

No, splitting 4025 into multipliers gives the following viable solution: 81x25... 81 satellites in 25 planes, perhaps 1 is spare, so...

This gives each plane a satellite every 500km(40000km/80), and if the planes are inclined 51.3 degrees to the equator, the distance between planes is also 500km.

increasing inclination to 60 degrees covers almost all world population except north european countries which have by coincidence the best internet in the world.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/20/2015 09:42 am
Bringing this over from the comparison thread:

Will these be launched on F9 or FH?
How many per shot?
What is the cost per launch?

For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.

I'm also thinking it might be useful to have two or more layers, for enhanced coverage when there's a failure.
64x64? The Earth is flat and square? :)

No, splitting 4025 into multipliers gives the following viable solution: 81x25... 81 satellites in 25 planes, perhaps 1 is spare, so...

This gives each plane a satellite every 500km(40000km/80), and if the planes are inclined 51.3 degrees to the equator, the distance between planes is also 500km.

increasing inclination to 60 degrees covers almost all world population except north european countries which have by coincidence the best internet in the world.
Yes, of course, that was just to get a quick idea, starting with polar...
Everything else stays....
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Misha Vargas on 01/20/2015 01:03 pm
[...]
For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.
[...]
[...]splitting 4025 into multipliers gives the following viable solution: 81x25... 81 satellites in 25 planes, perhaps 1 is spare[...]

Yes, of course, that was just to get a quick idea, starting with polar...
Everything else stays....

I was just about to ask why not 115 x 35, or 161 x 25, or 175 x 23. (Musk did say this number was probably overprecise.)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Llian Rhydderch on 01/20/2015 01:15 pm
Rather than have such back-ups on each and every member of the 4500 and counting satellite constellation, for the vast majority of which it will be an unneeded expense, it might be simpler and more cost-effective to have one or more dedicated dud satellite removal satellites which sidle up to your duds and by one means or another causes them to deorbit.
I think you're really overestimating the overhead from including tethers on everything, and underestimating the effort of creating a totally separate de-orbit design.
Very possibly! :) But there may be a cross-over point financially. And you need to consider reputational and regulatory risk from leaving any dud satellite up there. And of course such a dud removal satellite needn't be limited to SpaceX's duds!

Musk is talking about doubling the number of satellites. Concern about space junk is only going to grow with possible consequential regulatory action. Musk is very good at spotting such potential problems in advance, so I wouldn't be surprised if SpaceX hadn't given at least some thought to this.

Absolutely correct.  And this brings with it, as does time and understanding of the existing satellite situation, that the populace may become more aware, and it may become politically expedient for the political class to begin to regulate the negative externality of space debris. 

However, if this were to occur, then in my view it would not be a huge disadvantage to Elon and his new 4000+ satellite constellation, especially if launch licenses eventually become highly restrictive (like the proposed "one up, one down" legal regime that has been proposed theoretically).

If SpaceX just designs in a capability for eventual collection and disposal-- whether per satellite with tethers or ..., or with a collection tug in each of the constellations orbital planes, or something else--SpaceX will actually have a competitive advantage over many of the existing satellite makers and their products.  So I don't see how such a requirement for timely deorbit is a large adverse problem for SpaceX intentions here.

Will require engineering work, and the company incurs will incur a cost to do this, but assuming the legal regime applies to all new launches after some point in time, and not just the SpaceX mega-constellation, its not a competitive disadvantage to SpaceX (should such a legal regime be put in place).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/20/2015 03:15 pm
Any speculation on the frequency range? They have submitted an ITU filing, and Space News sources haven't noticed any in the Ku/Ka range. This leads me to believe that they will use the ~34-50 GHz range for down/uplink and use the highly directional antennas to power through rain fade.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 01/20/2015 03:25 pm
They have submitted an ITU filing,

Do we know that? I have not seen it, may have missed it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/20/2015 03:32 pm
Do we know that? I have not seen it, may have missed it.
I've not "seen" it either yet (still looking), but Musk said they filed it (17:24)
http://youtu.be/AHeZHyOnsm4?t=17m24s
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RedLineTrain on 01/20/2015 03:37 pm
But Space News is reporting that it is not yet out in the wild.

Quote
Peter B. de Selding ‏@pbdes  22 hours ago
Sat industry official: We're always on lookout for new ITU Ku/Ka filings. We havent seen any SpaceX networks...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/20/2015 03:43 pm
could it be above Ka?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pietro on 01/20/2015 03:47 pm
Some people are worried about the SkyNet sats blocking access to space, so I tried to do some calcs: if my napkin-calculations are correct, 4025 stats @ 1100 km LEO orbit would occupy 0.0000000573% of the sky - so not a real problem, I believe ;) (if they are in the same plane and each has an 10mx10m bounding area)

  ((10/1000)**2*4025)/(4*pi*(6371+1100)**2)*100 %

Does this look all-right?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: starsilk on 01/20/2015 04:12 pm
could it be above Ka?

atmospheric absorption becomes an issue above Ka. even Ka has problems with rain fade, Ku is better for that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: symbios on 01/20/2015 04:13 pm
But Space News is reporting that it is not yet out in the wild.

Quote
Peter B. de Selding ‏@pbdes  22 hours ago
Sat industry official: We're always on lookout for new ITU Ku/Ka filings. We havent seen any SpaceX networks...

They have three other bands to try
L-band 1-2GHz (Not likely) Satellite navigation and terrestrial mobile networks
S-band 2-4GHz,  As I understand it the S band is not that common for satellite use but has been allocated.
C-band 4-8GHz, Highly possible, used for satellite communication.
X-band 8-12GHz, as I understand it the X band is reserved for US gov/military.
K-bands >12GHz

They will probably go lower and not higher. The different K-bands are sensitive for weather. The lower bands are less sensitive (<10GHz).

As said in a comment bellow, lower frequency usually requires larger antennas. (see more on comment below). Does not know how this is affected by the choice of antenna that SpaceX has made (phased array).

Edit: expanded information.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: starsilk on 01/20/2015 04:22 pm
But Space News is reporting that it is not yet out in the wild.

Quote
Peter B. de Selding ‏@pbdes  22 hours ago
Sat industry official: We're always on lookout for new ITU Ku/Ka filings. We havent seen any SpaceX networks...

They have three other bands to try L(1-2GHz)/S (2-4GHz)/C(4-8GHz). The several K-bands starts at 12 GHz
As I understand it the X band is reserved for US gov/military.

They will probably go lower and not higher. The K-band is sensitive for weather. The lower bands are less sensitive.

As I understand it the S band is not that common for satellite use but has been allocated.

the big problem with that is they are talking about very high bandwidth, from small 'home mounted' dishes. the lower frequencies make that much harder to do - they'd need a wide slice of the band.

satellite TV uses Ka for the high def channels because of the bandwidth requirements, and push the power way up to be able to punch through rain fade - which they can do because they are using 'battlestar galactica' satellites. might not be possible with SpaceX's proposed small satellites (although they are far, far closer so perhaps they can..)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/20/2015 04:28 pm
But Space News is reporting that it is not yet out in the wild.
That's what I said. That's why I'm guessing it's neither Ka nor Ku.
They will probably go lower and not higher. The different K-bands are sensitive for weather. The lower bands are less sensitive (<10GHz).
Musk wants bandwidth in the "Gbps range" for every user, and 50% of long range internet traffic. I'm not seeing that with lower frequencies.
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
They are not using wall penetrating frequencies, and plan to use highly directional antennas. Isn't that enough to power through rain fade?
(http://abload.de/img/rainfadeo4u3a.jpg)
edit: Multiply the values by 5 for total rain fade.

My guess is ~34-50 GHz.

edit: I've written a long post on reddit, if someone wants to read that: https://www.reddit.com/r/spacex/comments/2t24zv/spacex_satellite_frequency_background_and/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 01/20/2015 04:34 pm
Do we know that? I have not seen it, may have missed it.
I've not "seen" it either yet (still looking), but Musk said they filed it (17:24)

Thanks. I see  how I could miss this. :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/20/2015 04:35 pm
Jim, the assertion was that if SpaceX anticipated regulation and designed in capability that they would be ahead. That does not suggest that they will be ahead. I took it as implicit in the caveat that between design and implementation there is execution risk as there is with all new activity. Before they put the cheese in orbit and brought it back there was extreme risk to the concept that they could build a spacecraft at all.

I think that, not without risk of anticipating wrong, SpaceX could very well influence regulation of satellite disposal rules in their activities leading up to and implementing this data communications satellite network. If they do influence it positively (ie their approach provides capabilities that closely match the final standards) then they definitely will have a competitive advantage. Even if that advantage only translates to the disposal technology and not the communications systems.  However, given the iterative nature of their proposed design and implementation process, I give them better odds than I give Irridium 2 or Orbcom's 2nd gen network at staying competitive. They have demonstrated the benefit of their iterative process, just not on communications, they certainly have demonstrated the flight control side engineering that we are discussing here.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/20/2015 05:08 pm
MeekGee I also remember some intense thread discussion on a possible constellation of communications satellites/GPS for Mars and what it would take to put it there and what the nature of the individual satellites would be like. This does not seem to be too far removed from those ideas either (though definitely larger satellites)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jim on 01/20/2015 05:08 pm

I think that, not without risk of anticipating wrong, SpaceX could very well influence regulation of satellite disposal rules in their activities leading up to and implementing this data communications satellite network. If they do influence it positively (ie their approach provides capabilities that closely match the final standards) then they definitely will have a competitive advantage.

They have been wrong before in anticipating regulation rules.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 01/20/2015 05:14 pm
Don't worry about it.

This is the same fanboism that was berated for suggesting over a  year ago that Musk is planning a very large telecom constellation, that will rely on smaller satellites, limited lifetime, and advanced technology over today's satellites.

Just keep fanboing, you're doing fine.

Could you please link me to that prediction? I'd love to read that discussion.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/20/2015 05:36 pm

I think that, not without risk of anticipating wrong, SpaceX could very well influence regulation of satellite disposal rules in their activities leading up to and implementing this data communications satellite network. If they do influence it positively (ie their approach provides capabilities that closely match the final standards) then they definitely will have a competitive advantage.

They have been wrong before in anticipating regulation rules.

Exactly, if they had never been wrong, and if I could point to a dozen cases or more, then I would not have added the caveat: "not without risk of anticipating wrong".  Just like MeekGee had his explicit (and a few, to me, implicit) caveats in his statement.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Mongo62 on 01/20/2015 05:39 pm
edit -- I see that ScepticMatt has already posted a direct link to this post upthread.

From ScepticMatt at Redditt (http://www.reddit.com/r/spacex/comments/2t24zv/spacex_satellite_frequency_background_and/):

SpaceX satellite frequency background and speculation

SpaceX has announced the development of a internet satellite constellation.
A question that often came up is what frequency and bandwidth they will use for down/uplink.
So as a newbie in this area I tried to gather all the background information to form an educated guess.

Considerations:

bandwidth needs


Musk wants majority of long distance traffic, 10% of local traffic. [1]

So how much traffic is this?

The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps

Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route

Source: Telegeography (https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf)

So I estimate that SpaceX would need 10 Tbps for each short and long distance communication in 2014.
Proportionally more if you account for growth rate during and after rollout.

In addition, he wants "Gigabit level access" [1]
Such high wireless bandwidths need high frequencies. For example, the upcoming wireless standard WiGig (IEEE 802.11ad) uses 60 Ghz for up to 7 Gbps line of sight data rate.

The Cancelled system Teledesic, was going to use 28.6-to-29.1-GHz for up and downlink, and 60 Ghz for intersatellite communications. 2 Mbps uplink, 64 Mbps downlink, 10 Gbps total. [3 (http://disi.unitn.it/~sacchi/EHF_SatComIEEEProceedings.pdf)]

In summary, this requires double-digit GHz frequencies, the higher the better.

Transmission Losses

A satellite internet system needs a strong enough signal to be received at the end.
Different frequencies have varying signal loss characteristics.

Free space losses

Musk wants to use a 1100 km orbit [1]
Satellite signal power is attenuated in vacuum by just distance and frequencyFrequency   Attenuation
10 GHz   173.228 db
20 GHz   179.248 db
30 GHz   182.77 db
40 GHz   185.269 db
50 GHz   187.207 db
60 GHz   188.791 db
70 GHz   190.13 db
80 GHz   191.29 db
90 GHz   192.313 db
100 GHz   193.228 db


So a 100 GHz signal is 100 times weaker than a 10 GHz signal.
Directionality of transmitter and receiver antenna reduces those losses [4]

Scattering losses

Transmission losses occur when millimeter waves traveling through the atmosphere are absorbed by molecules of oxygen, water vapor and other gaseous atmospheric constituents.

These losses are greater at certain frequencies, coinciding with the mechanical resonant frequencies of the gas molecules.

For current technology the important absorption peaks occur at 24 and 60 GHz. The spectral regions between the absorption peaks provide windows where propagation can more readily occur.

The transmission windows are at about 35 GHz, 94 GHz, 140 GHz and 220 GHz. [2]

Image: Atmospheric Absorption (http://abload.de/img/atmosphericabsorption2pkdg.jpg)

So in short, losses are very high between 50-70 GHz and above 120 GHz.
Effects on frequencies below ~20 GHz are negligible.

Rain fade

Millimeter wave propagation is also affected by rain. Raindrops are roughly the same size as the radio wavelengths and therefore cause scattering of the radio signal [2]

Image: Rain fade (http://abload.de/img/rainfadeo4u3a.jpg)

As you can see, rain losses can be huge, especially in the 94 GHz window.
Thankfully rain is intermittent, so this can be mitigated if you accept some downtime and/or have a lower frequency backup channel.

Image: USA Rain vs downtime (http://abload.de/img/raindowntimeaiuce.jpg)

So you need to account for 35 mm/h if you want to guarantee a 99.9% US wide downtime.
Or for 6mm/h if you want to guarantee a 99% downtime and use the backup for the 1%.

Licensing

To operate a communication satellite fleet, a ITU regulatory filing (https://www.itu.int/sns/) is required.
Frequencies are usually reserved on a first-come, first serve basis.
Looking at spektrumwiki [link], 10-30 GHz are "full".
For example, similar network O3b has reserved 17.8-19.3 and 27.6-29.1 GHz (Ka) ranges, and SkyBridge has reserved the 10.7-18.1 (Ku) GHz range.
SpaceX has submitted a ITU filing [1], but Space News Sources haven't noticed filing in the Ka/Ku ranges [Source]

Conclusion/Speculation

I believe SpaceX will be using ~34-50 GHz for up/downlink.
Those frequencies allow for high bandwidth and are not used as far as I can see.
They'll use highly directional antennas [1] to power through strong losses.
For communication between satellites, 60Hz is ideal, because it provides protection against ground-based intercept attempts due to strong oxygen absorption.

comments?

Resources:
1) Seattle announcement (https://www.youtube.com/watch?v=AHeZHyOnsm4)
2) FCC bulletin (http://transition.fcc.gov/Bureaus/Engineering_Technology/Documents/bulletins/oet70/oet70a.pdf)
3) EHF satellite paper (http://disi.unitn.it/~sacchi/EHF_SatComIEEEProceedings.pdf)
4) Fundamentals of satellite communication, part 2 (http://www.ieee.li/pdf/viewgraphs/fundamentals_satellite_communication_part_2.pdf)
Other parts: http://www.ieee.li/communications/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/20/2015 05:46 pm
About the filing system, what I found so far:

Quote
ITU RADIO REGULATORY FRAMEWORK FOR SPACE SERVICES

5.3 Advance Publication of Information
The aim of the API procedure prescribed under Section I of Article 9 of the RR is to inform all
administrations of any planned satellite system using a GSO or a non-GSO satellite and of its general Page 11 of 13
description. This mandatory (“starting a clock”) procedure provides a formal mechanism whereby any
administration can make a preliminary assessment of the effect that a planned satellite network is likely to
have on the stations of existing or planned satellite systems and their terrestrial stations in certain
frequency bands and comment accordingly. To this end, the administration responsible for the planned
satellite network has to submit to the Bureau, for API/A publication in the BR IFIC, the API data
stipulated in Appendix 4 to the RR not sooner than seven years and preferably no later than two years
before the planned date of bringing into use of the network or system.
If, upon receipt of the BR IFIC containing the API/A special section published under No. 9.2B of
the RR, any administration considers its existing or planned satellite systems or networks or terrestrial
stations to be affected, it may send its comments to the publishing administration, so that the latter may
take those comments into consideration when initiating the coordination procedure. A copy of these
comments may also be sent to the Bureau. Thereafter, both administrations shall endeavour to cooperate
in joint efforts to resolve any difficulties, with the assistance of the Bureau if so requested by either of the
parties, and shall exchange any additional relevant information that may be available
Source: http://www.itu.int/en/ITU-R/space/snl/Documents/ITU-Space_reg.pdf

So there should be a API filing (Section I of Article 9 of the regulation)

Now there are two data sources, the basic Space Network List and the full Space Network Systems database.
The latter has a site to search for APIs by frequency, but is password protected (ITU account, presumably)
https://www.itu.int/sns/advpub.html

The former is publicly available. check API, N-GSO, and use USA under adm. enter from date e.g. 1/1/2012
http://www.itu.int/net/ITU-R/space/snl/bsearchb/spublication.asp

I couldn't relate any of these abbreviations with SpaceX. If you knew, you can download the database file and open it with SpaceCap and get the frequency information.
Download link: http://www.itu.int/ITU-R/go/space-software-space-notification-system-pc-capture/en

Anyone with more luck identifying any abbreviations?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/20/2015 05:49 pm
Don't worry about it.

This is the same fanboism that was berated for suggesting over a  year ago that Musk is planning a very large telecom constellation, that will rely on smaller satellites, limited lifetime, and advanced technology over today's satellites.

Just keep fanboing, you're doing fine.

Could you please link me to that prediction? I'd love to read that discussion.

It's in various places, some have to do with a discussions of low-flying constellations (which Musk's isn't) and some not - but I'd rather only bring it up as a response to the "fanbois don't know jack" kind of posts, just to set the record straight.  Here's one recent (http://forum.nasaspaceflight.com/index.php?topic=34774.msg1201502#msg1201502) post (with an ill-formed last sentence...), and here's  another (http://forum.nasaspaceflight.com/index.php?topic=35292.msg1252693#msg1252693).

(Clearly we're not hitting all the parameters since these are just estimates, not on inside knowledge, but still...)

Otherwise, it's more constructive to keep the conversation on the obviously-happening constellation

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: symbios on 01/20/2015 06:00 pm
Questions:

1. Is it reasonable for a individual satellite to cover a 3 degree area from an altitude of 750 miles?

2 My orbital mechanics knowledge is low, so... to cover the earth they need not only the equator but they need x amounts of different inclinations.

If this is correct and the satellite had 3 degree coverage. If they want to cover an area of north 60 degree to south 40 degrees (24 degree axial tilt) total 100 degrees and 360 degrees around the earth

100/3 = ~33 different inclinations.
360/3 = 120 orbits?

33,33*120 = 4000 satellites + spares.

Is this a correct deduction?

If a customers Phased Array antenna can cover 20 degree of the sky each antenna could "see" ~30 satellites (circle not square?) at any given time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: randomly on 01/20/2015 08:44 pm
I think it's highly likely the satellite interlinks will be done with lasers. Iridium looked into using lasers in it's early development but microwave had sufficient bandwidth for them and was a lower risk technology at the time.

The required inter-satellite link speeds required of Musk's constellation are going to be enormously higher.

As to rain fade -
 To achieve high data rates from a given amount of spectrum you can use more complex modulation schemes. The tradeoff is you need higher signal to noise ratios to keep error rates manageable. With a sophisticated enough system you can accommodate reduced signal to noise ratios from rain fade by scaling back your modulation scheme. Instead of link drop out you just have reduced data rates.

For high bandwidth connections you are also going to need either phased arrays or tracking dishes. I'm betting phased arrays will be a centerpiece of the system with simultaneous reception from multiple satellites simultaneously. With 4000 satellites you will probably have at least 25 satellites in view (40-45 degree above the horizon) at any one time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Lars-J on 01/20/2015 08:56 pm
A repost from the generic SpaceX thread #11, since it may be related...
---------
A press release from SpaceX about raising an additional $1 billion from Google and Fidelity: http://www.spacex.com/press/2015/01/20/financing-round

Quote
Space Exploration Technologies (SpaceX) has raised a billion dollars in a financing round with two new investors, Google and Fidelity. They join existing investors Founders Fund, Draper Fisher Jurvetson, Valor Equity Partners and Capricorn. Google and Fidelity will collectively own just under 10% of the company. 

SpaceX designs, manufactures, and launches the world's most advanced rockets and spacecraft. This funding will be used to support continued innovation in the areas of space transport, reusability, and satellite manufacturing.

That would suggest that SpaceX is now valued at over $10 billion.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/20/2015 08:58 pm
satellite TV uses Ka for the high def channels because of the bandwidth requirements, and push the power way up to be able to punch through rain fade - which they can do because they are using 'battlestar galactica' satellites. might not be possible with SpaceX's proposed small satellites (although they are far, far closer so perhaps they can..)
Closer and more directional, GEO has to cover an entire hemisphere. In this case the satellites will pass over a storm in what, less than a minute? If their beams have a footprint of tens of kilometers on the ground you can power up handful of beams impacted by a storm for just as long as you need it. Can even exceed steady state power by drawing on batteries.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/20/2015 09:23 pm
I think it's highly likely the satellite interlinks will be done with lasers. Iridium looked into using lasers in it's early development but microwave had sufficient bandwidth for them and was a lower risk technology at the time.

The required inter-satellite link speeds required of Musk's constellation are going to be enormously higher.

As to rain fade -
 To achieve high data rates from a given amount of spectrum you can use more complex modulation schemes. The tradeoff is you need higher signal to noise ratios to keep error rates manageable. With a sophisticated enough system you can accommodate reduced signal to noise ratios from rain fade by scaling back your modulation scheme. Instead of link drop out you just have reduced data rates.

For high bandwidth connections you are also going to need either phased arrays or tracking dishes. I'm betting phased arrays will be a centerpiece of the system with simultaneous reception from multiple satellites simultaneously. With 4000 satellites you will probably have at least 25 satellites in view (40-45 degree above the horizon) at any one time.

Yes, that's key.  (multi-pointing)

It's beyond merely "make before break".  It's "use the best links at any given time", and that very handy when the weather's bad.

Optical inter-sat links, IMO, are definitely on the table.  Maybe bandwidth is sufficient with RF, but I believe power requirements are lower, and there have been insane advances in optical receivers that can completely eliminate the need for tracking.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/20/2015 09:32 pm
Questions:

1. Is it reasonable for a individual satellite to cover a 3 degree area from an altitude of 750 miles?

2 My orbital mechanics knowledge is low, so... to cover the earth they need not only the equator but they need x amounts of different inclinations.

If this is correct and the satellite had 3 degree coverage. If they want to cover an area of north 60 degree to south 40 degrees (24 degree axial tilt) total 100 degrees and 360 degrees around the earth

100/3 = ~33 different inclinations.
360/3 = 120 orbits?

33,33*120 = 4000 satellites + spares.

Is this a correct deduction?

If a customers Phased Array antenna can cover 20 degree of the sky each antenna could "see" ~30 satellites (circle not square?) at any given time.

From a ground station 20 degrees of the sky is NOT 20 degrees around the circumference of the earth. That depends on how high up the satellite is.  Nor is 20 degrees field of view from the satellite translatable to 20 degrees around the circumference of the earth.

I think it is more relevant to simply think about the spacing of the satellites themselves, if they were equally distributed around the globe then they would directly above points that were about 360km apart. That means that the nearest alternates are directly above points 360 km away when the primary is directly above you. The spread in degrees from any point on earth to find an alternate to one directly above you is the largest angle compared to finding alternates to ones that are closer to the horizon. However since the satellites if distributed evenly would be spaced 3.24 degrees apart around the circumference of the earth and since the furthest satellite you could see on the horizon would be above a point 31 degrees away there would be unequally about 17 satellites spaced over a 180 degree arc from horizon to horizon from any point and picking any arc. That would mean that at the horizon you would see about 11 degrees separation between satellites directly overhead the separation would be about 22 degrees. With there would always be a satellite within 11 degrees of direct vertical and 5 within 22 degrees of vertical.

Now that presumed even distribution and that is not actually all that likely, however it will not be a huge difference from that and it probably will be more concentrated than that at latitudes near the poles and below 55 degrees.  I would envisage putting a few near polar planes of satellites up and the rest of the planes lying between 28 and 55 degrees. I imagine some would launch from Vandenberg (<30%) and the majority from Florida.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jcc on 01/20/2015 09:40 pm
News break....
Google and Fidelity jointly invest $1 Billion for 10% stake in Spacex. "This funding will be used to support continued innovation in the areas pf space transport, reusabolity and satellite manufacturing."
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/20/2015 09:58 pm
I saw two FALCONSAT ITU filings...

... but they are 450 MHz Air force satellites.

lol

edit: here

http://tinyurl.com/mzk92n8
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/21/2015 01:23 am
FalconSAT is a cadet built sat at the USAF Academy.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mlindner on 01/21/2015 01:44 am
<snip>

I agree with ScepticMatt's analysis. I don't see them being able to do the volume of communications they propose unless they jump into higher frequency regimes or permanently take a giant swath of Ka/Ku bandwidth which I don't see as politically possible. Yes there's going to be some significant atmospheric attenuation but if the beams are highly directional as opposed to omnidirectional like most connections then this is certainly possible.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mlindner on 01/21/2015 01:48 am
Update from Google via NBC News (http://www.nbcnews.com/science/space/liftoff-spacex-gets-1-billion-google-fidelity-n289866)
Quote
Google said Don Harrison, the company's vice president of corporate development, would be joining SpaceX's board of directors. "Space-based applications, like imaging satellites, can help people more easily access important information, so we're excited to support SpaceX's growth as it develops new launch technologies," Google said in a statement.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Andy USA on 01/21/2015 01:54 am
I would remind members that if someone, i.e Jim, posts something that is against site rules, you REPORT the post. You do not then ask why it's allowed. Posts here are not pre-approved. We do not read every post. You need to bring it to a moderator's attention. You do not respond to it (i.e meekgee) causing two posts to be a problem.

There is no excuse for regular members not to know this.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/21/2015 01:57 am
Update from Google via NBC News (http://www.nbcnews.com/science/space/liftoff-spacex-gets-1-billion-google-fidelity-n289866)
Quote
Google said Don Harrison, the company's vice president of corporate development, would be joining SpaceX's board of directors. "Space-based applications, like imaging satellites, can help people more easily access important information, so we're excited to support SpaceX's growth as it develops new launch technologies," Google said in a statement.

A fast-refreshing Google Maps/Earth is actually a communication problem, not an imaging problem.  So it comes back to the comm constellation.

However, if the satellites taking photos over densely populated areas are downlinking them to data-centers that are away from populated areas, then they're using largely "free" bandwidth.

For Google, this is... how you say... a good "get"?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: butters on 01/21/2015 02:44 am
Another thing to consider about Google is that they operate a distributed storage system which demands massive amounts of bandwidth to synchronize and replicate the global storage volume across datacenters positioned all over the planet. Google's internal network uses custom hardware and their routing tables are dynamically reconfigured by software algorithms for maximum efficiency.

I'm sure that Google can come up with a clever way to use a satellite constellation as part of their internal network.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: enzo on 01/21/2015 05:20 am
This investment is going to be massively profitable for Google in the long run. The expansion of their developing world customer base, the satellite backbone bandwidth, and the prestige of going to Mars. I wonder if this investment carries any contractual agreements, e.g. usage of the satellites, discounts on launch services. With one man on the board, is there an expectation of favors to be paid?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: macpacheco on 01/21/2015 05:37 am
The vast majority of the US (outside of the urban/major markets) will NEVER get gigabit speeds, using fiber to the home/business. Wireless tech is the only hope. Not to mention the developing world is even in worse shape.

GoogleFiber will only be in a few US markets. And the US cable companies, that roll out DOCSIS 3.1 over the next few years, will only give 1-3Gbps downstream. There is no plan in the US for massive rural downstream, and massive urban and rural upstream bandwidth. Musk sees a problem, finds a solution. Now I just have to see it work, for which I have my doubts. Not really related to tech, nor even the capital.

Remember Gwynne shotwell has been complaining about the broadband divide for years.
I know a thousand times more about telecom than space. GEPON (passive splitted fiber at 2.5Gbps per fiber branch) provides massive bandwidth, and land routers will always be orders of magnitude cheaper than space ones. Fiber costs have been dropping, and today labor costs to install dominate fiber deployment costs. And there's the GEPON upgrade to 10Gbps speeds. Satellites will never be able to match those. Old school telecoms will eventually be forced to go full fiber, but they would rather do it over many decades.

But, there is a significant market for remote internet, remote MPLS, that is forced to pay as much as 50x more than a dedicated link in the big city for business to get dedicated connectivity to their central office. Something like US$ 2000 / month per Mbps, while suffering 1-2 second latency. Just this out in the middle of nowhere dedicated connectivity would migrate to Elon Sat @ the same cost to get 2x the bandwidth at the same price, while dropping latency to 50-100ms. That's a market worldwide worth billions per month. Having lots of high bandwidth sats @ 1000Km orbits = affordable to simply bridge a remote area to the local internet land hub using a single satellite at a time. Then there's affordable internet @ every ship away from line of sight with land. Then there's worldwide mobile voice @ a sane price. Those are all captive markets that only compete with other satellite networks, no need to compete with land prices directly.

Then there's basic universal broadband internet in places outside dense metro areas (like internet @ Gwynne ranch). People would happily pay US$ 200/month for 5Mbps best effort service (in places where they have no choice). Here it would also make sense to just bridge customers to the closest land internet hub.

In all of those cases I'm intentionally avoiding the need for a Space router, instead just a basic land to land system with the routing taking place in the ground, perhaps no need for satellite to satellite laser links.

Ok, perhaps I'm being way too conservative. That's not how Elon rolls. But he seems to be ok with incremental functionality. Just showing how much potential there is for a first gen Elon Sat constellation.

Oh, and even if SpaceX actually ends up managing on LEO commsat constellation, it doesn't mean they wouldn't be very interested in supplying the birds for competing constellations, after all, that's what he's done with Tesla (opened up all Tesla patents to the competition), in this case he would still be getting paid to make and launch the satellites (and make tens of billions along the way).

How will Elon pay for this ? The question that needs asking is how much will it cost to make the first dozen 1st gen functional elon sats ? I bet it's not many billions. He can probably build and launch a proof of concept for much less than a billion bucks. Once that works, he'll have tens of VC investors fighting over funding for this. Finally, nothing prevents SpaceX from spinning current SpaceX (everything but sats) from a separate SpaceXTelecom, that might be a publicly traded company (helps with getting funding).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: macpacheco on 01/21/2015 05:52 am
What he said was:
i.e., he hopes SpaceX can win a majority of long distance traffic, but only a very small amount of local traffic - as, ya know, going to space and back doesn't make a lot of sense for that.
It does makes sense for markets where existing fiber capacity is taken or there is no fiber. Specially if they offer point to point laser relay (laser means no spectrum required). I don't know if space to ground gigabit (and up) laser is feasible right now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/21/2015 08:08 am
It does makes sense for markets where existing fiber capacity is taken or there is no fiber. Specially if they offer point to point laser relay (laser means no spectrum required). I don't know if space to ground gigabit (and up) laser is feasible right now.
I'm sure it's been done, it'd just be super weather sensitive.

You can get multi-gigabit in Ka+ bands, no? Probably easier to do that and still have a chance of powering through rain. With those wavelengths at that altitude you could have some pretty small beams without needing huge antennas. Gigabits for small beams (if a satellite has to cover a circle of diameter 500 km, say there's 19 beams, that's ~65 km diameter per beam) isn't bad for anything other than a major city, and even for a city you'd probably be able to see multiple satellites with the adjacent ones over under-utilized areas and therefore able to lend surplus capacity to the city. This sounds like decent DSL speeds even in cities with a significant number of people using it, even faster in rural areas, could easily have faster internet than home if it's a cruise ship out to sea or airplane over the ocean.

Here it would also make sense to just bridge customers to the closest land internet hub.
Was thinking about the routing, Musk kept talking about the latency advantages. There's no latency advantage if you just terminate the connection at a local ISP. The satellites will have to know the global routing table and know which satellites are near which ground stations.

That doesn't prevent the service from being marketed in multiple ways, but in principle something that looks like a dedicated connection to the home office could just as easily be implemented entirely on the customer equipment. Probably better to avoid complicating the satellites if at all possible.

In all of those cases I'm intentionally avoiding the need for a Space router, instead just a basic land to land system with the routing taking place in the ground, perhaps no need for satellite to satellite laser links.
Satellite to satellite links can be millimeter wave and still have many gigabits I should think. There's a number of really strong absorption bands from water and oxygen making those bands completely useless for talking to Earth, use those.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/21/2015 09:55 am
There's no latency advantage if you just terminate the connection at a local ISP.

It seems to me that the satellite network will be the ISP.

The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.

In 2000 I worked on an internet routing project for Nortel (part of a team of 7), I did the UI and control software for the demo system. The idea is that at the edge of the network the IP packets are inspected and tagged with a series of tags and sent off to internal nodes. At the internal nodes the first tag are used to do the routing and popped (or swapped for another tag). The routes between nodes are set up by a global entity that manages bandwidth so that there is always enough bandwidth for these routes through the network. This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees). There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).  The difficult bit is then the global ('god') system that allocates bandwidth to paths through the network and tells the edge nodes how to tag packets due to their destinations. This is naively a O(n^3) problem, so doesn't scale well without tricks, for this network n=4025 so scaling is not a problem, just whether it is tractable at a size of 4025. It is easy to see how such an idea can be applied to a satellite network.

The variable size of IP packets is a problem, so on a lower level we transported them within ATM cells, and developed a cleaver way of mapping the tags onto the ATM VPI and VCI routing. Some such scheme (perhaps not using ATM but a larger cell) would be ideal for the satellite network.

[Before working on this project I worked on WISDOM which was an EU funded project to develop a broadband satellite network with Matra-Marconi Space (now Astrium part of EADS) and various universities and consultancies. We mainly looked at MEO satellites, but also LEO and GEO sats. Nortels interest was in the ground segment, I worked on the general system design, uplink and downlink protocols and network control centre. Another part of Nortel worked on demonstration ground systems (breadboard level) while Astrium did the satellite breadboard. About 1998, Nortel decided not to continue with the project, partly because to take it further would require significant investment, partly because they reckoned that they could make more money by just being a ground network supplier to all the satellite networks that were being proposed at the time. - So it is easy to see my interest in these large satellite systems.]
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/21/2015 12:13 pm
You can get multi-gigabit in Ka+ bands, no? Probably easier to do that and still have a chance of powering through rain
My thought is to use ~35 GHz and up (upper edge of Ka band, maybe slightly above).
It is directly in the middle of the atmospheric window, and not used as far as I can find.
Quote
Satellite to satellite links can be millimeter wave and still have many gigabits I should think. There's a number of really strong absorption bands from water and oxygen making those bands completely useless for talking to Earth, use those.
You want inter-satellite communication to be hard to intercept.

60 GHz would be nice, right in the oxygen absorption peak. It also going to be swamped with next-gen Wifi signals (IEEE 802.11ad). It was previously used for communication between military satellites, and was the frequency Teledesic was planning to use for the same purpose.

The next big absorption peak is the 183 GHz water line. At that frequency and above it's really hard to design emitters.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/21/2015 01:12 pm
Musk's comments about latency do not preclude first sats being sat-to-ground only, just demoing those capabilities plus bus, longevity, and de-orbiting.

As for sat-to-sat, given the very large number of satellites you want to speak with and listen too, isn't optical a clear winner in terms of power and mass?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: PerW on 01/21/2015 01:15 pm
Very exiting indeed!

From: http://www.spacex.com/news/2015/01/20/financing-round
Space Exploration Technologies (SpaceX) has raised a billion dollars in a financing round with two new investors, Google and Fidelity. They join existing investors Founders Fund, Draper Fisher Jurvetson, Valor Equity Partners and Capricorn. Google and Fidelity will collectively own just under 10% of the company. 

SpaceX designs, manufactures, and launches the world's most advanced rockets and spacecraft. This funding will be used to support continued innovation in the areas of space transport, reusability, and satellite manufacturing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 01/21/2015 02:26 pm

-Awesomeness-

Well put, Matt, but could you shrink it down a wee tad? It's hugely enormous and messes with the margins. Thanks a bunch for uploading however.


I'm wondering how far the process for Google and Fidelity investing in SpaceX goes? I can imagine a lot of long term scheming over that one - Google's chieftain has always been fairly supportive of Musk and SpaceX in general, from what I remember.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: SwissCheese on 01/21/2015 02:44 pm
My thought is to use ~35 GHz and up (upper edge of Ka band, maybe slightly above).
It is directly in the middle of the atmospheric window, and not used as far as I can find.

35 GHz is being used increasingly for synthetic aperture radar systems, especially for military purposes (you can miniaturize "high frequency" radar systems such that they can be installed on medium size UAVs). Heavy use of this frequency band might not please everybody...

For the technical side, the attenuation caused by rain at 35 GHz is usually not a major issue. Only the most severe tropical rain really hampers the signal transmission.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/21/2015 03:39 pm
And the global routing table will have to be updated every few seconds. This is a hard problem.
This is BGP's hard problem though, no? It doesn't seem like SpaceX has to solve it.

There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).
That describes most routers handling large scale traffic on the internet, no? Software can't do line rate forwarding with 100 gigabit so there is no choice. The way things are scaling seems to push things farther in that direction over time, these days large datacenters are doing L3 (iBGP etc) to the top of rack.

This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees).
Unless otherwise indicated ElonSat seems like a "best effort" system.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: A12 on 01/21/2015 03:57 pm
The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.


Nowadays there are several proposal/techniques for handle this and others issues. See for example https://tools.ietf.org/html/rfc5177.

IMO, what is (slightly) more concerning, is the depletion of IPv4 address space, but, who knows, maybe this could be the 'killer application' for a real deployement of IPv6.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/21/2015 04:19 pm
Regarding routing, the sat network is dealing with a specialized problem, since it is known a-priory that you'll never go  outside of the network except for the start node and end node.  (This might have been called a compact subnet in school...  where the cost metric between any two point in the subnet is guaranteed to be lower than through any route that exits it)

So really, you only have to keep one global (and time varying) solution for best-path, and even in orbital speeds, this solution only varies in the single-second time frame, which is nothing compared to making routing decisions for individual packets.

Also, the global routing solutions can be known in advance (based on geometry and usage pattern).  So you can distribute the solutions in advance, and switch from one to the next on the precise whole-second boundary, and that's that.  It is no longer a distributed problem.

General-purpose IP is designed to operate in a very heterogeneous environment, where it is assumed that each router is a free agent.  In this case, you have absolute control over each of the satellites, so it's more like a 4000-node switch fabric then a portion of the internet.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 01/21/2015 05:28 pm
I wonder what portion was Fidelity, and what portion was Google.  Is Larry mainly giving Google's name to provide brand equity while Fidelity puts in $950 million for example?

Also, I suspect this is just opening the door to get things where they are a lot more investible as a subsidiary later.  Annnnnd... It's Elon.  3 years from now, he could provide a low interest corporate bond for $10+ billion.  If he buys a pile of them himself, people will shrug and figure it's probably safe - assuming the Gigafactory(s) and Tesla are doing well, and the satellite biz and BFR are a lot further along. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/21/2015 05:46 pm
I wonder what portion was Fidelity, and what portion was Google.  Is Larry mainly giving Google's name to provide brand equity while Fidelity puts in $950 million for example?

Also, I suspect this is just opening the door to get things where they are a lot more investible as a subsidiary later.  Annnnnd... It's Elon.  3 years from now, he could provide a low interest corporate bond for $10+ billion.  If he buys a pile of them himself, people will shrug and figure it's probably safe - assuming the Gigafactory(s) and Tesla are doing well, and the satellite biz and BFR are a lot further along.

I am curious about the apportionment as well, I was more thinking it was the other way around. Fidelity is what lends credibility both to SpaceX and Elon who may be a household name but not necessarily taken seriously in the investment world. Google has a big war chest (60G$) and they have a history of making big investments in technology that is tangential to their existing operations.  Fidelity on the other hand does not get named like this in 'exotic' ventures usually. They are investors in a wide variety of things, and often seen as astute investors (I take notice when Fidelity, Berkshire Hatheway,  Ontario Teachers Pension Plan, GE and a few other large institutional investors who seem to use similar investing logic as me) suddenly invest in something I already pay attention to.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: macpacheco on 01/21/2015 05:56 pm
And the global routing table will have to be updated every few seconds. This is a hard problem.
This is BGP's hard problem though, no? It doesn't seem like SpaceX has to solve it.

There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).
That describes most routers handling large scale traffic on the internet, no? Software can't do line rate forwarding with 100 gigabit so there is no choice. The way things are scaling seems to push things farther in that direction over time, these days large datacenters are doing L3 (iBGP etc) to the top of rack.

This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees).
Unless otherwise indicated ElonSat seems like a "best effort" system.
Much easier to just use an ethernet+vlan transport system. The ground system establish routing protocols, find other routers, when the ground sends a packet, it tells the sat system which router this is intended to (MAC address of the router). Then the whole satellite cloud would just need to function as a huge ethernet switch, implementing an rSTP protocol (wayyy simpler than BGP). VLAN is used to separate one dedicated customer from another. The whole forwarding plane must be implemented as FPGA or ASIC in order to handle a hundred Gbps of aggregate bandwidth per sat. Many large Cisco routers are implemented as FPGA, allows lower clock speeds with orders of magnitude more forwarding speed than pure software solutions.
PS: I'm a performance guy, that tends to do everything KISS. SpaceX might have more complex ideas.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/21/2015 06:13 pm
The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.


Nowadays there are several proposal/techniques for handle this and others issues. See for example https://tools.ietf.org/html/rfc5177.

IMO, what is (slightly) more concerning, is the depletion of IPv4 address space, but, who knows, maybe this could be the 'killer application' for a real deployement of IPv6.

What I was trying to describe was MPLS switching (e.g. http://www.rfc-editor.org/rfc/rfc3031.txt) and its application to ATM (e.g. http://www.rfc-editor.org/rfc/rfc3035.txt). What we did fed through into those and similar MPLS standards before 2002, when Nortel gave this up (most of the team was made redundant in 2000).

So the satellite could be something like a Label Switch Router, where every satellite is both an ingress/egress node and a transit node, where the logical transit node handles the inter-satellite links (and one link to the logical ingress/egress node). The label tables need to be updated every few seconds, but they can be precomputed in advance by the network control centre. I'm 15 years out of date (since then I've worked on speech recognition, mapping software, mobile phone software and am currently working on a IDE for an analytic database) so I'm pretty rusty on this now, but I don't think BGP is a particularly good match for a dynamic satellite network.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 01/21/2015 06:55 pm

Much easier to just use an ethernet+vlan transport system. The ground system establish routing protocols, find other routers, when the ground sends a packet, it tells the sat system which router this is intended to (MAC address of the router).

The problem with this is how does it know the MAC address of the router it is intended for. That is at the other end of the satellite network. So you need to distribute these MAC addresses. You also need to distribute the IP to router information.

An ethernet like system does not work well on a shared bandwidth uplink where none of the transmitters can hear the others. The only way they will know about collisions is when the far end sends a TCP retransmission request.

The way we solved this for WISDOM is to have a fixed frame for the uplink consisting of a large number of cells. Each ground station first requests some bandwidth on shared cells (a collision means that they don't get the bandwidth and have to request it again after some random timeout). The satellite allocates some of the uplink cells to each ground station that has requested it, they can ask for more, or relinquish it depending on their past data traffic uplink profile. These changes of bandwidth come in the cells allocated to the ground station, so there are no further collisions. We used ATM cells with a simple propriety routing protocol, there were lots of beams and few intra-satellite links, almost all the time data would flow from one of the beams to another on the same satellite. Some of the beams were dedicated to links to network interconnect ground stations which had full use of the entire beam. The satellites were basically an ATM switch, with very little software, both the ATM switching and uplink bandwidth allocation could be done entirely in hardware.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Space Ghost 1962 on 01/21/2015 07:01 pm
Google has long had significant space ambitions for years. This is an adequate vehicle for them to exploit.

Back in the late 90's with the emergence of the "dot.com's", bandwidth was priced by providers as a proportion of the "site to user" revenue chain. This hobbled Google and others in its infancy for communicating distributed updates, which were done between coasts then on a weekly basis to reduce cost. Microsoft's Jim Gray told me at the time that Microsoft mailed boxes of tapes and hard drives by parcel at that time because it was cheaper.

Then Google started a process of buying up "dark fiber", which was then extremely scarce because telco's wanted to hold-off more ISPs that were killing the ILEC/CLEC scene. They used the peerage as a secondary revenue channel to pay off the acquisitions, and started doing distributed updates much more frequently. The excess bandwidth allowed them to acquire YouTube and undercut everyone on video bandwidth pricing, using the prior mentioned bandwidth pricing model to slay competition, including those in Hollywood/media like Sony et al.

Google Earth is a GIS that can convey sensor data in real time. They can connect bandwidth and information to a wide variety of channels/consumers/needs worldwide. It's a high growth move for them, like dark fiber was.

So look at the comparison. Boeing and Lockheed have a deep pocket in the US govt for significant sat tech on orbit, with SC/LV operations for that flow. Its the primary point of ULA.

Google is the deep pocket for SpaceX next gen SC/LV operations for much larger operations than Boeing/Lockheed/ULA - by a factor of 6x! It has the potential of being the largest space business on the planet, all vertically integrated. IF it plays out.

So its does not matter if ULA holds on to US govt SC business. In other words, commercial SC business just got disrupted like LV business is being.

And in like way, if the economics for commercial launch severely undercut govt/institutional launch, the govts/institutions costs radically increase until unaffordable. Both LV's, and the SC's that make use of them!

This is a big deal! In effect that 1B is "seed" investment for a trillion dollar market bet. And Google sees it as a "no brainer".
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: macpacheco on 01/21/2015 07:02 pm

Much easier to just use an ethernet+vlan transport system. The ground system establish routing protocols, find other routers, when the ground sends a packet, it tells the sat system which router this is intended to (MAC address of the router).

The problem with this is how does it know the MAC address of the router it is intended for. That is at the other end of the satellite network. So you need to distribute these MAC addresses. You also need to distribute the IP to router information.

An ethernet like system does not work well on a shared bandwidth uplink where none of the transmitters can hear the others. The only way they will know about collisions is when the far end sends a TCP retransmission request.

The way we solved this for WISDOM is to have a fixed frame for the uplink consisting of a large number of cells. Each ground station first requests some bandwidth on shared cells (a collision means that they don't get the bandwidth and have to request it again after some random timeout). The satellite allocates some of the uplink cells to each ground station that has requested it, they can ask for more, or relinquish it depending on their past data traffic uplink profile. These changes of bandwidth come in the cells allocated to the ground station, so there are no further collisions. We used ATM cells with a simple propriety routing protocol, there were lots of beams and few intra-satellite links, almost all the time data would flow from one of the beams to another on the same satellite. Some of the beams were dedicated to links to network interconnect ground stations which had full use of the entire beam. The satellites were basically an ATM switch, with very little software, both the ATM switching and uplink bandwidth allocation could be done entirely in hardware.
MAC learning is a basic function performed by every ethernet switch. Every ethernet packet has a source MAC address, the MAC address is saved with the incoming port as the return path. The downside of this method is that when packets are sent to a unlearned/offline MAC address, it must be handled as a broadcast (sent to all nodes). But using VLAN means each VLAN has its own MAC learning space and only ports authorized for that VLAN get broadcasts.

ATM is an obsolete protocol in land based networks. The trend is everything ethernet. In many cases packets have their ethernet headers anyways, while ATM is a totally extra layer. Today most routing is already done using gigabit/10 gig/40 gig/100 gig pure ethernet links, typically running on top of WDM/DWDM optical multiplexing systems.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/21/2015 09:44 pm
RE Fidelity Investment in SpaceX

http://money.cnn.com/2015/01/21/investing/spacex-fidelity-elon-musk/ (http://money.cnn.com/2015/01/21/investing/spacex-fidelity-elon-musk/)

Ignore the video (at least as far as finding out Fidelity's involvement), read the article, Fidelity will disclose which of its mutual funds bought how much of SpaceX when they publish their month or quarter end reports.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/21/2015 09:55 pm
Much easier to just use an ethernet+vlan transport system.
Disagree, I think an ethernet type scheme would be untenable at this scale. The reliance on broadcast and the size of the forwarding tables would be deal breakers.

The whole forwarding plane must be implemented as FPGA or ASIC in order to handle a hundred Gbps of aggregate bandwidth per sat. Many large Cisco routers are implemented as FPGA, allows lower clock speeds with orders of magnitude more forwarding speed than pure software solutions.
PS: I'm a performance guy, that tends to do everything KISS. SpaceX might have more complex ideas.
Due to Moore's Law the winning trades are constantly shifting. The reason L3 schemes have pulled ahead in the last few years is due to this. The software to implement BGP is not trivial, however the hardware required to host the software is. Your phone would be overkill unless you exercise really exceptional asceticism.

The way you get the insane forwarding rates with BGP or another L3 scheme is having BGP build the forwarding tables and program them into TCAMs. The bigger you want the TCAM, the slower it is. This is why datacenters are going to L3 to the top of rack, the decision is driven by performance. I believe eg Facebook has published notes on this. The TCAM required to hold the MAC tables for the datacenter is slower than the TCAM required to hold the routing tables for the datacenter.

I think similar considerations would apply for the satellite network.

I'm pretty rusty on this now, but I don't think BGP is a particularly good match for a dynamic satellite network.
I agree with the reasoning behind this, but still think the satellites will need full global routing tables which are only defined in terms of BGP. This implies a two layer scheme.

On ingress of ground traffic the packets would be encapsulated in some scheme that's internal to the satellite network. I'm not sure any existing scheme that I know of is a good match for this, so let's say for the sake of argument that this is part of the engineering work to be done for the system. Either way the encapsulation has to be done with some knowledge of where the egress has to happen, even if that's a network not directly connected to the satellite network. This is why global routing tables are needed. Once it's in though, the packet is forwarded satellite to satellite using the internal scheme, and only decapsulated when it reaches the egress point.

So there'd be a forwarding plane that only deals with the internal satellite routing protocol, and there'd be the ingress/egress plane that knows the global routing table.

One trade I thought about would be customer equipment having the full global routing table and doing the encapsulation itself, but that's probably untenable to keep updated (and you can't trust the customer equipment in any case).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/21/2015 10:08 pm
There's no latency advantage if you just terminate the connection at a local ISP.

It seems to me that the satellite network will be the ISP.

The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.

In 2000 I worked on an internet routing project for Nortel (part of a team of 7), I did the UI and control software for the demo system. The idea is that at the edge of the network the IP packets are inspected and tagged with a series of tags and sent off to internal nodes. At the internal nodes the first tag are used to do the routing and popped (or swapped for another tag). The routes between nodes are set up by a global entity that manages bandwidth so that there is always enough bandwidth for these routes through the network. This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees). There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).  The difficult bit is then the global ('god') system that allocates bandwidth to paths through the network and tells the edge nodes how to tag packets due to their destinations. This is naively a O(n^3) problem, so doesn't scale well without tricks, for this network n=4025 so scaling is not a problem, just whether it is tractable at a size of 4025. It is easy to see how such an idea can be applied to a satellite network.

The variable size of IP packets is a problem, so on a lower level we transported them within ATM cells, and developed a cleaver way of mapping the tags onto the ATM VPI and VCI routing. Some such scheme (perhaps not using ATM but a larger cell) would be ideal for the satellite network.

[Before working on this project I worked on WISDOM which was an EU funded project to develop a broadband satellite network with Matra-Marconi Space (now Astrium part of EADS) and various universities and consultancies. We mainly looked at MEO satellites, but also LEO and GEO sats. Nortels interest was in the ground segment, I worked on the general system design, uplink and downlink protocols and network control centre. Another part of Nortel worked on demonstration ground systems (breadboard level) while Astrium did the satellite breadboard. About 1998, Nortel decided not to continue with the project, partly because to take it further would require significant investment, partly because they reckoned that they could make more money by just being a ground network supplier to all the satellite networks that were being proposed at the time. - So it is easy to see my interest in these large satellite systems.]

Is this such a hard problem? It seems to be more geometry than computer science.

Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.

If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.

For faster moving objects, like a jetliner, the network could route to the last known lat/long. If the destination is not in range, that node could then forward the request to each of its neighbouring nodes, several of which should have the plane in range. The nodes which can see the target could coordinate their info to derive a new lat/long, then propagate that to correct the routing tables.

But, this assumes dumb ground stations. A jetliner could report its GPS location, heading, and even any future course changes programmed into the autopilot. Once this info is propagated once, it might not need any updates for hours unless the plane deviates for something like bad weather.

Each sat has a good idea of the location of any ground stations in range due to the phased array. Several sats together could pinpoint something even more accurately, using a sort of GPS-in-reverse.  Any station spoofing its GPS location could be rejected from the network.

cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/21/2015 10:28 pm
Do not count any revenue from NASA on this. Maybe if they manage to snag something for the Mars part that NASA has a request out for.

ISS has a 300mbps down / 25mbps up appetite for bandwidth today, that is currently accommodated via ku-band to TDRS sats in GEO. (On general principles, it will probably need another speed boost eventually):-
http://www.nasaspaceflight.com/2013/04/iss-communications-overhaul-boost-scientific-output/

ISS is well below this constellation, and ISTM it could utilise it as long as the sats can accommodate a "ground" station moving that quickly. (See my previous post about fast-moving objects providing position and velocity state to the network.) Perhaps the comms module could even be a peer on the network, IE just another sat in the constellation.

I don't know if other LEO sats might find it useful to just buy a commodity comms module which interfaces into the constellation. If they currently use ground stations for comms, that would avoid that requirement. It might be one way that this reduces the design and operations costs for many sats in the future.

ISTM that a constellation of near-real-time ground-observation sats would have huge appetite for bandwidth. Unless these sats themselves might also include Earth observation - Google might find a use for that.

cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 01/21/2015 10:50 pm
I wonder what portion was Fidelity, and what portion was Google.  Is Larry mainly giving Google's name to provide brand equity while Fidelity puts in $950 million for example?

Google has a 7.5% share in SpaceX (source (http://www.benchmarkmonitor.com/tesla-motors-inc-nasdaqtsla-founders-spacex-got-investment-of-1-billion-from-google-inc-nasdaqgoogl-fidelity/183988/)), with Fidelity presumably therefore having 2.5%.

Fidelity at least will want an ultimate exit from the company, but that can easily be accomplished by (as suggested) establishing the satellite operation as a publicly-listed subsidiary.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/21/2015 10:51 pm
There's no latency advantage if you just terminate the connection at a local ISP.

It seems to me that the satellite network will be the ISP.

The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.

In 2000 I worked on an internet routing project for Nortel (part of a team of 7), I did the UI and control software for the demo system. The idea is that at the edge of the network the IP packets are inspected and tagged with a series of tags and sent off to internal nodes. At the internal nodes the first tag are used to do the routing and popped (or swapped for another tag). The routes between nodes are set up by a global entity that manages bandwidth so that there is always enough bandwidth for these routes through the network. This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees). There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).  The difficult bit is then the global ('god') system that allocates bandwidth to paths through the network and tells the edge nodes how to tag packets due to their destinations. This is naively a O(n^3) problem, so doesn't scale well without tricks, for this network n=4025 so scaling is not a problem, just whether it is tractable at a size of 4025. It is easy to see how such an idea can be applied to a satellite network.

The variable size of IP packets is a problem, so on a lower level we transported them within ATM cells, and developed a cleaver way of mapping the tags onto the ATM VPI and VCI routing. Some such scheme (perhaps not using ATM but a larger cell) would be ideal for the satellite network.

[Before working on this project I worked on WISDOM which was an EU funded project to develop a broadband satellite network with Matra-Marconi Space (now Astrium part of EADS) and various universities and consultancies. We mainly looked at MEO satellites, but also LEO and GEO sats. Nortels interest was in the ground segment, I worked on the general system design, uplink and downlink protocols and network control centre. Another part of Nortel worked on demonstration ground systems (breadboard level) while Astrium did the satellite breadboard. About 1998, Nortel decided not to continue with the project, partly because to take it further would require significant investment, partly because they reckoned that they could make more money by just being a ground network supplier to all the satellite networks that were being proposed at the time. - So it is easy to see my interest in these large satellite systems.]

Is this such a hard problem? It seems to be more geometry than computer science.

Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.

If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.

For faster moving objects, like a jetliner, the network could route to the last known lat/long. If the destination is not in range, that node could then forward the request to each of its neighbouring nodes, several of which should have the plane in range. The nodes which can see the target could coordinate their info to derive a new lat/long, then propagate that to correct the routing tables.

But, this assumes dumb ground stations. A jetliner could report its GPS location, heading, and even any future course changes programmed into the autopilot. Once this info is propagated once, it might not need any updates for hours unless the plane deviates for something like bad weather.

Each sat has a good idea of the location of any ground stations in range due to the phased array. Several sats together could pinpoint something even more accurately, using a sort of GPS-in-reverse.  Any station spoofing its GPS location could be rejected from the network.

cheers, Martin

Thinking about this some more, each sat has a bunch of active ground stations that it's talking to. It will have to do a handoff to another sat as it gets near the station's horizon, but I wonder how often it can just hand over to the following sat in the same plane? That could make the coordination somewhat simpler.

As the earth rotates under the constellation, there will come a point where the ground station has to be handed off to the adjacent plane of sats to the East.

There may be reasons for a ground station to switch "randomly" between any of the sats that it sees in its sky, but maybe just "next in plane", "hop Eastwards", and "hop Westwards" (for load balancing) is all that is needed?

cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/21/2015 10:57 pm
[...]
For lack of any other information, I am assuming 64 planes x 64 satellites, and about 500 kg per each.
[...]
[...]splitting 4025 into multipliers gives the following viable solution: 81x25... 81 satellites in 25 planes, perhaps 1 is spare[...]

Yes, of course, that was just to get a quick idea, starting with polar...
Everything else stays....

I was just about to ask why not 115 x 35, or 161 x 25, or 175 x 23. (Musk did say this number was probably overprecise.)

IIUC, each different plane has to operate at a different altitude since they cross each other, so that might be a reason to keep the number of planes down.

Twice as many planes may reduce the inter-plane distance to half, which makes it that much easier for one sat to stray up or down and cause a nasty collision.

cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/21/2015 11:01 pm
There's no latency advantage if you just terminate the connection at a local ISP.

It seems to me that the satellite network will be the ISP.

The satellites will have to know the global routing table and know which satellites are near which ground stations.

And the global routing table will have to be updated every few seconds. This is a hard problem.

In 2000 I worked on an internet routing project for Nortel (part of a team of 7), I did the UI and control software for the demo system. The idea is that at the edge of the network the IP packets are inspected and tagged with a series of tags and sent off to internal nodes. At the internal nodes the first tag are used to do the routing and popped (or swapped for another tag). The routes between nodes are set up by a global entity that manages bandwidth so that there is always enough bandwidth for these routes through the network. This means that when a packet is tagged at the network it is guaranteed to reach the far end (with quality of service guarantees). There is also only limited and bounded queuing at internal nodes, the small tag and limited queuing means that the switching can be performed all in hardware, with software only used to update the tag to route map(s).  The difficult bit is then the global ('god') system that allocates bandwidth to paths through the network and tells the edge nodes how to tag packets due to their destinations. This is naively a O(n^3) problem, so doesn't scale well without tricks, for this network n=4025 so scaling is not a problem, just whether it is tractable at a size of 4025. It is easy to see how such an idea can be applied to a satellite network.

The variable size of IP packets is a problem, so on a lower level we transported them within ATM cells, and developed a cleaver way of mapping the tags onto the ATM VPI and VCI routing. Some such scheme (perhaps not using ATM but a larger cell) would be ideal for the satellite network.

[Before working on this project I worked on WISDOM which was an EU funded project to develop a broadband satellite network with Matra-Marconi Space (now Astrium part of EADS) and various universities and consultancies. We mainly looked at MEO satellites, but also LEO and GEO sats. Nortels interest was in the ground segment, I worked on the general system design, uplink and downlink protocols and network control centre. Another part of Nortel worked on demonstration ground systems (breadboard level) while Astrium did the satellite breadboard. About 1998, Nortel decided not to continue with the project, partly because to take it further would require significant investment, partly because they reckoned that they could make more money by just being a ground network supplier to all the satellite networks that were being proposed at the time. - So it is easy to see my interest in these large satellite systems.]

Is this such a hard problem? It seems to be more geometry than computer science.

Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.

If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.

For faster moving objects, like a jetliner, the network could route to the last known lat/long. If the destination is not in range, that node could then forward the request to each of its neighbouring nodes, several of which should have the plane in range. The nodes which can see the target could coordinate their info to derive a new lat/long, then propagate that to correct the routing tables.

But, this assumes dumb ground stations. A jetliner could report its GPS location, heading, and even any future course changes programmed into the autopilot. Once this info is propagated once, it might not need any updates for hours unless the plane deviates for something like bad weather.

Each sat has a good idea of the location of any ground stations in range due to the phased array. Several sats together could pinpoint something even more accurately, using a sort of GPS-in-reverse.  Any station spoofing its GPS location could be rejected from the network.

cheers, Martin

Thinking about this some more, each sat has a bunch of active ground stations that it's talking to. It will have to do a handoff to another sat as it gets near the station's horizon, but I wonder how often it can just hand over to the following sat in the same plane? That could make the coordination somewhat simpler.

As the earth rotates under the constellation, there will come a point where the ground station has to be handed off to the adjacent plane of sats to the East.

There may be reasons for a ground station to switch "randomly" between any of the sats that it sees in its sky, but maybe just "next in plane", "hop Eastwards", and "hop Westwards" (for load balancing) is all that is needed?

cheers, Martin

Such schemes are easier for a human to wrap their brain around, since they are "intuitive", but for a computer, it's all the same.

All that matters is that under central control, at some time T0, everyone switches to a new switching table, which was distributed in advance, and clearly anticipates any geometry changes.  All packets that enter the system at time T>T0 use the new tables.  I expect changes to occur about once every 30 seconds or so.  (Just looking at 5400 seconds orbital period, 80 sats per plane (as has been proposed upthread) and doing it every half-interval). 

The updates can also occur every single second and still be glacially slow compared with the rate at which data packets are handled.

This is very different than a bunch of independently managed routers that try to reach a routing solution in a distributed manner.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/22/2015 05:10 am
All that matters is that under central control, at some time T0, everyone switches to a new switching table, which was distributed in advance, and clearly anticipates any geometry changes.  All packets that enter the system at time T>T0 use the new tables.  I expect changes to occur about once every 30 seconds or so.  (Just looking at 5400 seconds orbital period, 80 sats per plane (as has been proposed upthread) and doing it every half-interval).
I think that probably doesn't work as different people will have different weather blowing through and have different obstructions to deal with.

Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.
I think this also probably doesn't work as lat/long won't allow you to infer a priori which satellite can reach someone, for similar reasons.

If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.
The volume of data seems quite large though - probably hundreds of millions of stations changing every few seconds. That gets untenable fast. And updates would still have a race where they would be behind some of the packets.

All this state information seems like something you'd want to keep localized to a few satellites, the ones that would be the possible handoff candidates. They can forward the traffic if they receive it erroneously and send a redirect message to the source satellite (like an ICMP redirect).

Can probably also parcel up the Earth into IPv6 prefixes well enough to let you generate a set of possible satellites for a given address. IPv4 can be supported via RFC6877 (meaning ignored entirely in the satellites).

This scheme wouldn't directly handle mobile stations, but that can be separately handled a number of ways. A special set of prefixes could be set aside for stations known to be mobile and SpaceX could market that at a premium and actually do the work to track them (along with the global routing tables). A few hundred thousand of those would probably work. A few hundred million, probably not. Another option would be a VPN type system to allow the wifi on the plane to not have to renumber every few minutes by tunneling somewhere else.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/22/2015 06:41 am
All that matters is that under central control, at some time T0, everyone switches to a new switching table, which was distributed in advance, and clearly anticipates any geometry changes.  All packets that enter the system at time T>T0 use the new tables.  I expect changes to occur about once every 30 seconds or so.  (Just looking at 5400 seconds orbital period, 80 sats per plane (as has been proposed upthread) and doing it every half-interval).
I think that probably doesn't work as different people will have different weather blowing through and have different obstructions to deal with.

Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.
I think this also probably doesn't work as lat/long won't allow you to infer a priori which satellite can reach someone, for similar reasons.

If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.
The volume of data seems quite large though - probably hundreds of millions of stations changing every few seconds. That gets untenable fast. And updates would still have a race where they would be behind some of the packets.

All this state information seems like something you'd want to keep localized to a few satellites, the ones that would be the possible handoff candidates. They can forward the traffic if they receive it erroneously and send a redirect message to the source satellite (like an ICMP redirect).

Can probably also parcel up the Earth into IPv6 prefixes well enough to let you generate a set of possible satellites for a given address. IPv4 can be supported via RFC6877 (meaning ignored entirely in the satellites).

This scheme wouldn't directly handle mobile stations, but that can be separately handled a number of ways. A special set of prefixes could be set aside for stations known to be mobile and SpaceX could market that at a premium and actually do the work to track them (along with the global routing tables). A few hundred thousand of those would probably work. A few hundred million, probably not. Another option would be a VPN type system to allow the wifi on the plane to not have to renumber every few minutes by tunneling somewhere else.

Yes, there's that too.   

The predictable situation I described holds only for sat-to-sat orbital traffic.  You still have to choose the best end nodes, and the selection process involves the ground nodes, and so yes, it gets complicated.

I didn't mean to imply this is simple, just to imply that figuring out the sat-to-sat route, even though the geometries change very rapidly, is actually IMO the simpler of the issues they face.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 01/22/2015 12:19 pm


Most of the ground stations will be static, or nearly so (a car doesn't move far vs orbital speeds). If you want to talk to an IP address, this can be translated to a lat/long, then calculate the best routing to the least busy sat that's visible to that location. The system needs to maintain a table of IP vs lat/long, but most of these won't change very rapidly.
I think this also probably doesn't work as lat/long won't allow you to infer a priori which satellite can reach someone, for similar reasons.

I take your point re dynamically changing obstructions (eg urban canyons). Areas with a lot of tall buildings may be untenable for this service, but then it should be able to fall back to 4G / 5G.

However, lat/long will still identify the set of sats which are above the horizon for a ground station. Assuming each sat has a direct link to its near neighbours, they could coordinate, moment-to-moment which one has the baton, and forward any packets to that sat.

The long-range routing might well then become "get this packet to any sat that's above the horizon for this lat/long", and then that sat can forward to the one with the baton.


If 4,000 sats each have a direct link to 25 other sats that's 100,000 links. ISTM this state info could be broadcast across the network quite easily - 1mbps reserved on each link would propagate the info in seconds to every other node, I'd think.
The volume of data seems quite large though - probably hundreds of millions of stations changing every few seconds. That gets untenable fast. And updates would still have a race where they would be behind some of the packets.

No, this is a fundamental difference between terrestial networks and this sat network.

Terrestrial networks have machines at arbitrary locations, with very complex networks of routers also at arbitrary locations.

A fixed ground station never needs to have its lat/long updated, the sats just need to figure out between themselves which is the best route to get to that place, given the known location of every intermediate hop. That's a lot of data to remember, but it's essentially static, at least for the fixed ground stations.

The route will change continuously, but everything except traffic loads on the sat-to-sat links can just be calculated from "static" data.

"Static" in quotes, because the ephemeris / almanac ( http://en.m.wikipedia.org/wiki/GPS_signals ) will need occasional updates.

Cheers, Martin

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MechE31 on 01/22/2015 01:48 pm
For those asking who put how much, http://blogs.wsj.com/digits/2015/01/21/spacexs-valuation-rockets-to-12-billion-with-google-investment/

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 01/22/2015 03:07 pm
As somebody who was there for the birth of digital cellular, then CDMA through 3-4G I really hate it when folks conflate apples and oranges.   Thinking multiple protocol stacks and within the RF one multiple stacks is the best way to look at it.   The IP packets don't need to know anything about the various space packets.

Some of this comes from the primitive hard handoff concepts that early protocols like GSM had.
Perhaps CDMA with its connections to multiple bases(s) or more accurately Base-Carrier-Sector-WalshCodes is a more appropriate to this multitude of satellites in view for a short period.     

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   

In the first phase of the rollout you could have a CBSC style controller that covers areas of the sky.    In following rollouts the more complex options for work and such with  individual packets and the site registrations just like was done in Cellular.    This speaking as an inventor of one of those low level entry to a call protocols.   

If you want to consider it as an ISO-Layer because you have to - consider all this taking place at the MAC layer or (Zeroth) layer.   (think MAC hardware electric and collision layer on your physical LAN Ethernet)

Again the IP layer does NOT need to know anything about the radio layers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 01/22/2015 03:14 pm
As somebody who was there for the birth of digital cellular, then CDMA through 3-4G I really hate it when folks conflate apples and oranges.   Thinking multiple protocol stacks and within the RF one multiple stacks is the best way to look at it.   The IP packets don't need to know anything about the various space packets.

Some of this comes from the primitive hard handoff concepts that early protocols like GSM had.
Perhaps CDMA with its connections to multiple bases(s) or more accurately Base-Carrier-Sector-WalshCodes is a more appropriate to this multitude of satellites in view for a short period.     

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   

In the first phase of the rollout you could have a CBSC style controller that covers areas of the sky.    In following rollouts the more complex options for work and such with  individual packets and the site registrations just like was done in Cellular.    This speaking as an inventor of one of those low level entry to a call protocols.   

If you want to consider it as an ISO-Layer because you have to - consider all this taking place at the MAC layer or (Zeroth) layer.   (think MAC hardware electric and collision layer on your physical LAN Ethernet)

Again the IP layer does NOT need to know anything about the radio layers.


Was about top post a question about this. Why does the satellite network need to work the same as the ground network? At the uplink, tell the packet which physical direction it needs to go to. First satellite knows its closet neighbours (wild assumption, but should be fairly easy to implement - these things don;t move randomly). Sends the packet to the one in the right direction. Continues like that until reaches the destination satellite, where is downlinks and continues using terrestrial addressing.

Or, the uplink can work out the route through the satellite network before uplinking the pack, and it routes itself according to that table. Thereby only basestations need to have routing tables, satellite just need to know their and their nearest neighbour locations, and maybe not even that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 01/22/2015 03:52 pm
As somebody who was there for the birth of digital cellular, then CDMA through 3-4G I really hate it when folks conflate apples and oranges.   Thinking multiple protocol stacks and within the RF one multiple stacks is the best way to look at it.   The IP packets don't need to know anything about the various space packets.

Some of this comes from the primitive hard handoff concepts that early protocols like GSM had.
Perhaps CDMA with its connections to multiple bases(s) or more accurately Base-Carrier-Sector-WalshCodes is a more appropriate to this multitude of satellites in view for a short period.     

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   

In the first phase of the rollout you could have a CBSC style controller that covers areas of the sky.    In following rollouts the more complex options for work and such with  individual packets and the site registrations just like was done in Cellular.    This speaking as an inventor of one of those low level entry to a call protocols.   

If you want to consider it as an ISO-Layer because you have to - consider all this taking place at the MAC layer or (Zeroth) layer.   (think MAC hardware electric and collision layer on your physical LAN Ethernet)

Again the IP layer does NOT need to know anything about the radio layers.


Was about top post a question about this. Why does the satellite network need to work the same as the ground network? At the uplink, tell the packet which physical direction it needs to go to. First satellite knows its closet neighbours (wild assumption, but should be fairly easy to implement - these things don;t move randomly). Sends the packet to the one in the right direction. Continues like that until reaches the destination satellite, where is downlinks and continues using terrestrial addressing.

Or, the uplink can work out the route through the satellite network before uplinking the pack, and it routes itself according to that table. Thereby only basestations need to have routing tables, satellite just need to know their and their nearest neighbour locations, and maybe not even that.

It's likely that the user's station will need location and accurate time just to transmit and aim its active phased array antenna and send the registration messages with power strength measurements to the system (akin to the constant PSM messages your cellphone puts out -when powered up.   If the system didn't have these messages it wouldn't know where to send a ring message when you have a call.   It's also how your movements can be tracked by the powers that be.)

Look at it this way.   Each sat will be putting out pilot signals. These are evaluated by each user device as power strength measurements, timing, location of the each sat, etc.  and provide oodles of both direct and derived information needed by the user device to figure out when and where it wants to request resources from or transmit.     Imaginary resources (base carrier sector) are useful too but will get you in hot water with the spy services.   Why?  Because it's useful to idle a high bandwidth connection in imaginary resources than spend the overhead (LAG) creating a new connection.  But, imaginary resources can't be spied upon by 3rd parties so they will fight them tooth and nail. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 01/22/2015 04:36 pm
For the IP part of your internet connection think RADIUS server for intellectually simplicity.

Now the radius server can be a physical device that brings all sorts of extra lag with it.
First gen it might be in a geosynchronous sat until the rest of the system kinks are worked out, however only first gen due to the lag.

In later generations the Radius sever could be just a context space  in the satellite cloud and perhaps it would be a virtual application context handed off to sats as they enter and leave a serviced region.   
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/22/2015 05:05 pm
Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.
How straightforward is it for the phased array antenna to have an open session with multiple satellites?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 01/22/2015 05:21 pm
Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.
How straightforward is it for the phased array antenna to have an open session with multiple satellites?

Very easy.  They'd have a dedicated channel for it, but they may not need it.  If the sats and the terrestrial handsets are talking to each other via similar open sessions, the phased array can track each of the multiple sessions and even hand off between sectors no sweat.

There are however bounds on this.  The processing power and transmit power have to be there, and both get limited as the number of sessions increases.  Current terrestrial equivalents can handle thousands of sessions before more equipment is needed.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 01/22/2015 07:57 pm
Considering the multiple sats and short time on station I think CDMA cellular (both infrastructure and subscriber units) would be a good starting point for the software.

A useful concurrent project would be to port a radius server to an SCTP or similar distributed object protocol, perhaps with the master server images on geosynchronous sats.   

The man who ported SCTP to Linux is currently a top Google engineer so it's a match.

Alternative is modify Reliable Server Pooling (RSerPool)

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/22/2015 08:57 pm

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   


And so this constellation is the same thing, only without the spider, yes?

(Sorry)

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 01/22/2015 09:18 pm

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   


And so this constellation is the same thing, only without the spider, yes?

(Sorry)
CDMA has the spider.
The CDMA phone watches between 20 - 60 beacons.
Most CDMA handoffs are not hard but rather some degree of soft/softer.     
If you pay for more quality (yes most carriers don't openly offer it) the spider can have up to say 18 legs.
Typical is 3 or 4 legs and the phone or bases just grab the best (or first) packets.     Individual protocol up a leg is CDMA proprietary and has retries but if the phone or land side gets it better on another leg... who cares.  It can be done voting wise so you are always sure of the result.

It got well executed by bullet trains in Japan.   Many were in cell coverage regions less than 30 seconds so hard handoffs would have had huge dropped call rates for the train passengers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: king1999 on 01/23/2015 12:47 am
Elon's SkyNet will do to the fiber what the cellular network did to the landline. Period.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/23/2015 02:14 am
Very easy.  They'd have a dedicated channel for it, but they may not need it.  If the sats and the terrestrial handsets are talking to each other via similar open sessions, the phased array can track each of the multiple sessions and even hand off between sectors no sweat.
That simples things up. If location can be inferred from the address then destination satellites can be calculated a priori, with forwarding used in case of obstruction on the path from a particular satellite to the customer equipment (big tree or whatever).

There are however bounds on this.  The processing power and transmit power have to be there, and both get limited as the number of sessions increases.  Current terrestrial equivalents can handle thousands of sessions before more equipment is needed.
Do you think the satellites would go with phased array or just a bunch of directional beams? The beams could be pretty narrow at the higher wavelengths. I think the customer equipment can probably bring pretty substantial computing resources to bear, it'll be the satellites that are constrained.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/23/2015 03:26 am

Think of the end user having spider-legs running from it to the various satellites and it slowly walks its legs from sat to sat one careful leg at a time always keeping several on the various sats.   


And so this constellation is the same thing, only without the spider, yes?

(Sorry)
CDMA has the spider.
The CDMA phone watches between 20 - 60 beacons.
Most CDMA handoffs are not hard but rather some degree of soft/softer.     
If you pay for more quality (yes most carriers don't openly offer it) the spider can have up to say 18 legs.
Typical is 3 or 4 legs and the phone or bases just grab the best (or first) packets.     Individual protocol up a leg is CDMA proprietary and has retries but if the phone or land side gets it better on another leg... who cares.  It can be done voting wise so you are always sure of the result.

It got well executed by bullet trains in Japan.   Many were in cell coverage regions less than 30 seconds so hard handoffs would have had huge dropped call rates for the train passengers.

Sorry, I was referencing that old explanation (http://quoteinvestigator.com/2012/02/24/telegraph-cat/) of how the telegraph works.  :s/dog/spider/g
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Zed_Noir on 01/23/2015 04:20 am
Elon's SkyNet will do to the fiber what the cellular network did to the landline. Period.

No network of Musk will be call SkyNet  :o, more likely NetX.

Besides there is already a SkyNet in operation with the British MoD.  ;D
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Jakusb on 01/23/2015 04:58 pm
Just wondering what all the discussion is all about...
So, the 4000+ sats are not stationary, right? So ground stations need to keep track of multiple to keep a link and do handover in time.
I guess they will be flying in formation? So at any moment it can be determined where the receiving party will be, so which (group of) sat(s) to send the package. If a group is used, all can broadcast the package redundantly if needed?!
Also Elon was very explicit that the receiver needs to be in direct line of sight. Not rooftop penetrating....
So trees and such are not in question and not something they need to take into account. It is for the receiver to 'manage' trees, likely by not positioning under one. ;p
If a receiver would be on a (fast) moving platform, would this even be supported?
How fast would it need to move, before its roaming becomes too fast?
Just curious. :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/23/2015 06:49 pm
There's no need for mechanical tracking if they use phased array antennae and several satellites are overhead. They can electronically aim them. They've already mentioned this much.

Simplifying this could be the recent development of chip-level phased arrays. Commercialize that and a shoebox sized antenna box with a few arrays gives you most of the sky coverage. Below is prototype array chip.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/23/2015 07:07 pm
There's no need for mechanical tracking if they use phased array antennae and several satellites are overhead. They can electronically aim them. They've already mentioned this much.

Simplifying this could be the recent development of chip-level phased arrays. Commercialize that and a shoebox sized antenna box with a few arrays gives you most of the sky coverage. Below is prototype array chip.
And googling this you see a lot of activity around their applications in eg automotive radar etc at extremely short wavelengths (W-band). Meaning there's a lot of work happening totally regardless of what SpaceX does or doesn't do. Plenty of companies to acquire, PHD's to hire, and off the shelf tech likely to be available by the time SpaceX wants to actually start launching things.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RedLineTrain on 01/23/2015 07:09 pm
Peter de Selding at Space News has a very informative article about how ITU filings work and a rundown of the filings that were submitted by various players.  It appears that not much is known publicly about which is SpaceX's horse at this time.

http://spacenews.com/signs-of-satellite-internet-gold-rush/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: A12 on 01/23/2015 07:27 pm
Interesting discussion ongoing.

Just let me add a little bit of information: the more advanced ethernet switches are not anymore relying on the old concept of MAC learning, broadcast, spanning tree protocol and so on.
The most important vendors are now using TRILL or some other more advanced concept.

For a very good introductory article on TRILL see: http://www.cisco.com/web/about/ac123/ac147/archived_issues/ipj_14-3/143_trill.html

Curiously, who has developed this new ideas is the same person that more forty years ago invented the STP.

Finally, take care with ethernet switches: at least the terrestrial ones are quite power hungry devices.
I don't know if that same technologies, as they are, can be used on satellites;
let alone they are built using ASIC and FPGA with an high scale of integration, that make me doubtful about  survivability in space.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/23/2015 08:36 pm
http://qz.com/330537/claims-about-the-boom-in-satellite-internet-schemes-in-order-of-apparent-implausibility/

Quote
>
Update, (1/23): While the International Telecommunications Union had described Wyler’s firm and Google as joint rights-holders to a block of satellite radio spectrum, Wyler tells Quartz that his company fully owns the spectrum and has no relationship with Google. As to possible arrangements to share that spectrum with the Google/SpaceX enterprise, Wyler says “my crystal ball is broken. Nothing is off the table, the mission is to connect people.”
>
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 01/23/2015 11:39 pm
Peter de Selding at Space News has a very informative article about how ITU filings work and a rundown of the filings that were submitted by various players.
After some testing, I've posted a reddit tutorial on how to find and view ITU filings:
https://www.reddit.com/r/spacex/comments/2tgh4m/itu_filing_tutorial_rumored_spacex_steam/

Anyways, here are the frequencies for some of the constellations mentioned by Space News:

STEAM-1
10.7-12.75 GHz
12.75-13.25
13.75-14.5
STEAM-2
17.7-18.6
17.3-17.7
27.5-30.0
18.8-20.2
STEAM-0
10.7-12.75
12.75-13.25
13.75-14.5
17.7-18.6
18.8-20.2
17.3-17.7

MCSAT-LEO:
(http://abload.de/img/mcsatleo-fqrqcb7.jpg)
MC2SAT-LEO:
(http://abload.de/img/mc2satleo-fq22e08.jpg)
MC3Sat:
(http://abload.de/img/mcsat3-fqtrzea.jpg)

As a reference, here is the list of OneWeb filings.
goo.gl/BkmmSk (http://goo.gl/BkmmSk)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 01/25/2015 07:56 am
Peter de Selding at Space News has a very informative article about how ITU filings work and a rundown of the filings that were submitted by various players.  It appears that not much is known publicly about which is SpaceX's horse at this time.

http://spacenews.com/signs-of-satellite-internet-gold-rush/
Perhaps there's some launch price elasticity after all.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 01/25/2015 02:21 pm
Peter de Selding at Space News has a very informative article about how ITU filings work and a rundown of the filings that were submitted by various players.  It appears that not much is known publicly about which is SpaceX's horse at this time.

http://spacenews.com/signs-of-satellite-internet-gold-rush/
Perhaps there's some launch price elasticity after all.
However, it only costs $40,000 to start the ball rolling and file the request that blocks anyone from using interfering bands until your filing fails to meet the continuance requirements.  This is in analogous to mineral leases that oil and gas companies vie for. I would argue that the price needs to go up to weed out frivolous, but strategic, applications. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 01/25/2015 02:36 pm
Peter de Selding at Space News has a very informative article about how ITU filings work and a rundown of the filings that were submitted by various players.  It appears that not much is known publicly about which is SpaceX's horse at this time.

http://spacenews.com/signs-of-satellite-internet-gold-rush/
Perhaps there's some launch price elasticity after all.
However, it only costs $40,000 to start the ball rolling and file the request that blocks anyone from using interfering bands until your filing fails to meet the continuance requirements.  This is in analogous to mineral leases that oil and gas companies vie for. I would argue that the price needs to go up to weed out frivolous, but strategic, applications.

The system is faulty if it allows a company to block someone/company/organization with a viable product simply by putting down 40k; thereby tying up the spectrum for years.  First come first served (get it into orbit and transmitting), after fees are paid and proposal accepted, should be the policy with revocation of spectrum in the offing if there is no or little follow through.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 01/25/2015 05:00 pm
From another thread this is a item that has a possible design varriation for the sat network.


The fundamental change is when it is a cheap as building a stadium ($1.5B say) and that there are several other 'national' stations as well as a few commercial ones (and not just hotels).  Right now there is no compelling reason to put up a 'national' station - but there is political reason to put up a competing 'international' station which China is doing (see ESA takes steps to send astronaut to Chinese space station (http://spacenews.com/esa-hikes-budget-takes-steps-to-send-astronaut-to-chinese-space-station/).  If we assume that we have a man rated launch capability at $20M per person, and FHR1 gross cargo rates at $1,000/kg to LEO - then building a station with the capacity of living and working space and power budget of the ISS becomes the cost of a stadium or a high rise office building. ($500M to launch, $1B for constructing and initial provisioning, $300M a year to keep six people there continuously without the 'science' budget).  Say that is 5 years from now, by 10, it could easily be half that price if the same general elements are borrowed from the first implementation.  I wonder how much it cost to build, and how much it costs to operate the indoor ski hill in Dubai?

http://forum.nasaspaceflight.com/index.php?topic=36384.msg1320788#msg1320788 (http://forum.nasaspaceflight.com/index.php?topic=36384.msg1320788#msg1320788)

A station or other satelite could link in to the in-space links and then recieve a very high data rate connection continuously anywhere in orbit.  For Dargon operations this would be an advantage because SpaceX would no longer need TDRS. 

By using very high frequencies >64GHz in the high atmospheric absorption bands where the ITU currently does not manage or require fillings, these inspace links could support not only the sat to sat links at high throughput but thousands of satelites and stations (even support ISS operations past 2020 causing TDRS to become obsolete) with continuous internet connections.

What this means to LEO sat operators is the cost of communications at a high data rate 1Gb or greater becomes trivial $300/month or $3,600/year. This is compared to the 10's or 100's of thousands of dollars for less than continuous communications currently available to LEO operators.

Another note about TDRS is that these sats are very expensive to put in place $1B+ each and then cost $M's/year to operate.  NASA operations could reduce spending on the costs associated with operations due to  communications. 4 sats at total of $6B divided by operational life of 15 years is $400M/year plus $10-50M to operate the satelites and thier dedicated downlink centers or a charge of ~$450M/yesr for worldwide communications that could be replaced with a system that would charge them for the same basic services at $1M or less per year.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 01/26/2015 03:18 pm
There are however bounds on this.  The processing power and transmit power have to be there, and both get limited as the number of sessions increases.  Current terrestrial equivalents can handle thousands of sessions before more equipment is needed.
Do you think the satellites would go with phased array or just a bunch of directional beams? The beams could be pretty narrow at the higher wavelengths. I think the customer equipment can probably bring pretty substantial computing resources to bear, it'll be the satellites that are constrained.

Phased array is pretty common now and the controlling software is many generations into maturity.  Beam steering using a solid-state phased-array ("smart antenna") gives so many power utilization advantages that it'd be inconceivable to me that they wouldn't use one in favor of a "dumb" antenna system for a system where both ends of the connection are in motion.   

I also expect it will be a FDD system instead of a TDD system.  The distances between satellites are too great for TDD to be of any use (the communications window would close before many packets could be sent).  http://electronicdesign.com/communications/what-s-difference-between-fdd-and-tdd  In practice when I was deploying TDD systems, they were unusable beyond 18 miles.  The alternating send/receive communications windows can certainly be lengthened, but the objective of duplexing your radios is to allow two-way communications, so the window has to be kept short enough to still allow a two-way conversation.

Now why did I bring this up?  Because FDD systems are inherently more complex and lend themselves less to beamforming (though they can and do), and more importantly, they are huge bandwidth hogs compared to TDD.  And why does that matter?  Because both SpaceX and OneWeb are going to need bandwidth to operate their systems.  An FDD system requires that they have more bandwidth and more complicated bird than a TDD system.  That in turn begs the all important question, what bandwidth are they going to use, and where are they going to get it?

Keep in mind that bandwidth "ownership" or licensing of the same frequencies varies from country to country, and also each country varies the acceptable uses of frequencies according to its own laws and regulations.  There is no common usage or ownership- though many countries have similar and/or identical usage plans and treaties in place. 

So that means, bear with me, that both satellite systems will have to use unlicensed spectrum (which also varies from country to country) or they will need to somehow convince every country they fly over to not only allow them to license common spectrum (unlikely to the point of laughability), but also change their usage regulations to allow the same chunk of spectrum to be used for their system in every country.  An FDD system needs more spectrum, so the problem gets a bit larger.

Now, there are some alternatives.  First, either system could use the "white spaces" in licensed spectrum as Google has been proposing for a decade, but the owners of that unused spectrum simply will never allow it in this country or any other without due compensation.  Second, and more likely IMHO, is that both proposed systems are shopping around for currently utilized satellite spectrum from a fellow satellite vendor. 

Why you ask?  Because the heavy lifting of setting aside specific chunks of licensed spectrum for use across borders has already been done by these entities.  It's in use by different applications right now, but that can always change, and the different applications may even be able to coincide if the white spaces are used.

In other words, don't build when you can buy.  If I were a betting man, I'd purchase stock in Irridium or Sirius/XM (just to name two).  Their phones have undoubtedly been ringing with requests to license the spectrum outright, or utilize their white space- IMHO.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 01/26/2015 03:38 pm
There are however bounds on this.  The processing power and transmit power have to be there, and both get limited as the number of sessions increases.  Current terrestrial equivalents can handle thousands of sessions before more equipment is needed.
Do you think the satellites would go with phased array or just a bunch of directional beams? The beams could be pretty narrow at the higher wavelengths. I think the customer equipment can probably bring pretty substantial computing resources to bear, it'll be the satellites that are constrained.

Phased array is pretty common now.  Beam steering using a solid-state phased-array ("smart antenna") gives so many power utilization advantages that it'd be inconceivable to me that they wouldn't use one in favor of a "dumb antenna system" on a system where both ends of the connection are in motion.   

I also expect it will be a FDD system instead of a TDD system.  The distances between satellites are too great for TDD to be of any use (the communications window would close before many packets could be sent).  http://electronicdesign.com/communications/what-s-difference-between-fdd-and-tdd  In practice when I was deploying TDD systems, they were unusable beyond 18 miles.  The alternating send/receive communications windows can certainly be lengthened, but the objective of duplexing your radios is to allow two-way communications, so the window has to be kept short enough to still allow a two-way conversation.

Now why did I bring this up?  Because FDD systems are inherently more complex and lend themselves less to beamforming (though they can and do), and more importantly, they are huge bandwidth hogs compared to TDD.  And why does that matter?  Because both SpaceX and OneWeb are going to need bandwidth to operate their systems.  An FDD system requires that they have more bandwidth and more complicated bird than a TDD system.  That in turn begs the all important question, what bandwidth are they going to use, and where are they going to get it?

Keep in mind that bandwidth "ownership" or licensing of the same frequencies varies from country to country, and also each country varies the acceptable uses of frequencies according to its own laws and regulations.  There is no common usage or ownership- though many countries have similar and/or identical usage plans and treaties in place. 

So that means, bear with me, that both satellite systems will have to use unlicensed spectrum (which also varies from country to country) or they will need to somehow convince every country they fly over to not only allow them to license common spectrum (unlikely to the point of laughability), but also change their usage regulations to allow the same chunk of spectrum to be used for their system in every country.  An FDD system needs more spectrum, so the problem gets a bit larger.

Now, there are some alternatives.  First, either system could use the "white spaces" in licensed spectrum as Google has been proposing for a decade, but the owners of that unused spectrum simply will never allow it in this country or any other without due compensation.  Second, and more likely IMHO, is that both systems are shopping around for spectrum that's already currently being used by a satellite vendor. 

Why? You ask?  Because the heavy lifting of setting aside specific chunks of licensed spectrum for use across borders has already been done by these entities.  It's in use by different applications right now, but that can always change, and the different applications may even be able to coincide if the white spaces are used.

In other words.  Don't build when you can buy.  If I were a betting man, I'd purchase stock in Irridium or Sirius/XM (just to name two).  Their phones have undoubtedly been ringing with requests to license the spectrum outright, or utilize their white space- IMHO.

Whatever happened to Code-domain?   Especially if you're going to maintain links with multiple satellites?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 01/26/2015 05:28 pm
Whatever happened to Code-domain?   Especially if you're going to maintain links with multiple satellites?

If you are referring to "code-division" (as in CDMA), the coding allows multiple sessions within the same band separated by codes.  That is implemented separately from whether or not the duplexing scheme is TDD or FDD within a CDMA-type channel access system, and doesn't affect multiple links between satellites (I think.  No one has ever built such a system in orbit TMK). 

For my own money, and speaking from experience with terrestrial systems, if I were the product designer on this system I'd implement a similar channel access scheme called "OFDMA" where sub-bands can be allocated per user to enable differing levels of radio bandwidth to be purchased by a service subscriber. Because bandwidth is scarce, it is at a premium for any satellite system, and bandwidth allocations should therefore form the basis of any level-of-service differentiators offered by SpaceX or OneWeb.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 01/28/2015 07:15 am
These new semiconductors maybe one part of solution to reducing build cost of these satellites.

http://nextbigfuture.com/2015/01/gallium-nitride-is-ten-times-better.html
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 01/28/2015 09:34 am
These new semiconductors maybe one part of solution to reducing build cost of these satellites.

http://nextbigfuture.com/2015/01/gallium-nitride-is-ten-times-better.html

Or the recent development of very small radio wave circulators

PhysOrg.com.... (http://m.phys.org/news/2014-11-lighter-cheaper-radio-device-telecommunications.html)

Quote
Researchers at the Cockrell School of Engineering have built a radio wave circulator that has the potential to double the useful bandwidth in wireless communications by enabling full-duplex functionality, meaning devices can transmit and receive signals on the same frequency band at the same time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 01/28/2015 05:04 pm
Can anyone toss out/rule out ISM as a possibility?  If they can get the range up from Project LOON (google; using unlicensed 2.4 and 5.8 GHz ISM bands) there are several world wide spectrum set asides already (http://en.wikipedia.org/wiki/ISM_band).

Frequency range    Bandwidth    Center frequency    Availability
6.765 MHz    6.795 MHz    30 kHz    6.780 MHz    Subject to local acceptance
13.553 MHz    13.567 MHz    14 kHz    13.560 MHz    Worldwide
26.957 MHz    27.283 MHz    326 kHz    27.120 MHz    Worldwide
40.660 MHz    40.700 MHz    40 kHz    40.680 MHz    Worldwide
433.050 MHz    434.790 MHz    1.74 MHz    433.920 MHz    Region 1 only and subject to local acceptance
                                                                                          (within the amateur radio 70 cm band)
902.000 MHz    928.000 MHz    26 MHz    915.000 MHz    Region 2 only (with some exceptions)
2.400 GHz    2.500 GHz    100 MHz    2.450 GHz    Worldwide
5.725 GHz    5.875 GHz    150 MHz    5.800 GHz    Worldwide
24.000 GHz    24.250 GHz    250 MHz    24.125 GHz    Worldwide
61.000 GHz    61.500 GHz    500 MHz    61.250 GHz    Subject to local acceptance
122.000 GHz    123.000 GHz    1 GHz    122.500 GHz    Subject to local acceptance
244.000 GHz    246.000 GHz    2 GHz    245.000 GHz    Subject to local acceptance

Like this little bob here: A tiny device that can transmit up to 5km, can it be scaled up?
http://www.matrix.es/eng/M2M-Communications-Industrial-Networking/RF-ISM-BANDS/Components-Embedded-Modules/Long-Range-RF-868MHZ-Modules

This on says wireless range of fifty miles:  could something like this be scaled up to reach leo with tolerable latency?
http://www.afar.net/wireless/ethernet-bridge/?gclid=CjwKEAiArqKmBRCOj_qfmuqinnYSJAAkAYwGQ48IfyY_DM_1c9BrwefQMPlej0HTsUVE48OHThrZpRoCnf_w_wcB
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 01/28/2015 07:35 pm
There's a lot more to communicating with a system like this than scaling up some point to point wireless link. With geo sats, it's not so hard because they don't move. With  this system you'll have to manage hundreds of satellites zipping every which way at thousands  of mph. That  means any sort of time shared uplink, like most ku band VSATs use, will be almost impossible because of the rapidly changing relative distances between the sats and remotes. Spot beams, restricting that transponder coverage to a particular area would help in that case, but it's not the only solution. They might also do like Iridium, before it got oversold, and keep multiple paths up so a drop to one sat won't interrupt things. Harder with high gain antennas, but not too bad if they come up with a dual or tri beam phased.
 I'm thinking it's going to be an entirely new technology. When you want high speed data in a system like this, a lost packet is probably going to be treated differently than with existing schemes.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/02/2015 02:55 pm

If a receiver would be on a (fast) moving platform, would this even be supported?
How fast would it need to move, before its roaming becomes too fast?
Just curious. :)

Everything on the ground and flying through the air is essentially stationary with respect to this constellation.  With 24 sats in the plane that you are on, one would be passing overhead about every 4 minutes.  If driving, you would have covered roughly 4 miles between sats -- flying, about 10 times that much.  The sats would have covered over 1000 miles relative to you on the Earth's surface.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/02/2015 02:59 pm
New article concerning competition with ESA/EU... now shifting to satellites.

Quote
Increased competition will challenge ESA’s space authority

Quote
Last month, news surfaced that Elon Musk, working with and funded by Google, want to create a constellation of about 4,000 small satellites to provide global Internet access. A new SpaceX factory dedicated to manufacturing satellites will be created as part of that effort. The Virgin Group, together with OneWeb, has similar ambitions. It is very likely that this development will challenge the next major sector in space: satellite manufacturing. It is easy to imagine that Europe’s satellite manufacturing industry will experience a shock similar to what the launcher industry is experiencing now. It is also possible that ESA will, again, try to increase efficiency by disregarding the GeoReturn principle.

http://www.thespacereview.com/article/2687/1

Also:
Quote
ESA was successful in the old world where space funding was provided by governments, resulting in a stable foundation for European (and global) space activities. The money for the space industry was secure and did not encourage risk-taking in the development of new space technologies. As a result, the space landscape has not changed much in the last 30 years. Technological details may have improved, but nothing much changed in the big picture. With a few exceptions, such as electric propulsion, innovation was limited to increasing efficiency of existing technologies. Access to space remained expensive, satellites continued to be big boxes incorporating many sensors, and the components of spacecraft and rockets became little more advanced than those used in the Apollo era. Additionally, we are not palpably closer to go beyond Earth orbit than 30 years ago.

This environment meant that launcher industries around the world were utterly surprised by the rise of SpaceX. They were surprised because the need to evolve launcher technology by a giant leap was not apparent to them. SpaceX shows that technology has advanced sufficiently in the last 30 years to enable new, game changing approaches to space access. Why did none of the other launch providers even start to tap into this technological potential?

The answer is that they were too comfortable in their positions. They were not in an existential crisis because their launches were guaranteed by government contracts that do not encourage risk taking.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 02/02/2015 03:54 pm
Can anyone toss out/rule out ISM as a possibility?  If they can get the range up from Project LOON (google; using unlicensed 2.4 and 5.8 GHz ISM bands) there are several world wide spectrum set asides already (http://en.wikipedia.org/wiki/ISM_band).

Read the link you posted for the answer. :)

Those bands are specifically set aside to provide space for extremely powerful radio disruptions caused by non-communicating machines (like microwave ovens) that can cause noise in those bands. There is a need for keeping communications and interference-causing machines separate- like keeping trains on railroad beds and trucks on roadways.  It's not a good idea to mix the two.

We call them "junk" bands, and for large-scale or long range deployments they really are.  I wouldn't build a business model around ISM-using equipment.

In fact, I can give you a real world example.  Back in 2003, I was a product developer testing a 2.4GHz unlicensed omni radio (because the 2.6GHz licensed radios weren't ready yet) for two-way mobile data communications.  The radios worked just fine even though being unlicensed we were operating at 1 watt (versus 5 watts with the licensed versions!) except that every few hours the system would be completely swamped by radio interference and become unusable.  After some sleuthing about with radio sniffers, we discovered that there was a building in a nearby commercial park that housed a commercial uniform laundering company, and they were using large microwave dyers to speed up the laundering process for the uniforms they cleaned.  If I had deployed the unlicensed system (not that I ever would have) I- and my customers- would have had to accept a totally unusable network for several hours at random times each day- and that was just from that one interference source.



Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 02/02/2015 05:16 pm
I wouldn't be surprised if this satellite company ends up on SpaceX or Google shopping list. Google especially have shown a willingness to purchase companies with technology they need

http://www.parabolicarc.com/2015/02/02/york-space-systems/
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 02/02/2015 05:19 pm
New article concerning competition with ESA/EU... now shifting to satellites.

Quote
Increased competition will challenge ESA’s space authority

Quote
Last month, news surfaced that Elon Musk, working with and funded by Google, want to create a constellation of about 4,000 small satellites to provide global Internet access. A new SpaceX factory dedicated to manufacturing satellites will be created as part of that effort. The Virgin Group, together with OneWeb, has similar ambitions. It is very likely that this development will challenge the next major sector in space: satellite manufacturing. It is easy to imagine that Europe’s satellite manufacturing industry will experience a shock similar to what the launcher industry is experiencing now. It is also possible that ESA will, again, try to increase efficiency by disregarding the GeoReturn principle.

http://www.thespacereview.com/article/2687/1

Also:
Quote
ESA was successful in the old world where space funding was provided by governments, resulting in a stable foundation for European (and global) space activities. The money for the space industry was secure and did not encourage risk-taking in the development of new space technologies. As a result, the space landscape has not changed much in the last 30 years. Technological details may have improved, but nothing much changed in the big picture. With a few exceptions, such as electric propulsion, innovation was limited to increasing efficiency of existing technologies. Access to space remained expensive, satellites continued to be big boxes incorporating many sensors, and the components of spacecraft and rockets became little more advanced than those used in the Apollo era. Additionally, we are not palpably closer to go beyond Earth orbit than 30 years ago.

This environment meant that launcher industries around the world were utterly surprised by the rise of SpaceX. They were surprised because the need to evolve launcher technology by a giant leap was not apparent to them. SpaceX shows that technology has advanced sufficiently in the last 30 years to enable new, game changing approaches to space access. Why did none of the other launch providers even start to tap into this technological potential?

The answer is that they were too comfortable in their positions. They were not in an existential crisis because their launches were guaranteed by government contracts that do not encourage risk taking.

The key in all of this is not to get in a war with the operators.  They are the business, have the customers and access to local markets, etc. These guys are regional by nature, and so it should be relatively easy to get them on board.

That's why I worry a little bit about the way the satellite project was presented. I hope there are behind the scenes talks in that direction.

About a year ago, SES's CTO commented something to the effect of "this company is going to completely revolutionize the business".  I wonder how much he knew back then.  (SES is involved in oneWeb, right? This is back when oneWeb was still associated with Google)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/03/2015 02:36 am
Elon's SkyNet will do to the fiber what the cellular network did to the landline. Period.
If I were a betting man, I'd probably bet you were right.

But I wouldn't be so confident. Long-haul fiber is expensive and not low-latency.

And how many people 30 years ago would've said cell would do to the landline what it has done? It's always amazing the confidence people have that things won't change, and then they do. I could list off half a dozen reasons right now why this constellation could make a huge difference even for long-haul fiber.


On a separate note, people mentioned switching capacities of 100s of Gigabits per satellite. Try Terabits or 10s of Terabits (or later on, Petabits/s) per satellite.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 02/03/2015 02:59 am
Just a little perspective.

http://www.extremetech.com/computing/96827-the-secret-world-of-submarine-cables (http://www.extremetech.com/computing/96827-the-secret-world-of-submarine-cables)

Quote
Across these cables, which span distances of up to 13,00o km (8,000 miles) and have total lengths over 21,000 km (13,000 miles), terabits of information squirt from one side of the planet to another. To get from London to Tokyo, your packets can traverse Europe, the Mediterranean, the Red Sea, the Arabian Sea, the Indian Ocean, and finally the South China Sea — or they can hop across the Atlantic, the entirety of continental North America, and then long haul over the Pacific.

Quote
Finally, we would be remiss to ignore geostationary, orbital communications satellites. While satellite data links can be in the gigabit range, the high latency of bouncing a signal through a point that is rather far away (35,000 km, 22,000 miles) makes them unsuitable for many consumer internet services. This same latency will pose some big problems when we start colonizing other planets and need to create an interplanetary (or intergalactic) internet, or galnet for short.

Dated Sept 2011.

http://www.submarinecablemap.com/ (http://www.submarinecablemap.com/)

http://www.cablemap.info/


http://www.extremetech.com/computing/187872-google-invests-in-60-terabit-300-million-trans-pacific-cable-to-protect-its-growth-in-asia (http://www.extremetech.com/computing/187872-google-invests-in-60-terabit-300-million-trans-pacific-cable-to-protect-its-growth-in-asia)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/03/2015 03:02 am
Can anyone toss out/rule out ISM as a possibility?  If they can get the range up from Project LOON (google; using unlicensed 2.4 and 5.8 GHz ISM bands) there are several world wide spectrum set asides already (http://en.wikipedia.org/wiki/ISM_band).

Read the link you posted for the answer. :)

Those bands are specifically set aside to provide space for extremely powerful radio disruptions caused by non-communicating machines (like microwave ovens) that can cause noise in those bands. There is a need for keeping communications and interference-causing machines separate- like keeping trains on railroad beds and trucks on roadways.  It's not a good idea to mix the two.

We call them "junk" bands, and for large-scale or long range deployments they really are.  I wouldn't build a business model around ISM-using equipment.

In fact, I can give you a real world example.  Back in 2003, I was a product developer testing a 2.4GHz unlicensed omni radio (because the 2.6GHz licensed radios weren't ready yet) for two-way mobile data communications.  The radios worked just fine even though being unlicensed we were operating at 1 watt (versus 5 watts with the licensed versions!) except that every few hours the system would be completely swamped by radio interference and become unusable.  After some sleuthing about with radio sniffers, we discovered that there was a building in a nearby commercial park that housed a commercial uniform laundering company, and they were using large microwave dyers to speed up the laundering process for the uniforms they cleaned.  If I had deployed the unlicensed system (not that I ever would have) I- and my customers- would have had to accept a totally unusable network for several hours at random times each day- and that was just from that one interference source.

I asked mainly because its already being used for internet/wi-fi...by (Google's Project Loon uses ISM bands (specifically 2.4 and 5.8 GHz bands) for balloon-to-balloon and balloon-to-ground communications.  At altitudes of 32km at 3g speeds)).  In 1st world countries lots of interference is to be expected.  There are Billions in areas with no such interference. 

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: DanielW on 02/03/2015 03:07 am
Elon's SkyNet will do to the fiber what the cellular network did to the landline. Period.
If I were a betting man, I'd probably bet you were right.

But I wouldn't be so confident. Long-haul fiber is expensive and not low-latency.

And how many people 30 years ago would've said cell would do to the landline what it has done? It's always amazing the confidence people have that things won't change, and then they do. I could list off half a dozen reasons right now why this constellation could make a huge difference even for long-haul fiber.


On a separate note, people mentioned switching capacities of 100s of Gigabits per satellite. Try Terabits or 10s of Terabits per satellite.

While I don't necessarily disagree. It is important to have a bit of perspective. The cellular was not alone in killing paired copper, much of the replacement infrastructure is also cable and fiber. 4G internet is great and all but I am not about to give up my land connection until something with a more intelligent pricing structure arrives. This constellation may be that something but I suspect that fiber will find a niche to fill for a while yet. It will still be lower latency and higher band-width for anything under 1000 miles or so.
 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/03/2015 03:13 am
Indeed. Musk himself doesn't imagine the constellation would replace local fiber, only edging out long-haul and supplementing local fiber.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 02/03/2015 03:31 am
http://www.csgnetwork.com/us30_100ghz.html (http://www.csgnetwork.com/us30_100ghz.html)



http://www.ntia.doc.gov/files/ntia/publications/spectrum_wall_chart_aug2011.pdf (http://www.ntia.doc.gov/files/ntia/publications/spectrum_wall_chart_aug2011.pdf)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/03/2015 07:46 am
But I wouldn't be so confident. Long-haul fiber is expensive and not low-latency.
Neither of these things seem particularly accurate. It's expensive but the cost is spread around enormously, and you can go around the world in a few hundred milliseconds. Noticeable in some contexts, but hardly prohibitive.

And how many people 30 years ago would've said cell would do to the landline what it has done?
This seems like it misses the mark, you're one step behind the obsolescence you're mocking. Cellphones can replace landline telephone service, but at this point both are only incidentally for voice service. For IP, cell doesn't have enough suitable spectrum exist to replace copper, let alone fiber.

On a separate note, people mentioned switching capacities of 100s of Gigabits per satellite. Try Terabits or 10s of Terabits (or later on, Petabits/s) per satellite.
Not sure how you got to these numbers, just based on what switching is possible in wired networks on likely power budgets this seems several magnitudes high even with 10+ years of improvement.

Fiber's getting to petabit speeds by putting multiple cores in a single strand, but you don't expect to handle that amount of traffic in single compact device.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/03/2015 12:50 pm
Internet traffic is currently more like 200Tbps (lower than I thought, so it's likely my earlier estimates for per-satellite switching capacity were too high), only about half of that long haul. ViaSat-1 is 140Gbps, ViaSat-2 is 2.5x that or so, around 350Gbps. If you've got 4000 ViaSats, you could easily carry the majority of long haul, if you take those figures at face value.

BTW, what is high bandwidth but can't be handled by content delivery networks? Only thing I can think of is something like Skype video or IP calling and the like. That IS very latency sensitive and would definitely benefit from the significantly lower latency a LEO constellation could give you.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: dror on 02/03/2015 07:29 pm
How is Inmarsat global express diffrent from the LEO constelation?
Apperrantly it cost less to establish ($1.6B) , it delivers the same (broadband communication) and starts this year.
Do we know the speed it will deliver, the capacity of each sat or the prices for the service?


<iframe width="450" height="337" src="https://www.youtube.com/embed/8nwiiRnCec0" frameborder="0" allowfullscreen></iframe>
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 02/03/2015 07:49 pm
How is Inmarsat global express diffrent from the LEO constelation?
Apperrantly it cost less to establish ($1.6B) , it delivers the same (broadband communication) and starts this year.
Do we know the speed it will deliver, the capacity of each sat or the prices for the service?


Inmarsat has such insane pricing, I'm not sure how they've retained customers, and it's been getting worse. On demand ISDN is close to $10 a megabyte. Not gigabyte. Megabyte. You can do better with MPDS, but it's completely useless for large parts of the day. They almost went under from the Lightsquared partnership and are still hurting.
 They claim Global Express will be better, but they claimed that with other services, and failed to deliver much better deals.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/03/2015 07:50 pm
Latency. Without the latency issue, you could do this in GSO, sure. But latency is a big deal.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/03/2015 08:20 pm
Internet traffic is currently more like 200Tbps (lower than I thought, so it's likely my earlier estimates for per-satellite switching capacity were too high)
IMO SpaceX's internet venture can do without the performance boasts where no one has even attempted to arrive at a reality-based number.

Only thing I can think of is something like Skype video or IP calling and the like. That IS very latency sensitive and would definitely benefit from the significantly lower latency a LEO constellation could give you.
There'd be a noticeable improvement for applications like that, though it's not clear to me how the plumbing works for this. Fiber's going to remain cheaper for long haul traffic so it seems unlikely most telecoms are going to pay to switch transit over to the SpaceX network.

What would have to happen for eg Skype would be building up local points of presence in different markets so people on local copper/fiber ISPs have a place the terminate connections that will use ElonSats.

I think most of the "long haul" stuff SpaceX anticipates is actually users in areas where the options aren't great that will switch to SpaceX. Then even CDN traffic will be long haul. This would add up pretty fast even though it's largely not latency sensitive. It would also have to anticipate that this service will bring many new internet users online, though that's probably a reasonable assumption.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/03/2015 08:27 pm
Companies already pay a pretty penny for interconnect fees. Having an inexpensive low-latency-long-haul alternative to the local entrenched monopolies would make a lot of companies very happy (ask me how I know).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 02/03/2015 08:50 pm
Companies already pay a pretty penny for interconnect fees. Having an inexpensive low-latency-long-haul alternative to the local entrenched monopolies would make a lot of companies very happy (ask me how I know).

How do you know?  :o Granted, the question is rather self-evident, but I'm curious to see if there's any direct specifics you can shed as to exactly which companies would benefit directly, perhaps to the point of directly investing. They're still going to need another 9 billion.

What would be extremely helpful if somebody (hate asking to people to do stuff I'm fairly unlikely to do myself) would make a list of possible entity types (as opposed to individual entities, that'd go on forever), I.E: various industries/consumer groups that would benefit/lose out from this constellation? Who is this going to endear and who is it going to annoy? There's been lots of (useful) speculation but no general consensus yet.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ScepticMatt on 02/03/2015 09:49 pm
Internet traffic is currently more like 200Tbps
2014: 137 Tbps total
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
Source: https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/03/2015 09:55 pm
Internet traffic is currently more like 200Tbps
2014: 137 Tbps total
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
Source: https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf
I think this measures the wrong thing for talking about ElonSats.

Aggregate Internet bandwidth suffers from the coastline paradox. When you count up all the bandwidth, what do you measure? People's wifi at home? Probably not, but where do you draw the line? What about CDN content that's hosted inside an ISP's network?

Your link measures international, so it's way low for our purposes.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Elmar Moelzer on 02/04/2015 12:24 am
BTW, what is high bandwidth but can't be handled by content delivery networks? Only thing I can think of is something like Skype video or IP calling and the like. That IS very latency sensitive and would definitely benefit from the significantly lower latency a LEO constellation could give you.
Online gaming, cloud based gaming and cloud based applications. The more things move to the cloud, the more important latency and data rate become. I think that the bad internet service in the US has been a big preventer of cloud based services taking off. I will be one of the first to drop my current cable internet service as soon as something (anything) competitive comes up that is not a complete rip off.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/04/2015 01:23 am
BTW, what is high bandwidth but can't be handled by content delivery networks? Only thing I can think of is something like Skype video or IP calling and the like. That IS very latency sensitive and would definitely benefit from the significantly lower latency a LEO constellation could give you.
Online gaming, cloud based gaming and cloud based applications. The more things move to the cloud, the more important latency and data rate become. I think that the bad internet service in the US has been a big preventer of cloud based services taking off. I will be one of the first to drop my current cable internet service as soon as something (anything) competitive comes up that is not a complete rip off.

If someone paid for this type of service, would they have access to it no matter where they were? (except of course on commercial aircraft if present rules don't change)  This would be worth a premium price -- and the wealthiest would be the first to subscribe.  If so, then this service wouldn't only be for the ill-served in remote locations...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 01:31 am
Internet traffic is currently more like 200Tbps
2014: 137 Tbps total
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
Source: https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf
Thanks! I was basing it off of estimates by Cisco (did you guys really think I just made up the 200Tbps figure? ;) ), but I think your figures are more relevant.

It's worth noting that's peak total International capacity, which is different from both total capacity and average throughput. But in either case, it's a good figure to use, here.

ViaSat-1 has a capacity of 140Gbps, by the way, which is almost exactly 1/1000th of that figure (ViaSat-2 will launch soon and has 2.5x that capacity, last I heard). SpaceX wants their satellites to be just as capable as large satellites, but instead wants 4000 of them. If you had 4000 ViaSat-2s, you could easily handle the entire international Internet traffic (granted, in this case of a LEO constellation you'd likely need a few hops, but that's true for international Internet capacity as well). In either case, it's clear that if they are able to build such a constellation as they've laid out, it would indeed be capable of taking on a large chunk of back-haul Internet traffic (and depending on the exact distribution of users and content and specs of the satellite, even a majority).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 01:35 am
Internet traffic is currently more like 200Tbps
2014: 137 Tbps total
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
Source: https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf
I think this measures the wrong thing for talking about ElonSats.

Aggregate Internet bandwidth suffers from the coastline paradox. When you count up all the bandwidth, what do you measure? People's wifi at home? Probably not, but where do you draw the line? What about CDN content that's hosted inside an ISP's network?

Your link measures international, so it's way low for our purposes.
No, it's not because Elon specifically said the constellation isn't intended to replace the majority of local traffic (i.e. traffic to CDN), only long-haul, which is just what this figure is talking about (although it doesn't count East-to-West coast).

And this figure looks similar to other estimates of actual Internet traffic, especially of long-haul traffic, see here: http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/VNI_Hyperconnectivity_WP.html

And it's not until recently that CDN traffic has exceeded long-haul traffic.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 01:38 am
Internet traffic is currently more like 200Tbps
2014: 137 Tbps total
Quote
The annual rate of capacity growth has fallen from 54 percent in 2010 to 32 percent in 2014 (see Figure: International Internet Bandwidth Growth, 2010–2014). Though its significance is masked by this declining pace of growth, 100 Tbps of new international Internet capacity was deployed between 2010 and 2014, bringing total global international Internet capacity to 137 Tbps
Since TeleGeography began tracking international Internet capacity in 1999, the highest capacity inter-regional routes had always been Europe-U.S. & Canada and Asia-U.S. & Canada. This changed in 2014 as capacity on the Latin America-U.S. & Canada route rose 43 percent to reach 12.6 Tbps, topping the 10.5 Tbps of capacity on the Europe-U.S. & Canada route
Source: https://www.telegeography.com/page_attachments/products/website/research-services/global-internet-geography/0005/1382/GIG_Executive_Summary.pdf
I think this measures the wrong thing for talking about ElonSats.

Aggregate Internet bandwidth suffers from the coastline paradox. When you count up all the bandwidth, what do you measure? People's wifi at home? Probably not, but where do you draw the line? What about CDN content that's hosted inside an ISP's network?

Your link measures international, so it's way low for our purposes.
Average TOTAL Internet traffic (including CDN) is projected to be about 310Tbps by 2018 according to Cisco. So however you draw the line, that figure is around the right order of magnitude (which is all we can hope for anyway, given the lack of specific technical details about the constellation): http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/VNI_Hyperconnectivity_WP.html
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bstrong on 02/04/2015 02:01 am
Elon sounded awfully confident about his marketshare projections. I think there's a good chance that comes from Google saying, "sign us up for 100 Tbps" (or something along those lines), which makes the whole venture decidedly less risky than everyone is assuming.

To me, the interesting question is whether Google has a specific use-case in mind that requires this network, or if this is just part of their general plan to make sure they have the long-haul capacity they'll need 5-10 years from now.

One possible use-case I've been kicking around: Google cares a lot about serving requests with low latency and builds datacenters around the world to ensure that they can. For any given latency target, you could route requests to servers that are geographically farther away via satellite than via fiber. This would allow them to make more efficient use of their current datacenter capacity by balancing traffic across datacenters in larger geographic regions than they do today, and it would allow them to build fewer (larger) datacenters in the future, because they wouldn't need to be quite as geo-distribtued.

Google spends double-digit billions per year building and operating datacenters, so utilizing them more efficiently could plausibly justify the whole constellation. I'm not saying this is the actual justification, only that it's interesting to spend some time thinking about how Google in particular would benefit from it, particularly in ways that wouldn't make sense for any other company.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 02:09 am
And heck, I wonder if it'd be worth it putting a slice of their most popular content (or perhaps some ultra-latency-sensitive application) in a cache on each satellite?

(My bet would be no, but this could change.)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bstrong on 02/04/2015 02:24 am
And heck, I wonder if it'd be worth it putting a slice of their most popular content (or perhaps some ultra-latency-sensitive application) in a cache on each satellite?

(My bet would be no, but this could change.)

I could see this making sense for serving users directly connecting to the constellation, but for backbone traffic I think it will always make more sense from both a cost and latency point of view to put the caches at the ground stations.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 02:27 am
Yeah, that's what I was thinking (for users directly connecting). Like a cache of Netflix or something, since that's fairly limited, changes fairly slowly/predictably, and makes up a large portion of total traffic. (It isn't at all latency-sensitive, though, so probably doesn't make sense vs the ground.)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bstrong on 02/04/2015 02:38 am
Thinking about it more, I'm not sure it makes a lot of sense for the end-user case, either. Web traffic patterns are usually highly correlated to geography, so you're going to get much higher cache hit rates on a geographically fixed cache than one moving at 7km/s. So, even if latency is a little higher, you're probably still better off with ground-based caches.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 02:41 am
Thinking about it more, I'm not sure it makes a lot of sense for the end-user case, either. Web traffic patterns are usually highly correlated to geography, so you're going to get much higher cache hit rates on a geographically fixed cache than one moving at 7km/s. So, even if latency is a little higher, you're probably still better off with ground-based caches.
Yeah, it'd definitely have to be a special application, not a general cache.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 02/04/2015 02:48 am
Thinking about it more, I'm not sure it makes a lot of sense for the end-user case, either. Web traffic patterns are usually highly correlated to geography, so you're going to get much higher cache hit rates on a geographically fixed cache than one moving at 7km/s. So, even if latency is a little higher, you're probably still better off with ground-based caches.
Yeah, it'd definitely have to be a special application, not a general cache.
Go crazy...  Put the Netflix archive in GEO...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bstrong on 02/04/2015 02:58 am
A somewhat related question I've been pondering is whether IP multicast starts to make more sense with this constellation than it does on the terrestrial internet (where it effectively doesn't work at all).

No concrete ideas there, but it seems to me there could be some interesting applications.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/04/2015 03:05 am
No, it's not because Elon specifically said the constellation isn't intended to replace the majority of local traffic (i.e. traffic to CDN), only long-haul
That CDN traffic might not be long haul now for most people, but it's going to have to be on ElonSats. CDN traffic has to terminate somewhere, and unless Akamai starts eg building a datacenter on Midway Island what else is it going to do?

This network is going to bring the next billion people online, and a lot of them don't have internet access right now because the local infrastructure. Meaning, yes, CDN traffic is going to go a long way.

When he says local traffic I think he means it's probably not going to be able to displace incumbents in densely populated areas. The share of long distance traffic is largely going to come from traffic that doesn't exist right now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 03:09 am
Yeah, I think that can be surmised.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: bstrong on 02/04/2015 03:22 am
That CDN traffic might not be long haul now for most people, but it's going to have to be on ElonSats. CDN traffic has to terminate somewhere, and unless Akamai starts eg building a datacenter on Midway Island what else is it going to do?

Depending on how cheap the bandwidth winds up being, it's possible that CDNs will become less distributed in general once the constellation is live, since you'll be able to provide the same latencies from a smaller number of larger POPs (meaning most CDN traffic will become long-distance traffic, reversing current trends).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/04/2015 03:29 am
Elon sounded awfully confident about his marketshare projections. I think there's a good chance that comes from Google saying, "sign us up for 100 Tbps" (or something along those lines), which makes the whole venture decidedly less risky than everyone is assuming.
I think it's extremely unlikely Google would use this for internal traffic. The variability in performance would make it pathological to work with. And, it's not going to be cheaper than fiber, especially not to Google.

To me, the interesting question is whether Google has a specific use-case in mind that requires this network
Bringing the next billion people online.

One possible use-case I've been kicking around: Google cares a lot about serving requests with low latency and builds datacenters around the world to ensure that they can. For any given latency target, you could route requests to servers that are geographically farther away via satellite than via fiber. This would allow them to make more efficient use of their current datacenter capacity by balancing traffic across datacenters in larger geographic regions than they do today, and it would allow them to build fewer (larger) datacenters in the future, because they wouldn't need to be quite as geo-distribtued.
You'd never want to configure your capacity such that issues with a satellite can have impact that will take a long time to recover from.

And heck, I wonder if it'd be worth it putting a slice of their most popular content (or perhaps some ultra-latency-sensitive application) in a cache on each satellite?
I think the working set required to make a dent in anything important makes this a non-starter given how much extra hardware this implies for each satellite. If the latency helps skype and gaming it should be fine for virtually everything else.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/04/2015 03:42 am
Depending on how cheap the bandwidth winds up being, it's possible that CDNs will become less distributed in general once the constellation is live
Given hints (probably not competitive with copper in densely populated areas) I think that's unlikely.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ggr on 02/04/2015 04:08 am
How is Inmarsat global express diffrent from the LEO constelation?
Apperrantly it cost less to establish ($1.6B) , it delivers the same (broadband communication) and starts this year.
Do we know the speed it will deliver, the capacity of each sat or the prices for the service?


<iframe width="450" height="337" src="https://www.youtube.com/embed/8nwiiRnCec0" frameborder="0" allowfullscreen></iframe>

It's very easy to quote performance for a single user. But what happens when you have multiple (2, 10, 100x) users? Does your system still perform? Inmarsat satisfies (at relatively large cost) a very small market.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 02/04/2015 04:33 am
Companies already pay a pretty penny for interconnect fees. Having an inexpensive low-latency-long-haul alternative to the local entrenched monopolies would make a lot of companies very happy (ask me how I know).

How do you know?  :o Granted, the question is rather self-evident, but I'm curious to see if there's any direct specifics you can shed as to exactly which companies would benefit directly, perhaps to the point of directly investing. They're still going to need another 9 billion.

What would be extremely helpful if somebody (hate asking to people to do stuff I'm fairly unlikely to do myself) would make a list of possible entity types (as opposed to individual entities, that'd go on forever), I.E: various industries/consumer groups that would benefit/lose out from this constellation? Who is this going to endear and who is it going to annoy? There's been lots of (useful) speculation but no general consensus yet.
Well here's one anecdote: The company I worked at most recently had plans that necessitated a large remote construction camp in northern Alberta away from existing infrastructure.  To attract workers away from the Ft. Mac area and be a competitive employer, there were a few requirements.  One was bringing fiber into the area at a cost of roughly $6 million.  A small operator in the area had been using a series of radio towers, and ongoing operations could definitely do it that way, but workers all want to be able to Skype their families, stream Ayreon or 8K FUHD cat videos on youtube, watch space launches, or whatever.  http://en.wikipedia.org/wiki/8K_resolution#Resolutions



If available, Elon$at would be a no-brainer in a case like that. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Mongo62 on 02/04/2015 09:00 pm
There were additional job postings on the SpaceX News Twitter account (https://twitter.com/spacexnews/), but the linked pages no longer exist. The job openings must have already been filled.

Satellite Lead Analyst, Vehicle Loads (http://www.spacex.com/careers/position/6616)

RESPONSIBILITIES:
Identify and analyze all load conditions for satellite lifetime; terrestrial, launch, deployables activation, orbit raising, in-orbit, de-orbit.
Analyze load conditions and provide inputs and recommendations to design teams (bus, payload, components) for corrections, alternatives.
Demonstrate that the vehicle structure solution meets all strength and stiffness requirements in all loading conditions, and accommodates bus sub-systems and payload operational requirements.
Drive analyses plans and schedules to meet vehicle development targets.
Responsibility for technical competency of function within the design group.

TASKS
Partner with system and sub-system engineering units to fully comprehend all vehicle loading conditions, and their effects.
Identify all strength and stiffness requirements and demonstrate validity of design approach through analysis, testing, reliability requirements definition, safety factor determination as well as leveraging heritage knowledge and lessons learned.
Facilitate design teams goals to drive to lowest possible mass-fraction for entire space vehicle; payload, bus and all subsystems and components without compromise to reliability.
Iterate analyses for mass trades and document all approaches and results.
Compile analysis summary reports demonstrating adequate strength margins for all possible modes of failure.
Define terrestrial validation test methods to correlate analysis predictions with physical prototypes.
Engage testing facilities, and assist in developing internal test capability when needed, to assess and verify predictions for load response at component, subsystem and system levels.
Provide design support to engineering and manufacturing to ensure that strength and stiffness. requirements are met for any non-conforming conditions to design with appropriate corrective action.

REQUIRED SKILLS AND QUALIFICATIONS
10+ years of work experience in aerospace industry structural analyses roles.
BS in Aerospace, Mechanical Engineering or related fields.
Flight heritage experience, your work has been to space!

PREFERRED SKILLS AND QUALIFICATIONS
Advanced degree in Aerospace, Mechanical Engineering or related fields.
Extensive experience with the analysis techniques and methods for mass optimized spacecraft structures; Load case definition, Quasi-static and fatigue analyses, Dynamic/modal analyses, thermal induced loads analysis.
Experience in correlation of model predictions and terrestrial testing to launch conditions and in-orbit performance.
Extensive experienced and highly proficient in FEA design analysis tools: (strongly preferred; Ansys workbench, Nastran Femap).
Highly aware of the various bus and payload design drivers and associated, manufacturing issues.
Knowledge of and experience with Unigraphics CAD system NX highly desirable.
Excellent communication and presentation skills.

Satellite Lead Analyst, Thermal Loads (http://www.spacex.com/careers/position/6615)

RESPONSIBILITIES:
Identify and analyze all thermal load conditions for satellite vehicle lifetime; launch, orbit raising, in-orbit operation.
Analyze all load conditions and provide inputs and recommendations to design teams (bus, payload, components) for corrections, alternatives
Deliver an overall thermal solution that meets all operational and mass / cost target criteria, with development schedules.
Demonstrate that the vehicle thermal control system caters to all thermal loading conditions, and accommodates all bus sub-systems and payload operational requirements.
Drive analyses plans and schedules to meet vehicle development targets.

TASKS
Partner with system and sub-system engineering units to fully comprehend all vehicle and sub-system thermal loading limits, conditions, and effects.
Identify all thermal mitigation requirements and demonstrate validity of thermal control system through analysis, testing, reliability requirements definition, safety factor determination as well as leveraging heritage knowledge and lessons learned.
Facilitate design teams goals to drive to lowest possible mass-fraction for entire space vehicle; payload, bus and all subsystems and components with cost, risk, schedule, and performance trades.
Compile thermal analysis summary reports demonstrating adequate margins for all possible operational conditions at component, subsystem, and vehicle system levels.
Define and develop as needed, terrestrial validation test methods to correlate analysis predictions with physical prototypes.
Engage testing facilities, and/or assist in developing any additional internal test capability when needed, to assess and verify material performance predictions for thermal load response at component, subsystem and system levels.
Plan and drive thermal vacuum tests
Engage vendor/supply base as needed in development of thermal control hardware/techniques (materials, sensors, radiators, OSRs, MLIs, surface finishes, heat-pipes, heaters).
Provide design support to engineering and manufacturing to ensure that thermal control system requirements are met for any non-conforming conditions to design intent, with appropriate corrective action.

REQUIRED SKILLS AND QUALIFICATIONS
10+ years of work experience in space mission thermal analyses or satellite thermal architect roles.
BS in Aerospace, Mechanical Engineering or related fields.
Proven successful satellite design heritage experience, your work has been to space!

PREFERRED SKILLS AND QUALIFICATIONS
Advanced degree in Aerospace, Mechanical Engineering or related fields.
Extensive experience with the analysis techniques and methods for satellite spacecraft thermal control systems.
Highly experienced in correlation of model predictions and terrestrial testing to launch conditions and in-orbit performance.
Extensively experienced and highly proficient in design analysis tools: (MSC Sinda, Ansys, or equivalent)
Highly aware of and familiar with the various bus and payload design drivers and their operational limits; (Antennas, RF Payloads, Batteries, ADCS, Avionics, Solar Array, Structure, ISLs, Propulsion etc), and coordinating subsystem analyses into a system level solution.
Experience with Unigraphics NX CAD system desirable.
Excellent communication and presentation skills.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 02/04/2015 09:26 pm
It's worth noting that's peak total International capacity, which is different from both total capacity and average throughput. But in either case, it's a good figure to use, here.

If you had 4000 ViaSat-2s, you could easily handle the entire international Internet traffic

- Not really. One single LEO sat will have to service customers in an area of hundreds of miles across, even with a constellation of 4000. That's potentially tens of millions of people, depending on population density.

- ViaSat-2 costs $358m to make (without insurance, launch). 4000*$358m = $1432bn. And no, you could not easily handle the entire internet traffic, see the argument above.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/04/2015 09:59 pm
Not really. One single LEO sat will have to service customers in an area of hundreds of miles across, even with a constellation of 4000. That's potentially tens of millions of people, depending on population density.
Yeah, a major metro area is out of the question. Even if you had a router sitting on the ground I don't think you could do it with a power budget of a few kilowatts.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/04/2015 10:42 pm
It's worth noting that's peak total International capacity, which is different from both total capacity and average throughput. But in either case, it's a good figure to use, here.

If you had 4000 ViaSat-2s, you could easily handle the entire international Internet traffic

- Not really. One single LEO sat will have to service customers in an area of hundreds of miles across, even with a constellation of 4000. That's potentially tens of millions of people, depending on population density.

- ViaSat-2 costs $358m to make (without insurance, launch). 4000*$358m = $1432bn. And no, you could not easily handle the entire internet traffic, see the argument above.
That's false. With such a huge constellation, you will have dozens of satellites in view. Consider an Iridium constellation is almost 1/100th the number but still has at least one satellite visible overhead at any one time.

Go ahead, do the calculation yourself if you don't believe me!

Additionally, when did SpaceX ever say ALL Internet traffic? They didn't. Musk said eventually the majority of long-haul traffic. And neither did I, I said all /international/ Internet traffic, which is defined in the link above. Please acknowledge this, so we don't have to keep repeating this.

Huge pet peeve of mine when "gotchas" consist only of not carefully reading or understanding what the other person said.

(As far as cost, well, we can address that after these straightforward misunderstandings are rectified so we're all on the same page here.)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 02/05/2015 01:13 am
It's worth noting that's peak total International capacity, which is different from both total capacity and average throughput. But in either case, it's a good figure to use, here.

If you had 4000 ViaSat-2s, you could easily handle the entire international Internet traffic

- Not really. One single LEO sat will have to service customers in an area of hundreds of miles across, even with a constellation of 4000. That's potentially tens of millions of people, depending on population density.

- ViaSat-2 costs $358m to make (without insurance, launch). 4000*$358m = $1432bn. And no, you could not easily handle the entire internet traffic, see the argument above.
That's false. With such a huge constellation, you will have dozens of satellites in view. Consider an Iridium constellation is almost 1/100th the number but still has at least one satellite visible overhead at any one time.

Go ahead, do the calculation yourself if you don't believe me!

I'm pretty sure you want to minimize the satellite beam widths and overlaps in order to maximize power/capacity and minimize ground antenna size/cost and interference. Otherwise what's the point of having so many sats?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 02/05/2015 01:52 am
And heck, I wonder if it'd be worth it putting a slice of their most popular content (or perhaps some ultra-latency-sensitive application) in a cache on each satellite?

(My bet would be no, but this could change.)

I could see this making sense for serving users directly connecting to the constellation, but for backbone traffic I think it will always make more sense from both a cost and latency point of view to put the caches at the ground stations.

There has been some discussion as to the utility of putting datacentres in orbit, usually revolving around issues of cooling (don't quote me on the details) or security (harder for terrorists to blow up!). For the latter at least, latency is not the be-all and end-all.

This does point out that another set of customers is likely to be other LEO users. In fact the system could probably interface with users orbiting much further out; especially if the constellation contained some members dedicated to this purpose.

This in turn points out that there is unlikely to be a single design of satellite in this constellation. Although there are advantages in mass production there are also advantages in specialisation. Optimisation is likely to result in a (probably small) number of different designs, though this may well have many components in common. The two markets identified - many low bandwidth users in sparsely populated regions and a much smaller number of very high bandwidth users for long-distance traffic - seems ripe for such specialisation.

Given his track record, one suspects Elon is already thinking a few steps ahead of his currently disclosed plans!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 03:11 am
Quote
The two markets identified - many low bandwidth users in sparsely populated regions and a much smaller number of very high bandwidth users for long-distance traffic - seems ripe for such specialisation.
How's this even supposed to work? It's not geostationary, every satellite has a ground track that covers most of the world.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 02/05/2015 08:04 am
It's worth noting that's peak total International capacity, which is different from both total capacity and average throughput. But in either case, it's a good figure to use, here.

If you had 4000 ViaSat-2s, you could easily handle the entire international Internet traffic

- Not really. One single LEO sat will have to service customers in an area of hundreds of miles across, even with a constellation of 4000. That's potentially tens of millions of people, depending on population density.

- ViaSat-2 costs $358m to make (without insurance, launch). 4000*$358m = $1432bn. And no, you could not easily handle the entire internet traffic, see the argument above.
That's false. With such a huge constellation, you will have dozens of satellites in view. Consider an Iridium constellation is almost 1/100th the number but still has at least one satellite visible overhead at any one time.

Go ahead, do the calculation yourself if you don't believe me!

I'm pretty sure you want to minimize the satellite beam widths and overlaps in order to maximize power/capacity and minimize ground antenna size/cost and interference. Otherwise what's the point of having so many sats?

Part of the point of having so many sats is having several in view at all times. Most sites aren't going to have perfect views of the sky and need to have several sats to choose from as they zip accross the sky. Other remote sites will probably keep redundant links up to keep transition time to near zero when sat switching needs to occur.
 Another point is that several sats near concentrated numbers of customers can concentrate their efforts there instead of covering only the area right below them.
 It's all guesswork at the moment. There's no detail as to how identicle the sats will be frequency wise, how capable the antennas will be, how intelligent the whole system will be, how much the system will constantly evolve or a dozen other questions.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/05/2015 10:05 am
Satellites out over the oceans can serve the densely populated coastal areas for quite a while before switching to serve ships or planes over the ocean or become transocean relay sats. At an altitude of over 1000km they have quite a large area they can chose to serve using tight beams.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 02/05/2015 12:13 pm
Part of the point of having so many sats is having several in view at all times. Most sites aren't going to have perfect views of the sky and need to have several sats to choose from as they zip accross the sky.

If you look at "footprint maps" of LEO constellations (existing, planned or abandoned) they all seem to try to minimize overlap. Of course you need some overlap to have time to switch between sats and there is more overlap the closer you get to the poles (depending on orbits). Again, I suppose the reason is a that a tighter beam give you more power per area on the ground. With more sats at the same LEO altitude you can provide coverage at smaller satellite elevation angles (which qualifies as having a "better view").

Satellites out over the oceans can serve the densely populated coastal areas for quite a while before switching to serve ships or planes over the ocean or become transocean relay sats. At an altitude of over 1000km they have quite a large area they can chose to serve using tight beams.

That would require a steerable antenna and I suppose if you want different beam widths you'd need multiple antennas. Such sats exist of course, but I'm not sure whether its the way to go for a cheap LEO constellation.

Anyway, I'm not an expert on comsats, maybe someone with better knowledge can weight in.


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/05/2015 12:19 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/05/2015 12:30 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.

I knew phased array antennas can steer the beam. But they can generate different beams simulataneously?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Semmel on 02/05/2015 01:12 pm
I knew phased array antennas can steer the beam. But they can generate different beams simulataneously?

If the different beams are at different wavelength, this should be physically possible. But I dont know if the electronic is able to handle it. The reason for my statement is:

A phase array antenna consist of many antennas in a grid. If you excite them all with the same signal simultaneausly, you generate an interference pattern that has a maximum, perpendicular to the beam array plane. If you alter the time the signal is going to the individual antennas (shift the phase of the signal), you alter the direction at which the constructive interference is directed. In principle, you can excite the antennas with many different wavelength at once and by adjusting the phase of the different signals, the constructive interference is pointing in different directions. But you have to be able to generate the different phases to excite the antenna with the signal. I dont know how difficult it is to create the electronics that can support a range of independent phases.

Cheers,
Semmel
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 02/05/2015 01:25 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.

I knew phased array antennas can steer the beam. But they can generate different beams simulataneously?

Yes, but there are limits to how many (up in the thousands), and each new beam utilizes power, so the overall cell size starts to shrink as the transmitter divides up its available power.  That leaves your marginal customers (because of distance or terrain) without service during busy periods.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Oli on 02/05/2015 03:31 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.

Indeed I thougth only the ground terminals use a phased array.

Yes, but there are limits to how many (up in the thousands), and each new beam utilizes power, so the overall cell size starts to shrink as the transmitter divides up its available power.  That leaves your marginal customers (because of distance or terrain) without service during busy periods.

As you seem to be the expert here:

Is it possible to have beams with lower power without service degradation for individual users? I'm not sure whether lower power reduces capacity for the individual user or only capacity for the entire beam (independent of number of users). Intuitively I thought its the former...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 02/05/2015 04:33 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.

Indeed I thougth only the ground terminals use a phased array.

"Handsets" (ground terminals) and satellites can both have phased array as well, but they don't have to.  I used to deploy Navini "Ripwave" modems  which had two phased array antennas (one on each side) and one pop up omni antenna.  The secret sauce back then was that the DSP chips were coming down in price and they could add-up the  multi-path (signal echos) to get a stronger coherent signal, whereas cell phones at the time actually suffered with multi-path because signal echos arriving from different surfaces raised the noise floor and there was no DSP to add them up.  Cheap DSP chips were really the enabling technology, and they are becoming ubiquitous these days.


Yes, but there are limits to how many (up in the thousands), and each new beam utilizes power, so the overall cell size starts to shrink as the transmitter divides up its available power.  That leaves your marginal customers (because of distance or terrain) without service during busy periods.

Is it possible to have beams with lower power without service degradation for individual users? I'm not sure whether lower power reduces capacity for the individual user or only capacity for the entire beam (independent of number of users). Intuitively I thought its the former...

To your question about using lower power.  Absolutely you can lower the power, and you want to.  Signal quality is more important than power.  It's preferable to have the lowest power possible without going below the noise floor so that you can have the widest possible coverage zone.  Think of a radio station full of static.  If you crank up the volume (adding power) you can get a super loud station full of static that's no more coherent than before.

If you have a high-quality clear signal (see my comment about multi-path and DSP chips), you can back way off of the volume, and get a greater range per watt of power that you're expending, or if you have phased array you can devote the power you do have to marginal users who need it.  In a digital world, you also get fewer packets dropped with a clean signal, so your latency goes down.

I'll also point out- and my team made a multi-million dollar mistake once fully learning this lesson. :) -that in a situation where your "tower" (satellite in this case) has more power available than your handset, your cell size is limited by the handset's ability to yell "Hear I am!!!" to initiate communications with the tower (satellite).  Once the tower knows the handset is out there and trying to talk, it can synch up and initiate beamforming to that handset with it's vastly greater power resources.  The tower will keep pouring higher and higher power towards the handset to maintain a connection as long as it still has more power to send, but it will always use the lowest power required so that it can serve the most number of users with other beams (called "lobes").  In other words, the cell tower may have a range of 20 miles, but if the handset can only talk to the tower out at 10 miles, your cell range is only 10 miles, because any further than that, and your handset can let the tower know it's there.  Similarly, if you go into a building and your signal drops below the noise floor, you may be quite close to the tower but it doesn't know you are there to send more power your way and establish a connection.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 04:34 pm
I'm not sure whether lower power reduces capacity for the individual user or only capacity for the entire beam
It's a whole trade space unto itself. On one extreme contention for the channel will hurt performance more, on the other extreme having to back off the coding due to lower power will hurt performance more. Somewhere in the middle performance is maximized. You'd also need to understand density of your customers inside the beams, and make decisions about what density you want to optimize for, since the satellites will end up all over the world, you can't make one for London and another one for Siberia.

And I'm sure you can have a power level every beam can support steady state versus a power level fewer beams can support, like for the cruise ship or passenger aircraft in the open ocean scenario.

Basically it's a rich trade space and you probably need to decide whether you maximize revenue, global performance, etc. I doubt there's one obvious number, someone just needs to do the analysis.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/05/2015 05:33 pm
A key thing you're missing is that both sides will use phased array antennas. You can send multiple beams in completely different directions simultaneously with a phased array. The beams are steered totally electronically.

I knew phased array antennas can steer the beam. But they can generate different beams simulataneously?

Yes, but there are limits to how many (up in the thousands), and each new beam utilizes power, so the overall cell size starts to shrink as the transmitter divides up its available power.  That leaves your marginal customers (because of distance or terrain) without service during busy periods.
There's no fundamental limit to how many beams you can have, provided you have enough phased array elements (with too few elements, the beams kind of meld together) and good enough electronics. And obviously, energy is conserved.

But I'm assuming SpaceX has some ridiculously ambitious plans as far as how many phased array elements they'll have (over a million?) as well as the power per satellite (>10kW?).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 02/05/2015 05:58 pm
The two markets identified - many low bandwidth users in sparsely populated regions and a much smaller number of very high bandwidth users for long-distance traffic - seems ripe for such specialisation.
How's this even supposed to work? It's not geostationary, every satellite has a ground track that covers most of the world.

Just have the two types mixed in with each other so that there is at least one of each type overhead at all times. With 4,000+ satellites at 1100 miles this should be possible.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 06:19 pm
as well as the power per satellite (>10kW?).
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 02/05/2015 06:25 pm
as well as the power per satellite (>10kW?).
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.

I do expect SpaceX to attempt 10-15kW power for each Sat, but I doubt they will be buying Boeing Spectrolab panels. I assume SpaceX would probably attempt to use Solar City/Silveo solar panels and space rate them.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/05/2015 06:37 pm
A 10kW solar array doesn't need much mass. UltraFlex arrays are 100-150W/kg, so a 10-15kW array may have a mass of just 100kg. But there's room for improvement. One could build an array with a specific power of 500-1000W/kg, bringing the array mass down to tens of kilograms. At that point, you might opt for more power instead of less mass.

It will take them a few revisions to get up to full capability.

But remember, the satellites will have batteries and may not need as much power when over the ocean, so they could crank up the power when over highly populated areas and save up when over the Pacific.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 02/05/2015 07:00 pm
as well as the power per satellite (>10kW?).
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.

Another of Musk's ventures, Solar City, just acquired an advanced solar tech company and is now building a 'gigafactory' (1 GW/year worth of cells) in Buffalo NY to mass produce the tech as we type. Its supposed to open in 2016.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/05/2015 07:12 pm
For Dragon they don't even use spacegrade solar cells. The ordinary type is good enough. It may well be good enough for 5 years. 15 years would likely be too long.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 07:16 pm
I do expect SpaceX to attempt 10-15kW power for each Sat, but I doubt they will be buying Boeing Spectolab panels. I assume SpaceX would probably attempt to use Solar City/Silveo solar panels and space rate them.
Good chance because the unit price for each satellite has to be extremely low. And that's basically what they do with Dragon. But those are a good distance behind the best specialist space panels, so where's the extra power in less mass come from?

UltraFlex arrays are 100-150W/kg, so a 10-15kW array may have a mass of just 100kg.
Except:

a) The whole satellite will barely mass more than that.
b) The satellite has other jobs than generating electricity, like communications.
c) Commodity cells won't be as good as expensive specialist cells.

We see geostationary satellites with 10x the mass, more willingness to pay for specialist solar panels, and less need to buffer power in batteries with power budgets like 5 kw. If we have a pretty good idea of what power budgets are possible with bleeding edge solar technology, that should be taken into account when estimating power per satellite.

A more reasonable number is probably less than geostationary satellites now.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 07:21 pm
as well as the power per satellite (>10kW?).
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.

Another of Musk's ventures, Solar City, just acquired an advanced solar tech company and is now building a 'gigafactory' (1 GW/year worth of cells) in Buffalo NY to mass produce the tech as we type. Its supposed to open in 2016.
That doesn't really answer the question, the Silevo technology is incrementally better than existing commodity cells. That's great news for solar power on Earth but while very expensive, existing space solar arrays are far beyond that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/05/2015 07:23 pm
The super high specific power solar cells (ie well beyond UltraFlex) are actually also some of the cheapest: thin film (on a flexible substrate).

Musk specifically said these small sats would have the capability of regular large sats, which means a high power budget. Perhaps not 10kW at first, but definitely more than 1kW.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/05/2015 07:26 pm
I repeat. Those sats are expected to last 5 years, not 15 years. That alone greatly reduces demands put on the solar cells.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/05/2015 07:30 pm
Musk specifically said these small sats would have the capability of regular large sats, which means a high power budget.
Isn't the plan to jump ahead in integrated circuit technology? That has a lot of potential to increase capabilities on less power.

I repeat. Those sats are expected to last 5 years, not 15 years. That alone greatly reduces demands put on the solar cells.
That's true but I don't think that gets you 3x the power in 1/10 the mass.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 02/05/2015 09:08 pm
as well as the power per satellite (>10kW?).
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.

Another of Musk's ventures, Solar City, just acquired an advanced solar tech company and is now building a 'gigafactory' (1 GW/year worth of cells) in Buffalo NY to mass produce the tech as we type. Its supposed to open in 2016.
That doesn't really answer the question, the Silevo technology is incrementally better than existing commodity cells. That's great news for solar power on Earth but while very expensive, existing space solar arrays are far beyond that.

They don't have to last as long as state of the art, nor be as 'perfect' wrt efficiency. Just good enough to last ~5 year's without blowing the budget..
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AJW on 02/05/2015 09:56 pm
If the satellites are released by a dispenser, are there elements that must currently be incorporated into each individual sat, that could be removed and placed on the dispenser instead?  For example, could the mechanisms for extending the solar panels be removed from each sat if this is done by a single shared system on the dispenser?  This could save mass, cost, and simplify each sat.   Might you also reduce volume significantly if the solar panels are flat packed together and attached just prior to release.  Sats could then have a small plaque reading, "We Came In Pieces"
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/05/2015 09:59 pm
I just checked. The Dragon solar panels produce 5kW. So maybe double that is needed.

They won't fly off the shelf panels. Their Dragon panels are built in house. But they will IMO fly off the shelf cells that suit their needs and will design their own panels with them, just like they did for Dragon. By then they will have learned a lot from their Dragon panels.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: DanielW on 02/05/2015 10:29 pm
How flexible can thin film arrays be? Do they have a small enough radius to be deployed like those new-years party favors? i.e. unrolling with force provided by a handful of inflatable hoses? I am guessing they still can't roll that tightly.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: denis on 02/06/2015 12:09 am
a) The whole satellite will barely mass more than that.
b) The satellite has other jobs than generating electricity, like communications.
c) Commodity cells won't be as good as expensive specialist cells.

We see geostationary satellites with 10x the mass, more willingness to pay for specialist solar panels, and less need to buffer power in batteries with power budgets like 5 kw. If we have a pretty good idea of what power budgets are possible with bleeding edge solar technology, that should be taken into account when estimating power per satellite.

A more reasonable number is probably less than geostationary satellites now.
Actually large geo satellite have solar array that generate up to 20kw. AlphaSat recently launched has 31kW (at beginning of life): http://www.space-airbusds.com/en/equipment/solar-array-kz8.html

A 10kW solar array doesn't need much mass. UltraFlex arrays are 100-150W/kg, so a 10-15kW array may have a mass of just 100kg. But there's room for improvement. One could build an array with a specific power of 500-1000W/kg, bringing the array mass down to tens of kilograms. At that point, you might opt for more power instead of less mass.

The problem of going for a thin-film solution, is that you need larger panels (because the efficiency is half of the triple junction GaAs cells commonly used, see http://sroeco.com/solar/images/PVeff-rev100414.png).
It's true that the cells are lighter, which is why you could end up with a lower mass for a given power.

The problem of trying to use 10kw lightweight arrays on a 300kg satellite, is that you end up with huge and very flexible arrays on a small satellites. This is bad from a dynamic point of view and can have a serious impact on the attitude control performances, in particular for any thruster based mode (for orbit correction). It's already not such a trivial issue on big Geo satellites (which weight 3 tons and have smaller and more rigid arrays). It might be an issue, or not. I guess for normal pointing, you could simply have a very slow response system as you are always nadir pointing. However you'll also need to track the Sun with the arrays, so it's not so simple.

But remember, the satellites will have batteries and may not need as much power when over the ocean, so they could crank up the power when over highly populated areas and save up when over the Pacific.
Maybe, but on the other hand, you need to use solar power to recharge the battery so your satellite can still be useful in eclipse. It lasts for about 1/3 of the orbit, so at first order (assuming 10kW total), in daylight you could use 3kW to recharge the battery and 6kW for the payload. In eclipse, you would use 6kW from the battery for the payload. Eclipse duration is about 30mins, so you need to use 3kWh per orbit. On Saft website, they have Li-ion batteries at 135Wh/kg. So you need 45kg of battery (assuming a max discharge of 50% which allows to maintain its capacity over the years).

So we are around 150kg just for the arrays and the battery; that seems a lot for a satellite supposed to weight something like 300kg!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: denis on 02/06/2015 12:20 am
Musk specifically said these small sats would have the capability of regular large sats, which means a high power budget.
Isn't the plan to jump ahead in integrated circuit technology? That has a lot of potential to increase capabilities on less power.
That does not represent such a large part of the power budget for current space technology.
For example, this recent On-Board Computer for satellite uses 15W in total: http://www.space-airbusds.com/en/equipment/oscar.html
It's based on this CPU which uses < 2W: http://www.space-airbusds.com/en/equipment/scoc3.html

So by using non space qualified components, they can increase the processing capability, but not reduce the power that much. In any case, we are talking small values compared to the 10kW mentioned here!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/06/2015 12:31 am
A typical off the shelf k-band transmitter, already made for/rated for use in space, is 6 pounds, uses 30 watts and provides 1+gbs - not shabby if you like k-band

http://www2.l-3com.com/tw/pdf/datasheets/ML601_Rev%20D.pdf (http://www2.l-3com.com/tw/pdf/datasheets/ML601_Rev%20D.pdf)

12 would barely be ~30 kilos

I think SpaceX can do better sourcing or making their own though, pack thirty into a single shielded compartment to save on weight and space.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/06/2015 12:55 am
About dynamics: It may be that SpaceX deals with this by just learning how to accept a dynamically more sloppy system.

This would require some rethinking of control algorithms and standard satellite systems engineering, but that sort of thing hasn't stopped them before. There's opportunity for innovation here, I believe. Since the satellites are going to be relying on Hall thrusters, the low thrust would play better dynamically with floppy arrays than high thrust short pulse thrusters typically used on satellites. Also, a neat thing you can do with a phased array (which nevertheless may not be advisable) is that you can compensate for the wiggling of the satellite on-the-fly by adjusting the electronic beam steering (just like adaptive optics but without needing any moving parts).

Remember, SpaceX thinks there's room for significantly advancing the state of the art beyond the typical trade space.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/06/2015 02:34 am
They don't have to last as long as state of the art, nor be as 'perfect' wrt efficiency.
Agreed and agreed, the problem is neither of the above things serves to increase specific power. Your suggestion is economically favorable but cuts the power budget. 15 kw for a few hundred kg still looks pretty unrealistic.

Actually large geo satellite have solar array that generate up to 20kw.
Sure, I was thinking of the small end of the spectrum. I don't think this changes anything important if we're talking about ElonSats at a few hundred kg. Instead of 10x the mass this is 20-30x the mass. Scale it to a few hundred kg and you get the same answer, potentially under a kilowatt.

That does not represent such a large part of the power budget for current space technology.
For example, this recent On-Board Computer for satellite uses 15W in total: http://www.space-airbusds.com/en/equipment/oscar.html
It's based on this CPU which uses < 2W: http://www.space-airbusds.com/en/equipment/scoc3.html
If we're going to talk about the processing power budget we should at least talk about a computer that could handle an ElonSat's workload...

Remember, SpaceX thinks there's room for significantly advancing the state of the art beyond the typical trade space.
What Musk said was that much of the technology that goes to space is quite old compared to terrestrial standards, which is true in a number of areas. However the one field where this is unequivocally false is photovoltaics. This is the one area where ElonSats are virtually certain to be no better or even worse than contemporaries.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/06/2015 02:59 am
What could we do with 1250 watts of solar generated power and a small battery pack?  The Dragon capsule gets about 5000 watts from its dual 21 foot long panels, if you use the same stuff, upgrade its survivability in space but cut it down to a quarter that size deployed Space X gets its sat power system and economies of scale for both projects...

1250 watts , a pair of ~5'*10' panels, batteries

anyone know how much those buggers weigh in at?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/06/2015 03:50 am
Remember, SpaceX thinks there's room for significantly advancing the state of the art beyond the typical trade space.
What Musk said was that much of the technology that goes to space is quite old compared to terrestrial standards, which is true in a number of areas. However the one field where this is unequivocally false is photovoltaics. This is the one area where ElonSats are virtually certain to be no better or even worse than contemporaries.
(I wish you'd read context more carefully because I'm about to repeat a lot of things that either I or someone else just said.) Only if you look just at efficiency and nothing else. Efficiency helps, but specific power is what we're actually looking for, here. And in that area, there is, unequivocally ;) room for at least an order of magnitude improvement over the status quo, if you're willing to look at thin film arrays and novel structural concepts. And as I already said, thin film solar cells aren't just much lighter, they're also much cheaper than your typical Spectrolab triple-junction cells, by orders of magnitude. But feel free to make me repeat it again. :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AdrianW on 02/06/2015 06:54 am
If we're going to talk about the processing power budget we should at least talk about a computer that could handle an ElonSat's workload...
I'm almost 100% sure that the heavy lifting (100s of Gbps of data transfer, digital signal processing for the phased array antennas, etc.) will not be done by processors, but either by FPGAs or custom-built ASICs, which are much, much more power-efficient than general purpose processors for this kind of workload. If I had to guess, I would say that they'll use FPGAs for their first generation of satellites (much easier to prototype, faster time-to-market, lower NRE costs, but high unit costs, power hungrier, and more difficult to radiation-harden than ASICs).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/06/2015 06:56 am
I have no way of knowing how much energy those sats are going to need. However consider this: We are talking about 4000 satellites. Which means even with 1kW/satellite we have Megawatts total transmitting power. Power that can be concentrated dynamically where needed.



Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 02/06/2015 06:58 am
Can someone work out a rough mass budget for a 300-500kg comm sat?  We're seeing roughly 100kg for 10-15kW of solar.  How much would the corresponding batteries weigh---I'm assuming Li-ion is still the best, even in space.  How much for a "reasonable" Hall thruster, plus 5yrs of propellant (Xenon?) in a tank (is it a copv?).  Modern electronics are quite small/light---but how much does a (say) 5kW phased array Ku/Ka band transmitter weigh?  The magnetics tend to resist miniaturization, but high frequency is generally conducive to small/light magnetics.  What else is needed?  Thermal management?  How much does a small heater/radiator weigh?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/06/2015 07:12 am
Only if you look just at efficiency and nothing else.
Disagree, as everyone agrees specific power is the main goal for satellite power sources. This means it is an important result that the fancy multi-junction cells are competitive after decades of various space programs and companies exploring the trades.

And as I already said, thin film solar cells aren't just much lighter, they're also much cheaper than your typical Spectrolab triple-junction cells, by orders of magnitude.
Definitely true neatly rolled up or crumpled uselessly, doesn't seem so clear cut for arrays actually useful for satellites. Didn't Hubble use thin film? The technology has been there for a while without pulling decisively ahead.

You can talk about new algorithms for floppy arrays but now you have to deal with more propellant or more duty cycle on reaction wheels. Every step like this adds significantly to the probability of failure - especially without an abundance of funding and the short timetable.



Something that I think is important to understand is how fragile the process of bringing new technology to market is. Especially if you're not a well established company, those can afford to fail on one attempt or absorb delays. From the timeline and what Musk has said a lot can be inferred. Most things involved are already pretty low risk. Has to be utterly straightforward to be doable in 5 years, a launchable demonstrator a year or two sooner. That's barely enough time to put together a shelf from Ikea. Fitting something really novel into that timeline invites failure.

So most things won't be novel technology, SpaceX's fundamental advantage is that they can bet on the new rules of the launch business (they're betting on it already). Meaning they can rely on commodity ICs, not anything unprecedented, and accept the higher failure rate. If you think about it that's a really direct path to market in the manner of multiple other projects SpaceX has done.

In fact while I wouldn't say ElonSats will be Dragon derived, I think a similar hand will be obvious in the design, and Dragon experience is very important to their confidence in using commodity technology.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 02/06/2015 07:36 am
Having been involved in a project that attempted to do "seven new things" right off the bat---and ultimately managed to make only three or so of those "new things" work---I can echo the appreciation for Musk's ability to balance the long term "new thing" goals against the practicalities of making "version 1" work well enough to fund the future new things.

So I think both sides of this debate are right---I think version 1 is likely to be fairly conservative, with the minimum level of technical innovation needed to prove the general concept, but that the longer-term plan is likely to have many of the more speculative ideas floated here penciled in.

In particular, I'd expect comm sat version 1 to use effectively the same arrays as Dragon (v1 or v2, perhaps the dragon 2 trunk arrays already use the flexible thin film technology?), since that will have been space-proven by then.  But if there are better solar panel technologies out there, I'm sure some R&D group inside SpaceX is already (or shortly will be) working hard to prove them out.

But maybe the power system is actually one of the few critical "new things" needed to make version 1 work.  That's one of the reasons I asked above for a speculative mass budget---what are the really critical drivers here?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/06/2015 07:53 am
They can place prototypes in orbit whenever they want at little to no cost. That is going to be an advantage. They can start with launching an empty satellite bus, just solar panels, thrusters plus something off the shelf to transmit telemetry data. It would likely not be at the later operational altitude but still beats any testbed on earth.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/06/2015 07:59 am
I'm almost 100% sure that the heavy lifting (100s of Gbps of data transfer, digital signal processing for the phased array antennas, etc.) will not be done by processors, but either by FPGAs or custom-built ASICs, which are much, much more power-efficient than general purpose processors for this kind of workload.
Yes, agreed, would be very unusual to try to handle these data rates with a CPU directly.

The budget seems like could credibly attempt a custom ASIC but FPGAs are interesting too. Might be able to reconfigure an FPGA in response to radiation damage, also allows design work to slip without delaying launch.

This is something job postings may help us understand.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/06/2015 12:09 pm
Only if you look just at efficiency and nothing else.
Disagree, as everyone agrees specific power is the main goal for satellite power sources. This means it is an important result that the fancy multi-junction cells are competitive after decades of various space programs and companies exploring the trades.

And as I already said, thin film solar cells aren't just much lighter, they're also much cheaper than your typical Spectrolab triple-junction cells, by orders of magnitude.
Definitely true neatly rolled up or crumpled uselessly, doesn't seem so clear cut for arrays actually useful for satellites. Didn't Hubble use thin film? The technology has been there for a while without pulling decisively ahead.

You can talk about new algorithms for floppy arrays but now you have to deal with more propellant or more duty cycle on reaction wheels. Every step like this adds significantly to the probability of failure - especially without an abundance of funding and the short timetable.



Something that I think is important to understand is how fragile the process of bringing new technology to market is. Especially if you're not a well established company, those can afford to fail on one attempt or absorb delays. From the timeline and what Musk has said a lot can be inferred. Most things involved are already pretty low risk. Has to be utterly straightforward to be doable in 5 years, a launchable demonstrator a year or two sooner. That's barely enough time to put together a shelf from Ikea. Fitting something really novel into that timeline invites failure.

So most things won't be novel technology, SpaceX's fundamental advantage is that they can bet on the new rules of the launch business (they're betting on it already). Meaning they can rely on commodity ICs, not anything unprecedented, and accept the higher failure rate. If you think about it that's a really direct path to market in the manner of multiple other projects SpaceX has done.

In fact while I wouldn't say ElonSats will be Dragon derived, I think a similar hand will be obvious in the design, and Dragon experience is very important to their confidence in using commodity technology.
Your huge certainty about what the satellites will use seems completely misplaced. Such certainty is not at all justified when the stated goal is an enormous improvement in the whole trade space.

I don't have high certainty about what the satellites will use for solar power. I'm pointing out thin film and other tech as a way of introducing a more realistic uncertainty in your own estimates. :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/06/2015 12:23 pm
There is one thing I am positively certain about. They won't be using commercially available space grade arrays unless the producers find a way to sell them a huge lot cheaper than they do now.

They did not for Dragon and would do it much less for that satellite fleet.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 02/06/2015 12:44 pm
I have no way of knowing how much energy those sats are going to need. However consider this: We are talking about 4000 satellites. Which means even with 1kW/satellite we have Megawatts total transmitting power. Power that can be concentrated dynamically where needed.

Huh?   ???
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: abaddon on 02/06/2015 01:59 pm
I have no way of knowing how much energy those sats are going to need. However consider this: We are talking about 4000 satellites. Which means even with 1kW/satellite we have Megawatts total transmitting power. Power that can be concentrated dynamically where needed.

Huh?   ???

Satellite to satellite microwave power relays? :D

Yes I am kidding...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 02/06/2015 02:04 pm
Question: are these projects aiming to supply direct internet connectivity to the final user or need some infrastructures on the ground to distribute it? Cable internet features more connections where there are more users. I wonder how they can manage that from orbit.

Moreover, how to guarantee stable connections inside buildings or in closed environments.

I guess they considered this but I cannot find their answers. Tried to check this thread but is really too large.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/06/2015 02:19 pm
I have no way of knowing how much energy those sats are going to need. However consider this: We are talking about 4000 satellites. Which means even with 1kW/satellite we have Megawatts total transmitting power. Power that can be concentrated dynamically where needed.

Huh?   ???

I guess the Huh? refers to the last sentence. OK, not all of it can -  and should -  be concentrated on one location.  :)

I meant to say that this power need not be evenly distributed.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 02/06/2015 03:12 pm



Disagree, as everyone agrees specific power is the main goal for satellite power sources.

That's only true when launch cost dominates.

If launch cost per kg is low enough, then the price of the cells will begin to matter.

MJ cells, when not under concentration, are insanely expensive.

The metric to shoot for is $/Watt in orbit, deployed.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: sghill on 02/06/2015 04:49 pm
I have no way of knowing how much energy those sats are going to need. However consider this: We are talking about 4000 satellites. Which means even with 1kW/satellite we have Megawatts total transmitting power. Power that can be concentrated dynamically where needed.

Huh?   ???

I guess the Huh? refers to the last sentence. OK, not all of it can -  and should -  be concentrated on one location.  :)

I meant to say that this power need not be evenly distributed.

Oh.  I was hoping you meant thousands of space-based laser weapons platforms.  It's the perfect cover project if Elon finally succumbs and goes full-on super villain on us. This is far less dramatic.

Well only one, or perhaps a few satellites will cover any given spot at the same time, so we're talking tens of Kw to serve customers, not a megawatt.

Honestly, IMHO power needs won't be a problem outside with upwards pointing patch antennas, but I don't expect to see handsets that work (well) inside or in building canyons.  That gets dicey real fast.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/06/2015 05:06 pm
Your huge certainty about what the satellites will use seems completely misplaced. Such certainty is not at all justified when the stated goal is an enormous improvement in the whole trade space.
This seems like putting words in Musk's mouth, where was that stated? Probably the most important indication I'm aware of is here:

http://shitelonsays.com/transcript/musk-says-spacex-will-develop-satellites-in-seattle-2015-01-14

The main source of advancement he's talking about is replacing 5-10 year old electronics with current or leading edge electronics that's not space rated but can probably be made to work. Very important to call attention to the differences in risk and investment with what you're talking about.

Replacing 10 year old electronics with current is a matter of buying it off the shelf, and cutting edge/not in the hands of consumers is a matter of approaching the relevant companies and asking. It may be an allusion to the fact that FPGAs are often the first customers of a new semiconductor node, as their repetitive patterns are some of the first things manufacturers are able to get working. It's an aggressive posture with some risk but not much development effort to fund and very quick to test.

What you're talking about is quite different, a whole new approach to solar has to be developed, it's not immediately ready to fly, and if it doesn't work assumptions about the power budget would change which has ramifications on the rest of the satellite. It's more expensive and it's riskier. And, I don't think Musk actually said enormous improvement in the whole thing. From the other talk it's pretty clear he's explicitly not after an enormous improvement to the whole thing, other aspects of the satellite eg thrusters are pretty conventional. They'll do them in house because it's expensive to buy them externally, but conventional nonetheless.

http://shitelonsays.com/transcript/spacex-seattle-2015-2015-01-15

The "enormous improvement in the whole trade space" seems to come from you, without that there's very little SpaceX has to make from whole cloth which is exactly the right thing to do for a satellite they want to launch in 5 years.

That's only true when launch cost dominates.

If launch cost per kg is low enough, then the price of the cells will begin to matter.
Yup, I think this is something where SpaceX won't be any better and may actually regress compared to contemporary satellites.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 02/06/2015 05:07 pm
 ::) ;D 8)
If  Lockheed Skunkworks CFR could get up in time it would sure beat solar cells on powerlimits for these sats.
 8) 8) 8)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/06/2015 05:07 pm
Honestly, IMHO power needs won't be a problem outside with upwards pointing patch antennas, but I don't expect to see handsets that work (well) inside or in building canyons.  That gets dicey real fast.

That is not what they plan to service. Not inside buildings was explicitly said. Also that they don't plan to service the population centers. That's where cable/fibre and cell towers are better.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Eer on 02/06/2015 05:39 pm
Question: are these projects aiming to supply direct internet connectivity to the final user or need some infrastructures on the ground to distribute it? Cable internet features more connections where there are more users. I wonder how they can manage that from orbit.

Moreover, how to guarantee stable connections inside buildings or in closed environments.

I guess they considered this but I cannot find their answers. Tried to check this thread but is really too large.
I think everyone is wondering where it all is going.  This (http://www.washingtonpost.com/blogs/the-switch/wp/2015/02/06/google-is-serious-about-taking-on-telecom-heres-why-itll-win/ ) article at WaPo talks about Google working with Sprint and T-Mobile, but doesn't mention SpaceX nor the satellite constellation.  Wonder if it should have?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ludus on 02/07/2015 12:39 am
Question: are these projects aiming to supply direct internet connectivity to the final user or need some infrastructures on the ground to distribute it? Cable internet features more connections where there are more users. I wonder how they can manage that from orbit.

Moreover, how to guarantee stable connections inside buildings or in closed environments.

I guess they considered this but I cannot find their answers. Tried to check this thread but is really too large.
I think everyone is wondering where it all is going.  This (http://www.washingtonpost.com/blogs/the-switch/wp/2015/02/06/google-is-serious-about-taking-on-telecom-heres-why-itll-win/ ) article at WaPo talks about Google working with Sprint and T-Mobile, but doesn't mention SpaceX nor the satellite constellation.  Wonder if it should have?
It's pretty speculative but I think it fits very well. It's sort of the moon shot version of Google fiber. If SpaceX can pull off direct competition with the telcoms at a fraction of the capital cost they have a piece of it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/07/2015 12:51 am
Building out your own cell network is a lot easier if you don't have to run fiber all over (over even power lines, with battery-backed solar).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/08/2015 07:07 am
Something I'd be interested in is what other applications this spacecraft bus will be useful for.

It seems like for the initial version the unit price of each satellite is likely under a million, and that's with communications gear that would likely be hundreds of thousands on the ground. Very economical. At a few hundred kg it's also a great candidate for a secondary payload. This makes it seem like even a mid size lab at a mid size school could contemplate interesting missions, I'm sure there's plenty of science that hasn't been done yet either because the technology is new or because it's never been at the front of the line in terms of funding. Examples might be astronomy in wavelengths that don't make it through the atmosphere.

The radiation environment might limit the usefulness BEO with the lack of hardening, but even then there would be an opportunity to try new lower cost hardening technology. An ElonSat could easily catch a ride BEO as a secondary payload on a Falcon Heavy GTO launch.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 02/08/2015 07:18 am
There is one thing I am positively certain about. They won't be using commercially available space grade arrays unless the producers find a way to sell them a huge lot cheaper than they do now.

They did not for Dragon and would do it much less for that satellite fleet.

For Dragon - where it has to last a week in space - a number of things are utterly unimportant.
Radiation damage, for example, and thermal cycling are almost a non-issue.

You actually can use ground-based cells with little or no modification.

Where you may want to have long life (>>5y to 50% output), and want them to cope with radiation, and thermal cycling for that long - things may differ a bit.

Plus - If you're actually willing to plop down a wadge of cash for ten megawatts or so of triple junction cells up-front, things might change a bit.
Of course - on the other side - triple junction cells work well for solar concentrated arrays on earth, and having the capability to make those would be a benefit for other Musk buisnesses.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/08/2015 07:30 am
There is one thing I am positively certain about. They won't be using commercially available space grade arrays unless the producers find a way to sell them a huge lot cheaper than they do now.

They did not for Dragon and would do it much less for that satellite fleet.

For Dragon - where it has to last a week in space - a number of things are utterly unimportant.
Radiation damage, for example, and thermal cycling are almost a non-issue.

They proposed using Dragon as Dragonlab with 2 years in orbit. The manned version staying in orbit for at least 6 months in the function of emergency escape.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 02/08/2015 07:31 am
Most of those sats will be idling while flying over inhabited areas, while a few will be struggling for a few minutes with zillions of high-demanding users.

I can't see a sat internet network siding ground based competitors. I would rather see sats network backing up fibers and radio ones in semi-deserted areas. That would make price very high.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 02/08/2015 09:48 am
There may be areas where SpaceX can jump ahead of contemporary satellite technology by space qualifying the most advanced terrestrial stuff, but solar power is not one of them. That may actually regress, and there's a lot less mass budget to work with.

I do expect SpaceX to attempt 10-15kW power for each Sat, but I doubt they will be buying Boeing Spectrolab panels. I assume SpaceX would probably attempt to use Solar City/Silveo solar panels and space rate them.


If launch cost per kg is low enough, then the price of the cells will begin to matter.
Yup, I think this is something where SpaceX won't be any better and may actually regress compared to contemporary satellites.
[/quote]

Spectrolab must be supplying a pretty small market today.

If SpaceX offered them 10,000 orders (4k of sats + replenished twice), they might be willing to work with SpaceX on a roadmap towards high performance and low per-unit cost.

I'm not really sure if the size of the market SpaceX could offer will be enough to kick start new industries of this sort, or if Spectrolab could already access this sort of funding just due to the size of the Solar Panel market.


In a similar vein, if they are ordering many thousands of computers, might it be worthwhile to work with someone to produce rad-hardened CPUs and other components. Again, I'm not really clear if today's market for rad-hardened is already much bigger than this, or if this could be an opportunity for them.

Of course, SpaceX handles this with redundancy today, but there may be a mass optimisation available here. Also, Dragon can just reset a computer that produces a spurious result, which is fine for a computer that's producing a small number of external outputs (tens / hundreds per second???) But, I'm not sure this strategy translates so well into cross-checking the output of computers handling gbps. Not to mention the weight and power demands of running multiple heavyweight systems just to cross-check the results.

Of course, other strategies may be possible - parity checking, packet counting, and just existing retry-if-it-doesn't-get-there protocols.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/09/2015 02:33 am
If SpaceX offered them 10,000 orders (4k of sats + replenished twice), they might be willing to work with SpaceX on a roadmap towards high performance and low per-unit cost.
I'm sure it would be worth exploring. There's likely also a number of choices that are not cheap enough for consumer applications quite yet but would be fine for SpaceX.

As you say a large purchase agreement tends to open doors to things even the vendor didn't know were for sale.

In a similar vein, if they are ordering many thousands of computers, might it be worthwhile to work with someone to produce rad-hardened CPUs and other components.
Similarly agreed.

Again, I'm not really clear if today's market for rad-hardened is already much bigger than this, or if this could be an opportunity for them.
Could go either way I guess. Much of SpaceX's experience suggests established vendors may be resistant to fire sale prices even if they could theoretically reach them.

I read some of this: http://en.wikipedia.org/wiki/Radiation_hardening

And what SpaceX has done with Dragon here: http://aviationweek.com/blog/dragons-radiation-tolerant-design

It seems like a few relatively simple steps can improve tolerance quite a bit. For example choosing a SOI process for the chip (which many fabs use anyway) seems to help. The avionics would want to do more than the communications payload.

Of course, SpaceX handles this with redundancy today, but there may be a mass optimisation available here.
4000 satellites with multiple chips is getting to the sort of scale where an ASIC makes sense, this is plausible. You can build multiple cores onto a single chip and have them do their votes on the chip. Still need multiple chips but I suspect the part count can be reduced compared to the current Dragon system.

Also, Dragon can just reset a computer that produces a spurious result, which is fine for a computer that's producing a small number of external outputs (tens / hundreds per second???)
I think the worry is some radiation event that causes permanent damage. What's the mean time to failure of the computers they'll use in the orbital radiation environment? What about after a CME?

But, I'm not sure this strategy translates so well into cross-checking the output of computers handling gbps.
I don't think the communications payload needs it.

Of course, other strategies may be possible - parity checking, packet counting, and just existing retry-if-it-doesn't-get-there protocols.
Protocols used on the internet already tolerate packet loss and already do checksums. It happens all the time due to congestion if not space weather. If packet loss can be kept below about one out of every few thousand I don't think it'll be a problem for the service. Meaning individual radiation events can be essentially ignored, the problem will be if damage occurs that creates persistent problems.

I suspect the communications payload can have some extra capacity, and if part of that is persistently damaged then that chip can be shut off. Or maybe have a few duplicate paths on a chip and fuse one of them off. It's like flash memory, it always fails eventually, it's designed so that bad blocks are detected and shut down automatically.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: hrissan on 02/09/2015 09:15 pm
Most of those sats will be idling while flying over inhabited areas, while a few will be struggling for a few minutes with zillions of high-demanding users.

I can't see a sat internet network siding ground based competitors. I would rather see sats network backing up fibers and radio ones in semi-deserted areas. That would make price very high.
Those sats over pacific/atlantic are not idling, but serve as the backbone/transocean bridge for the "over land" sats. And if you think about how many paths via different sets of sats you can have over pacific (assuming the direct links can be established between sats orbiting not adjacent to each other), you realize Elon's desire to compete with transatlantic fiber...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/10/2015 03:49 am
Most of those sats will be idling while flying over inhabited areas, while a few will be struggling for a few minutes with zillions of high-demanding users.

I can't see a sat internet network siding ground based competitors. I would rather see sats network backing up fibers and radio ones in semi-deserted areas. That would make price very high.
Those sats over pacific/atlantic are not idling, but serve as the backbone/transocean bridge for the "over land" sats. And if you think about how many paths via different sets of sats you can have over pacific (assuming the direct links can be established between sats orbiting not adjacent to each other), you realize Elon's desire to compete with transatlantic fiber...
Very good point.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pagheca on 02/10/2015 07:23 am
Without the specifics of the project is difficult to say, but with 4,000 satellites moving on the sky there would be on average 1 every 130,000 km2. Let's do a tenth by optimization of orbital parameters to improve coverage of low-latitude areas. Still 13,000 km2. NYC area is, for comparison, 1,214 km2.

So, you may have just a bunch of satellites covering NYC (several of them may overlap). Free to think this is comparable the way ground based internet works (more nodes => more connections where needed) as a stand-alone, user-to-user solution, but I will not buy it without further confirmations.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 02/10/2015 08:23 am
Most of those sats will be idling while flying over inhabited areas, while a few will be struggling for a few minutes with zillions of high-demanding users.

I can't see a sat internet network siding ground based competitors. I would rather see sats network backing up fibers and radio ones in semi-deserted areas. That would make price very high.
Those sats over pacific/atlantic are not idling, but serve as the backbone/transocean bridge for the "over land" sats. And if you think about how many paths via different sets of sats you can have over pacific (assuming the direct links can be established between sats orbiting not adjacent to each other), you realize Elon's desire to compete with transatlantic fiber...
If they do go high speed laser or such between sats, this could make the undersea cable operators pretty nervous.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ChrisWilson68 on 02/10/2015 09:05 am
It seems like a few relatively simple steps can improve tolerance quite a bit. For example choosing a SOI process for the chip (which many fabs use anyway) seems to help. The avionics would want to do more than the communications payload.

[...]

4000 satellites with multiple chips is getting to the sort of scale where an ASIC makes sense, this is plausible. You can build multiple cores onto a single chip and have them do their votes on the chip. Still need multiple chips but I suspect the part count can be reduced compared to the current Dragon system.

I think you are greatly underestimating the cost of doing an ASIC comparable to off-the shelf multi-processor DSP chips.  An enormous amount of engineering goes into optimizing a chip like that.  4000 is a tiny, tiny volume for an ASIC.  And it's not just the cost (maybe $100 million to design an ASIC like that), it's the time it takes.  If they have to build a chip design team from scratch, don't expect the first production chips for 3 years.  There's also the risk -- they might not find out until 2 years in that their chip won't have the performance they projected.  Meanwhile, TI comes out with a new, improved multi-chip DSP that beats the ASIC.

The chips are tiny and very cheap.  This is exactly the kind of thing where SpaceX can leverage its cheap launch costs -- just use cheap off-the-shelf chips and pay for the extra mass it takes to have lots of redundant chips, shielding, or both.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: speedevil on 02/11/2015 11:07 am
The chips are tiny and very cheap.  This is exactly the kind of thing where SpaceX can leverage its cheap launch costs -- just use cheap off-the-shelf chips and pay for the extra mass it takes to have lots of redundant chips, shielding, or both.

Also, in principle, because most of the time it's not over a major metropolitan area, you could have 90% of the chips 'idle' - and using no power.
You only burst it to 20* nominal power when you really, really need it.
The power and thermal requirements are somewhat easier if you only need to do this once every 5 orbits (say) for 8 minutes.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 02/11/2015 12:51 pm
It seems like a few relatively simple steps can improve tolerance quite a bit. For example choosing a SOI process for the chip (which many fabs use anyway) seems to help. The avionics would want to do more than the communications payload.

[...]

4000 satellites with multiple chips is getting to the sort of scale where an ASIC makes sense, this is plausible. You can build multiple cores onto a single chip and have them do their votes on the chip. Still need multiple chips but I suspect the part count can be reduced compared to the current Dragon system.

I think you are greatly underestimating the cost of doing an ASIC comparable to off-the shelf multi-processor DSP chips.  An enormous amount of engineering goes into optimizing a chip like that.  4000 is a tiny, tiny volume for an ASIC.  And it's not just the cost (maybe $100 million to design an ASIC like that), it's the time it takes.  If they have to build a chip design team from scratch, don't expect the first production chips for 3 years.  There's also the risk -- they might not find out until 2 years in that their chip won't have the performance they projected.  Meanwhile, TI comes out with a new, improved multi-chip DSP that beats the ASIC.

The chips are tiny and very cheap.  This is exactly the kind of thing where SpaceX can leverage its cheap launch costs -- just use cheap off-the-shelf chips and pay for the extra mass it takes to have lots of redundant chips, shielding, or both.

Whilst the timescales are about right, you can do an ASIC for a lot less than that. Tapeout costs about $1M depending on process, so you got to add on design,layout, testing, software etc. I'm working in a team that does this sort of stuff (Although I'm software side), there are 7 of us, so not huge in labour costs. The chip we are doing will cost probably $3M to develop. And will be quite complex...but we never start from scratch, there are always libraries you bring in to save time. As an existing example, the Brcm2836, the chip in the new Raspberry Pi, took about 1.5yrs to develop, with a team of about 10. However, that reused a lot of work from the previous chip.

So the cost does depend on how much reuse you can have, and therefore how much new work is required, and how much of it there is to do. But $100M is much too high, probably double.

And, tbh, you would do this in an FPGA anyway, an ASIC would be much more expensive for the low volumes involved.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/11/2015 01:15 pm
BTW, I really doubt they'll develop their own ASIC.

But the volume will be much higher than 4000. They'll have at a very minimum 3 times that due to their redundancy/reliability strategy, and if you add in multiple chips to handle the workload, you easily could be talking over a dozen chips per satellite. Dragon uses dozens of chips, for instance.

So you could be talking 50k-100k chips (though likely of various types) for each constellation iteration.

But it's true that they'd want to use state of the art (non-space-sector-specialized) chips in order to stay on the cutting edge without having to invest a lot every year.

But where a custom ASIC (customized from a licensed existing ASIC design, most likely) would make sense is for the ground terminals, since there would be millions of them, perhaps eventually hundreds of millions.

But they'll need to invest in custom phased array technology (on both ends) which may itself be chip-scale. But given the number of elements needed (hundreds of thousands or even a million per satellite), they'd easily have enough volume to justify custom chips.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 02/11/2015 03:03 pm
I would think they would need make in excess of a million to make the decision to go with an ASIC, prior to that any recent FPGA would do (assuming they are robust enough).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/11/2015 03:27 pm
I think you are greatly underestimating the cost of doing an ASIC comparable to off-the shelf multi-processor DSP chips.
Well, there's two main areas. The avionics and the communications payload. Both have potential advantages to doing custom chips.

For the avionics it's mostly a matter of shrinking the overhead necessary for redundant voting computers, so it would be mostly pretty simple processors, licensable stuff. I thought about whether the layout and design rules might be different for radiation tolerance, but even if they are the actual computation power needed for this isn't huge, so simple synthesizable cores are viable options, so most of the changes here would happen via automated tools. This isn't the same as a highly optimized design.

It's different for the communications payload, but that's also a lot more forgiving of errors so COTS makes more sense. If customization were desired here, again, a lot of it is probably licensable. The customization might even mean going to TI and asking if they can do a custom version of a DSP with the same cores but slightly different features, like maybe spare cores for failures that don't happen on the ground, and going outside the normally economical die size envelope.

So the cost does depend on how much reuse you can have, and therefore how much new work is required, and how much of it there is to do. But $100M is much too high, probably double.
For the avionics it seems like it would be quite simple ARM cores and some SRAM and some communications to implement the redundant/voting architecture Dragon uses with less mass, power, and without a pressurized compartment.

At a few hundred kg I'd expect most of that to go to power, propulsion, antenna, with most of the rest going towards the communications payload. If a relatively easy ASIC can improve radiation tolerance while reducing a significant circuit board to 1-2 chips that may be worth it. Obviously subject to a better analysis of the trades than I can provide. But I do think this is going to have to optimize mass more aggressively than Dragon does.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/11/2015 11:02 pm
I would think they would need make in excess of a million to make the decision to go with an ASIC, prior to that any recent FPGA would do (assuming they are robust enough).
Considering there probably isn't a chip scale phased array solution that meets their needs, that seems like a fairly easy number for them to hit (not that the development effort will be easy!). They're going to need billions of phased array elements, total, for the constellation and billions more for the ground terminals.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/11/2015 11:26 pm
But the volume will be much higher than 4000. They'll have at a very minimum 3 times that due to their redundancy/reliability strategy, and if you add in multiple chips to handle the workload, you easily could be talking over a dozen chips per satellite. Dragon uses dozens of chips, for instance.
What made me think it might be worth it it is that Dragon has an embarrassment of excess performance and a handy pressure vessel to put everything in, and the computers have much less work to do. A satellite as mass sensitive as what's been proposed may want to shrink this.

Could mean limited customization of existing IP blocks or special orders from existing companies rather than doing something from whole cloth.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 02/12/2015 12:07 pm
I would think they would need make in excess of a million to make the decision to go with an ASIC, prior to that any recent FPGA would do (assuming they are robust enough).
Considering there probably isn't a chip scale phased array solution that meets their needs, that seems like a fairly easy number for them to hit (not that the development effort will be easy!). They're going to need billions of phased array elements, total, for the constellation and billions more for the ground terminals.

Slightly different from the FPGA/ASIC stuff though - this is the transmission side rather than the processing side.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/12/2015 02:08 pm
I would think they would need make in excess of a million to make the decision to go with an ASIC, prior to that any recent FPGA would do (assuming they are robust enough).
Considering there probably isn't a chip scale phased array solution that meets their needs, that seems like a fairly easy number for them to hit (not that the development effort will be easy!). They're going to need billions of phased array elements, total, for the constellation and billions more for the ground terminals.

Slightly different from the FPGA/ASIC stuff though - this is the transmission side rather than the processing side.
Well yeah, you can't use an FPGA for that. Has to be ASIC.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: gospacex on 02/12/2015 02:28 pm
There is one thing I am positively certain about. They won't be using commercially available space grade arrays unless the producers find a way to sell them a huge lot cheaper than they do now.

What is the price?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/13/2015 02:30 am
There is one thing I am positively certain about. They won't be using commercially available space grade arrays unless the producers find a way to sell them a huge lot cheaper than they do now.

What is the price?
I think it's >>$100/Watt, which would be a large portion of the satellite costs if using those traditional cells. Dragon's arrays are most certainly using cheaper cells than traditional.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/14/2015 02:05 pm
Going into the future a bit does anyone expect non solar power to be used on the sats around Mars?  It seems a perfect time to use an ASRG rather than Solar.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/14/2015 02:26 pm
Going into the future a bit does anyone expect non solar power to be used on the sats around Mars?  It seems a perfect time to use an ASRG rather than Solar.

Europa Clipper is being baselined with solar power... Jupiter is at 5.2AU.  Mars is at 1.5AU.
Inverse square law... Mars solar = 12x Jupiter solar.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/14/2015 02:32 pm
Going into the future a bit does anyone expect non solar power to be used on the sats around Mars?  It seems a perfect time to use an ASRG rather than Solar.

Europa Clipper is being baselined with solar power... Jupiter is at 5.2AU.  Mars is at 1.5AU.
Inverse square law... Mars solar = 12x Jupiter solar.

Which was a cost decision....old space.  18 square meter solar panels are a bit hard to repair on something around Mars if you want them to work for years eh?


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/14/2015 02:46 pm

18 square meter solar panels are a bit hard to repair on something around Mars if you want them to work for years eh?

The likelihood that something goes wrong with  a mechanical system over many years is orders of maginitude higher than for solar. Also solar would have likely at least two panels which leaves you with still half power if something goes wrong with one.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/14/2015 09:50 pm
Which was a cost decision....
If this is meant to suggest SpaceX won't make cost decisions for Mars satellites I think you're wrong.

18 square meter solar panels are a bit hard to repair on something around Mars if you want them to work for years eh?
No harder than Juno's 24 m2 panels around Jupiter, or any of the current Mars orbiters.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/15/2015 11:28 pm
Which was a cost decision....
If this is meant to suggest SpaceX won't make cost decisions for Mars satellites I think you're wrong.

18 square meter solar panels are a bit hard to repair on something around Mars if you want them to work for years eh?
No harder than Juno's 24 m2 panels around Jupiter, or any of the current Mars orbiters.

ESA and NASA are constrained by budget.  if Elon decides he'd rather have a 100 year power plant (Amerecium 241 has a half life od 400 years; in a stirling engine/rtg based power solution 100 years is not so hard)  in a Mars orbiting sat rather than solar panels he has the $ to do so, the ESA was not so lucky: they sought and were DENIED the budget to go RTG/ARTG.

edited: for spelling 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/15/2015 11:35 pm
How are we talking about RTGs, for a freaking Mars orbiter of all things? No way in heck that a Mars orbiter would be anything but solar at this point. Solar is just way more powerful (for a given mass), way cheaper, will generally last longer (don't forget degradation of the thermocouples or whatever of your RTG!), doesn't have dangerous radiation (most relevant for Americium), doesn't have the complication of needing to reject heat while sitting on the ground awaiting launch, and doesn't have any political or public relations risk.

Also, Musk is very bullish on solar and batteries, sitting on the board of SolarCity and being CEO of Tesla.

An RTG wouldn't make any sense for a Mars orbiter. Not. One. Lick.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/16/2015 12:31 am
How are we talking about RTGs, for a freaking Mars orbiter of all things? No way in heck that a Mars orbiter would be anything but solar at this point. Solar is just way more powerful (for a given mass), way cheaper, will generally last longer (don't forget degradation of the thermocouples or whatever of your RTG!), doesn't have dangerous radiation (most relevant for Americium), doesn't have the complication of needing to reject heat while sitting on the ground awaiting launch, and doesn't have any political or public relations risk.

Also, Musk is very bullish on solar and batteries, sitting on the board of SolarCity and being CEO of Tesla.

An RTG wouldn't make any sense for a Mars orbiter. Not. One. Lick.

This is a thread about...satellites.  Communication sats that will go around Earth and Mars.  While Earth might have a problem with rtg/srg sats Mars isn't so easy to send out a repair man or put up a new sat on a short timeline.

Satellites  ...  for use by sats, not a mars orbiter with people on board ffs.  SRG (http://en.wikipedia.org/wiki/Stirling_radioisotope_generator (http://en.wikipedia.org/wiki/Stirling_radioisotope_generator)  way more efficient than old RTG with thermocouples).

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/16/2015 01:12 am
Mars isn't so easy to send out a repair man or put up a new sat on a short timeline.
Not so easy for Earth either last time I checked. The only quick replacement strategy for Earth orbit is spares already in orbit.

ESA and NASA are constrained by budget.
SpaceX is far more constrained. Looks like SpaceX wants to be able to do satellites for less than ESA or NASA can even think about satellites. This is the opposite of not caring how expensive it is.

if Elon decides he'd rather have a 100 year power plant
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/16/2015 03:00 am
How are we talking about RTGs, for a freaking Mars orbiter of all things? No way in heck that a Mars orbiter would be anything but solar at this point. Solar is just way more powerful (for a given mass), way cheaper, will generally last longer (don't forget degradation of the thermocouples or whatever of your RTG!), doesn't have dangerous radiation (most relevant for Americium), doesn't have the complication of needing to reject heat while sitting on the ground awaiting launch, and doesn't have any political or public relations risk.

Also, Musk is very bullish on solar and batteries, sitting on the board of SolarCity and being CEO of Tesla.

An RTG wouldn't make any sense for a Mars orbiter. Not. One. Lick.

This is a thread about...satellites.  Communication sats that will go around Earth and Mars.  While Earth might have a problem with rtg/srg sats Mars isn't so easy to send out a repair man or put up a new sat on a short timeline.

Satellites  ...  for use by sats, not a mars orbiter with people on board ffs.  SRG (http://en.wikipedia.org/wiki/Stirling_radioisotope_generator (http://en.wikipedia.org/wiki/Stirling_radioisotope_generator)  way more efficient than old RTG with thermocouples).
Except that would be heavier and more expensive than just a solar array.

And if you want to use an SRG, you're now talking about moving parts. Better than RTG, but with moving parts greater probability of failure.

Anyway, it's simply a bad idea. It's less bad for surface power.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JamesH on 02/16/2015 11:40 am
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.

Er, takes a lot more time and money to replace a satellite around Mars than one around the Earth.  That completely changes the economics.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/16/2015 05:37 pm
Have we completely forgotten the KISS principle?
Solar power -- simple
SRG -- not simple
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/16/2015 06:59 pm
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.
Er, takes a lot more time and money to replace a satellite around Mars than one around the Earth.  That completely changes the economics.
I don't think it gets within an order of magnitude of relevant decision thresholds.

Put the $500k satellite on the rocket that launches people to Mars for $500k. Time to replace is easily dismissed because this is so cheap we just send new ones every window, a more expensive satellite with longer expected life just makes you that much more vulnerable to failure.

Some people have this strange notion that solar energy is an interim solution that will be superseded by something else once a real power source is needed. I think these people are going to spend the remainder of the 21st century feeling frustrated and confused.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/24/2015 03:02 am
This narrows the field in what Spacex is contemplating:

http://www.fiercewireless.com/tech/story/5g-proceeding-spacex-urges-fcc-protect-future-satellite-ventures/2015-02-22 (http://www.fiercewireless.com/tech/story/5g-proceeding-spacex-urges-fcc-protect-future-satellite-ventures/2015-02-22)

And this:

http://apps.fcc.gov/ecfs/document/view?id=60001030927 (http://apps.fcc.gov/ecfs/document/view?id=60001030927)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 02/24/2015 03:05 am
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.
Er, takes a lot more time and money to replace a satellite around Mars than one around the Earth.  That completely changes the economics.
I don't think it gets within an order of magnitude of relevant decision thresholds.

Put the $500k satellite on the rocket that launches people to Mars for $500k. Time to replace is easily dismissed because this is so cheap we just send new ones every window, a more expensive satellite with longer expected life just makes you that much more vulnerable to failure.

Some people have this strange notion that solar energy is an interim solution that will be superseded by something else once a real power source is needed. I think these people are going to spend the remainder of the 21st century feeling frustrated and confused.

And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 02/24/2015 12:35 pm
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.
Er, takes a lot more time and money to replace a satellite around Mars than one around the Earth.  That completely changes the economics.
I don't think it gets within an order of magnitude of relevant decision thresholds.

Put the $500k satellite on the rocket that launches people to Mars for $500k. Time to replace is easily dismissed because this is so cheap we just send new ones every window, a more expensive satellite with longer expected life just makes you that much more vulnerable to failure.

Some people have this strange notion that solar energy is an interim solution that will be superseded by something else once a real power source is needed. I think these people are going to spend the remainder of the 21st century feeling frustrated and confused.

And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Expensive ones have. Still much cheaper than a radioisotope power supply and with greater power for the same mass even at Mars.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 02/24/2015 01:57 pm
And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Expensive ones have. Still much cheaper than a radioisotope power supply and with greater power for the same mass even at Mars.

They will need to be conditioned for Mars. But that is not too hard. A micrometer glass coating - like done on plastic beer bottles - for abrasion protection. Maybe an additional UV filter as well. that is done on cheap transparent plastic roofs so they stand up for many more years. Both very standard very cheap processes done in mass production.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ArbitraryConstant on 02/24/2015 02:31 pm
That seems unlikely. They'd rather have a half million dollar satellite that lasts 5 years with frequent replacement and lots of spares. Can't see why Mars would be any different than Earth.
Er, takes a lot more time and money to replace a satellite around Mars than one around the Earth.  That completely changes the economics.
I don't think it gets within an order of magnitude of relevant decision thresholds.

Put the $500k satellite on the rocket that launches people to Mars for $500k. Time to replace is easily dismissed because this is so cheap we just send new ones every window, a more expensive satellite with longer expected life just makes you that much more vulnerable to failure.

Some people have this strange notion that solar energy is an interim solution that will be superseded by something else once a real power source is needed. I think these people are going to spend the remainder of the 21st century feeling frustrated and confused.

And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Wouldn't be surprised if they could, I think the radiation environment attacking the cheap computers is probably a bigger constraint on how long they last.

Either way, AFAICT Musk thinks satellites that last for decades are a false economy. Build in an aggressive replacement schedule and make them cheap.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 02/24/2015 05:30 pm
And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Expensive ones have. Still much cheaper than a radioisotope power supply and with greater power for the same mass even at Mars.

They will need to be conditioned for Mars. But that is not too hard. A micrometer glass coating - like done on plastic beer bottles - for abrasion protection. Maybe an additional UV filter as well. that is done on cheap transparent plastic roofs so they stand up for many more years. Both very standard very cheap processes done in mass production.

Solar City is currently building an advanced  tech solar cell "gigafactory" in Buffalo, NY with plans for two more even larger ones. They'll have much better than "cheap" solar arrays at Mars.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 02/25/2015 12:11 am
And some people seem to think cheap solar panels will last for decades around the Red planet.  Dream on.
Expensive ones have. Still much cheaper than a radioisotope power supply and with greater power for the same mass even at Mars.

They will need to be conditioned for Mars. But that is not too hard. A micrometer glass coating - like done on plastic beer bottles - for abrasion protection. Maybe an additional UV filter as well. that is done on cheap transparent plastic roofs so they stand up for many more years. Both very standard very cheap processes done in mass production.

Solar City is currently building an advanced  tech solar cell "gigafactory" in Buffalo, NY with plans for two more even larger ones. They'll have much better than "cheap" solar arrays at Mars.

So ya think they might know something about  solar cells?  And batteries?  And comm satellites? And electric vehicles? And enclosed tube/low pressure transportation?

We should just assume that when one of these known entities is involved, they will default to in-house (broadly speaking) technologies instead of exotics.  These fora frequently argue for sexier exotics, but SpaceX has shown repeatedly that they are using what is known and proven... maybe stretching the tech aspects but mostly applying it in novel ways.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 02/25/2015 12:14 am
What I think is that they'll explore using in-the-family solutions first, and if they'll work for the projected lifetime of the satellite or spacecraft that's where they'll go.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 02/25/2015 08:47 am
What I think is that they'll explore using in-the-family solutions first, and if they'll work for the projected lifetime of the satellite or spacecraft that's where they'll go.

Agreed - they have a long, proven and successful history of not going outside if the inside has a longer evolutionary tree.

It 'ain't broken, so they won't fix it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 02/27/2015 04:59 pm
they have a long, proven and successful history of not going outside if the inside has a longer evolutionary tree.
That was very well put! 

The distillate of this thread imo.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 03/03/2015 06:46 pm
Does anyone know if SpaceX owns the rights to the dispenser tech used for Iridium NEXT?  SpaceX was charged, as part of the contract, to develop/build the dispenser. 

http://www.iridium360.com/wp-content/uploads/2014/07/IRIDIUM_Test_Prep_183_KHarris.jpg (http://www.iridium360.com/wp-content/uploads/2014/07/IRIDIUM_Test_Prep_183_KHarris.jpg)
^ Large pic.

This monster holds 10 800kilo Iridium NEXT Sats inside one fairing.

How many could SpaceX, using the same general configuration, fit into there if the sats wee only 200ish kilos?  20 to 40? 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cro-magnon gramps on 03/03/2015 08:35 pm
Does anyone know if SpaceX owns the rights to the dispenser tech used for Iridium NEXT?  SpaceX was charged, as part of the contract, to develop/build the dispenser. 

http://www.iridium360.com/wp-content/uploads/2014/07/IRIDIUM_Test_Prep_183_KHarris.jpg (http://www.iridium360.com/wp-content/uploads/2014/07/IRIDIUM_Test_Prep_183_KHarris.jpg)
^ Large pic.

This monster holds 10 800kilo Iridium NEXT Sats inside one fairing.

How many could SpaceX, using the same general configuration, fit into there if the sats wee only 200ish kilos?  20 to 40?

the link is saying there is something wrong with the image; so I googled and got a screen shot of the Thumbnail,
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 03/03/2015 10:59 pm
Full image
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: GalacticIntruder on 03/10/2015 08:12 pm
There have been two articles recently claiming vaporware sat companies and or ITU speculators are screwing up the gears in satellite bandwidth filings. I guess that is downside for being high-profile and public.

http://motherboard.vice.com/read/spacex-warns-paper-satellites-could-disrupt-its-space-internet-plan (http://motherboard.vice.com/read/spacex-warns-paper-satellites-could-disrupt-its-space-internet-plan)

http://spacenews.com/proposed-leosat-constellation-aimed-at-top-3000/ (http://spacenews.com/proposed-leosat-constellation-aimed-at-top-3000/)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 03/10/2015 11:41 pm
Recalling the pad lease saga and the ridiculous patent, is Bezos known to be involved in one of the announced competing constellations?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/13/2015 03:34 pm
Bumping up the thread because several other ones veered to talk about the constellation.

One upside of not doing direct-to-handset is that you're not getting into war with the cell carriers.

You do station-2-station, and actually help them by doing tower-2-backbone.

It is easier to deal with fewer larger endpoints than with billions of individual cell phones.  Might be 100 or a 1000 less entities to deal with for the same amount of bandwidth.

Clever.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/13/2015 05:08 pm
Bumping up the thread because several other ones veered to talk about the constellation.

One upside of not doing direct-to-handset is that you're not getting into war with the cell carriers.

Agreed.

Quote
You do station-2-station, and actually help them by doing tower-2-backbone.

Unless you are talking about serving within remote areas (and they are), this is practical with with 1-way or "non-immediate" communications.  There is little need for it in populated areas.

Quote
It is easier to deal with fewer larger endpoints than with billions of individual cell phones.

This is only kinda-sorta true.  If you are only dealing with a few receiving entities, you're dealing with entities that are large and can afford services already in existence.  This scenario wouldn't pay for your constellation and is a solution in search of a problem.

Quote
Might be 100 or a 1000 less entities to deal with for the same amount of bandwidth.

Session handling done by computers doesn't care; it's a non-issue as far as computing power goes.  However, as far as bandwidth utilization goes, the utilization schemes could greatly favor single sessions or not.  It depends on what they go with.
I was talking about tower-level end stations.  Each tower is handling 100-1000 phones at a time?
 In remote areas, still use towers, but they are stand alone, requiring only power.

I think for a remote tower, the cost of running a landline to it is more than the cost of the tower, and microwave link chains are not cheap either. Here, one hop and you're on the backbone.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Dudely on 03/13/2015 05:56 pm
I posted this on another thread, but it should really go here. It explains the financial and economic feasibility of the proposed constellation, which as Elon said will aim to serve up to 90% of all "server-to-server" traffic. They don't even need to finish most of the ground stations before they would be making money.


If you can provide a huge, direct, low latency pipe between two servers on opposite ends of a continent you can sell that pipe at a premium because it is a much lower latency connection than the "regular" connection which is routed through hundreds of switches and routers. This sort of transaction is constantly happening between ISPs and tech companies all over the world. The company I work for is a great example. We buy a direct link from our main servers in Ontario to our satellite office on the East coast where we do development. This link is jaw-droppingly enormous (full-bore you can send an HD movie in one second) and I can only imagine how much they charge.

The market for this is 10x the end-user market. 4G is chump change.


A satellite constellation could, in theory, provide a link between San Francisco and New York without having to provide coverage for any of the physical area in between, and could charge the same rate ISPs currently charge for the low latency pipes that already exist between those two places. (In theory, you could make it even better than direct fiber since the speed of light is 40% slower in fiber optics compared to vacuum, or air).


. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/13/2015 06:35 pm

I was talking about tower-level end stations.  Each tower is handling 100-1000 phones at a time?
 In remote areas, still use towers, but they are stand alone, requiring only power.

I think for a remote tower, the cost of running a landline to it is more than the cost of the tower, and microwave link chains are not cheap either. Here, one hop and you're on the backbone.

Yes, and that is a great idea for one-way traffic and for remote area traffic.  Not so much for two-way traffic in non-remote areas because of the latency.  Backhaul alternatives are available.
I don't understand the link to one way traffic.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 03/13/2015 06:47 pm
Sghill: You do realize that a LEO constellation offers lower latency than just about any connection outside of nearby local caches and local data centers, right? You seem to be laboring under the assumption the SpaceX constellation would have worse latency, which is actually wrong.

This is a whole different ball game than the usual "satellite Internet" in that it actually can offer LOWER long-haul latency than any existing fiber solution.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: gosnold on 03/13/2015 06:55 pm
Sghill: You do realize that a LEO constellation offers lower latency than just about any connection outside of nearby local caches and local data centers, right? You seem to be laboring under the assumption the SpaceX constellation would have worse latency, which is actually wrong.

This is a whole different ball game than the usual "satellite Internet" in that it actually can offer LOWER long-haul latency than any existing fiber solution.

Low latency is useful for high-frequency trading: some banks have built dedicated microwave links to cut the latency and make more profit, satellite links may be the next step.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/13/2015 07:11 pm

I was talking about tower-level end stations.  Each tower is handling 100-1000 phones at a time?
 In remote areas, still use towers, but they are stand alone, requiring only power.

I think for a remote tower, the cost of running a landline to it is more than the cost of the tower, and microwave link chains are not cheap either. Here, one hop and you're on the backbone.

Yes, and that is a great idea for one-way traffic and for remote area traffic.  Not so much for two-way traffic in non-remote areas because of the latency.  Backhaul alternatives are available.
I don't understand the link to one way traffic.

Sorry.  In a context where latency matters, video or large downloads would be considered one-way and voice and video conversations and gaming would be considered two-way.

OK, so which configuration would allow lower latency than a satellite backbone?

Latency is a function of physical distance and number of routers you go through (queuing delay at each one), and physical distance starts to matter only beyond a few hundred miles.

So if a cell tower aggregates the "last mile" customers and then goes directly to a satellite, it's better (latency wise) than multiple MW hops just to get to the backbone, no?

Ideally you'd go all the way to the handset, but in practical terms, it seems you get 90% of the benefit from going to the last cell tower, but you've removed several major constraints from your system.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 03/13/2015 07:43 pm
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

Straight syncing of many servers doesn't sound like it's sensitive to latency.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: butters on 03/13/2015 07:54 pm
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

Straight syncing of many servers doesn't sound like it's sensitive to latency.

Cheers, Martin

Google operates a distributed data store which must replicate writes across a minimum number of facilities before the write request is considered complete from the perspective of the client application. Replication latency is a component of write latency.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 03/13/2015 08:03 pm
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

Straight syncing of many servers doesn't sound like it's sensitive to latency.

Cheers, Martin

Google operates a distributed data store which must replicate writes across a minimum number of facilities before the write request is considered complete from the perspective of the client application. Replication latency is a component of write latency.
I'd assumed the servers that needed low latency comms were in the same datacentre, and the data then replicated via a mechanism not needing low latency.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nessus on 03/13/2015 09:40 pm
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

Straight syncing of many servers doesn't sound like it's sensitive to latency.

Cheers, Martin

Google operates a distributed data store which must replicate writes across a minimum number of facilities before the write request is considered complete from the perspective of the client application. Replication latency is a component of write latency.

The great thing about write latency is that with a bit of attention to detail, how long a write take does not matter from the user's perspective. Especially if your application can handle serving stale data. The latency of completing a write can be hidden from the user for most use cases.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 03/15/2015 07:41 pm
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

But that's irrelevant.  "figuring out how" to updating servers and the pipe you use to connect them are two different things.  The company you mentioned already knows how to update their connected servers.

It may simply be cheaper. Plus Google has a much bigger financial interest in getting those billions of people who don't currently have access to the Internet online and using their search engine. That requires SpaceX to make a success of this constellation and throwing some business their way makes it more likely they will succeed.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: dorkmo on 03/17/2015 01:26 am
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/17/2015 03:30 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Zed_Noir on 03/17/2015 07:13 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.

The customer is always right. If they want LEO comsats. So be it.  ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/17/2015 07:36 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.

The customer is always right. If they want LEO comsats. So be it.  ;)

I meant the whole thing.

"Yeah, we're soon going to sign a large contract with someone, and GEO satellites are better... but LEO satellites are good too... and we just wanted to say that, because they're big... like Google big, or Facebook even... "

Just odd.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RonM on 03/17/2015 07:48 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.

The customer is always right. If they want LEO comsats. So be it.  ;)

I meant the whole thing.

"Yeah, we're soon going to sign a large contract with someone, and GEO satellites are better... but LEO satellites are good too... and we just wanted to say that, because they're big... like Google big, or Facebook even... "

Just odd.

GEO satellites are better unless you need low latency for two way communications. To minimize the delay you need LEO satellites, but then you have all the issues with constellations, switching, tracking, etc. It depends on what the customer wants.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 03/17/2015 08:38 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.

The customer is always right. If they want LEO comsats. So be it.  ;)

I meant the whole thing.

"Yeah, we're soon going to sign a large contract with someone, and GEO satellites are better... but LEO satellites are good too... and we just wanted to say that, because they're big... like Google big, or Facebook even... "

Just odd.

GEO satellites are better unless you need low latency for two way communications. To minimize the delay you need LEO satellites, but then you have all the issues with constellations, switching, tracking, etc. It depends on what the customer wants.
Yeah, we all know the trade offs...  I wasn't arguing that.  It's the reiteration  of  the basics, plus opinion, in the middle of a statement about a deal they can't tell us about that's really offbeat.

It's as if they're being defensive about GEO, but hey if the unknown customer wants it, etc.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 03/18/2015 12:37 am
. . . Let's see, are there any companies with lots of servers all over the world that currently spends hundreds of millions of dollars a year trying to figure out how to better keep the data between their many servers up to date? Say, maybe a company that recently invested a large amount of money in SpaceX?

Straight syncing of many servers doesn't sound like it's sensitive to latency.

Cheers, Martin

Going directly from a base station to air is common practice in the 3rd world.    Why?  Copper telecommunication wire is a common high value theft item.   Even back as far as the late 90s some of our customers would install base stations in electric barbwire enclosures with generators, batteries and solar.  The base stations communicated back to the central office switching centers with direct microwave links.    If these systems were done the way we do base stations in the first world within days they would be stripped and non-functional.   

So if you want ideas of dual air hop latencies look at some 3rd world high crime cellular coverage areas. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RonM on 03/18/2015 01:27 pm
"Boeing eyes satellite deal with tech giant this year"

http://www.reuters.com/article/2015/03/17/us-boeing-satellites-idUSKBN0MD00T20150317

"[Mark Spiwak, president of Boeing Satellite Systems International] said GEO satellites were larger and more expensive, but they would save the cost of more frequent maintenance and launch costs for a large network of low-earth orbit satellites. At the same time, he said Boeing was also prepared to build LEO satellites, depending on customer preference."
That is a very odd statement.

The customer is always right. If they want LEO comsats. So be it.  ;)

I meant the whole thing.

"Yeah, we're soon going to sign a large contract with someone, and GEO satellites are better... but LEO satellites are good too... and we just wanted to say that, because they're big... like Google big, or Facebook even... "

Just odd.

GEO satellites are better unless you need low latency for two way communications. To minimize the delay you need LEO satellites, but then you have all the issues with constellations, switching, tracking, etc. It depends on what the customer wants.
Yeah, we all know the trade offs...  I wasn't arguing that.  It's the reiteration  of  the basics, plus opinion, in the middle of a statement about a deal they can't tell us about that's really offbeat.

It's as if they're being defensive about GEO, but hey if the unknown customer wants it, etc.

I think you're reading too much into that statement. Note that it isn't even in quotes. The reporter may have incorrectly paraphrased what Mr. Spiwak said or it could even be from a second hand source. Now if that statement was in a press release it would be odd.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: The Amazing Catstronaut on 06/02/2015 04:15 pm
Reddit post where I located the article's link. ->
http://www.reddit.com/r/spacex/comments/387lfy/new_details_of_spacex_satellite_two_kuband/ (http://www.reddit.com/r/spacex/comments/387lfy/new_details_of_spacex_satellite_two_kuband/)

-

Direct application details:

 FCC application ->
https://apps.fcc.gov/oetcf/els/reports/442_Print.cfm?mode=current&application_seq=66082&license_seq=66693 (https://apps.fcc.gov/oetcf/els/reports/442_Print.cfm?mode=current&application_seq=66082&license_seq=66693)

Another document pertaining to the FCC application (summarises the mission informatively) ->
https://apps.fcc.gov/els/GetAtt.html?id=162991&x=. (https://apps.fcc.gov/els/GetAtt.html?id=162991&x=.)


This is some exciting stuff - it appears that SpaceX is going to launch two Ku band satellites on a demonstrator mission for their in-house constellation next year. Orbital inclination means that they're going up out of Vandy. All tracking stations used are SpaceX and Tesla facilities. The orbits used will be significantly lower than the orbits SpaceX will use for the functional constellation. The satellites have low enough mass to be a secondary payload, if that's what SpaceX decides to go for. Each satellite has a six-to-twelve month lifespan. This is the first of six-to-eight demonstrator missions according to SX.

All credit to Reddit user Shrubit for finding the above.

Edit: Grammar.

Edit 2x: Readability.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/02/2015 04:30 pm
Not surprised, as this is the sort of thing that is necessary in order to reach initial operations in 5 years. I think people often brush off SpaceX's long-term plans as being mere long-term and thus won't ever happen. SpaceX is not standing still. The constellation is really happening, they are taking the necessary steps to build it.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/02/2015 08:31 pm
Some things suggested by the information:

1) Freq bandwidth Ku U/D 100Mhz
2) With multi bit QAM a possible 1Gb U/D data rate (such as 4096QAM [12bits/hz][2 bits used for forward error correction coding (FEC)])
3) A 50db gain antenna (4w transmitter  to 1.1kw ERP) suggests a 20 degree beam width.
4) a 20 degree beam width means a illumination circle ~ equal to the orbit altitude (200km orbit -> 200km diameter  surface spot)
5) for a 200km orbit a connection time from one location ~25s
6) total sat power ~50-100w. (Telemetry X band downlink uses 20w)
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 06/02/2015 09:01 pm
Some things suggested by the information:

1) Freq bandwidth Ku U/D 100Mhz
2) With multi bit QAM a possible 1Gb U/D data rate (such as 4096QAM [12bits/hz][2 bits used for forward error correction coding (FEC)])
3) A 50db gain antenna (4w transmitter  to 1.1kw ERP) suggests a 20 degree beam width.
4) a 20 degree beam width means a illumination circle ~ equal to the orbit altitude (200km orbit -> 200km diameter  surface spot)
5) for a 200km orbit a connection time from one location ~25s
6) total sat power ~50-100w. (Telemetry X band downlink uses 20w)
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.
4 watts to 1.1kw would be about 25db gain.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/02/2015 09:22 pm
Some things suggested by the information:

1) Freq bandwidth Ku U/D 100Mhz
2) With multi bit QAM a possible 1Gb U/D data rate (such as 4096QAM [12bits/hz][2 bits used for forward error correction coding (FEC)])
3) A 50db gain antenna (4w transmitter  to 1.1kw ERP) suggests a 20 degree beam width.
4) a 20 degree beam width means a illumination circle ~ equal to the orbit altitude (200km orbit -> 200km diameter  surface spot)
5) for a 200km orbit a connection time from one location ~25s
6) total sat power ~50-100w. (Telemetry X band downlink uses 20w)
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.
4 watts to 1.1kw would be about 25db gain.
Yes you are correct, confusion with voltage gain (20db for factor of 10) (power gain 10db for factor of 10).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: denis on 06/02/2015 09:30 pm
Some things suggested by the information:
...
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.

What information suggests you that ?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/02/2015 10:14 pm
Some things suggested by the information:
...
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.

What information suggests you that ?

Simplicity and small size. The antenna looks to be only about 20cm in diameter with the other telemetry downlink and command up link antennas to be whisker like omnidirectional. A 50w solar array .5X.5m sq for power says that it is approximately a 3cu sized cubesat. But most likely a custom small-sat than a cubesat design.

Using gravity gradient stabilization you just need a cold gas set of thrusters to handle the initial alignment. With a possible single reaction wheel to control the solar pannel alignment (roll). This ends with a 3 axis stabilized platform whose main antenna is always point at Earth and solar array is nominally pointed at the sun within some reasonable angle.

Remember this is a quick satellite to test the communication technologies of both the space and ground segments to do a probable 1Gb data rate per spot (This antenna would be equivalent to a spot on a larger sat that would have a lot more transmitters.) 16 such spots on a larger sat would give ~ 16Gb data throughput for the sat and use ~5w per spot * 16 or 80w of power. A 1mX1m sq solar array ~150w would be big enough for such a satellite making these satellites small and lightweight.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: denis on 06/02/2015 10:38 pm
Simplicity and small size. The antenna looks to be only about 20cm in diameter with the other telemetry downlink and command up link antennas to be whisker like omnidirectional. A 50w solar array .5X.5m sq for power says that it is approximately a 3cu sized cubesat. But most likely a custom small-sat than a cubesat design.

Using gravity gradient stabilization you just need a cold gas set of thrusters to handle the initial alignment. With a possible single reaction wheel to control the solar pannel alignment (roll). This ends with a 3 axis stabilized platform whose main antenna is always point at Earth and solar array is nominally pointed at the sun within some reasonable angle.
...

Ah ok, so you are not basing this assumption on specific information, just on the simplicity.

From a simplicity point of view, I'll scrap the idea of using thrusters (even cold-gas) and go for a magnetic-based attitude control for the initial rate damping; it's cheap, extremely simple and avoid all the annoyance of having thrusters (possible mechanical failures, need for filling/pressurization before launch...).
For gravity gradient, it all depends if the other constraints (mass/volume...) allow for such solution with the required pointing accuracy. It means a satellite which has the correct shape (mass distribution) to allow for such gravity gradient stabilization.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/02/2015 10:45 pm
Since they are doing a phased array, you would want the different transmitters linked up together to give you better gain. Each spot would be using the whole transmitter array simultaneously. This reduces the power requirements and increases the possibility for frequency reuse. Also power could be reduced by using larger ground phased arrays.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/02/2015 10:53 pm
Simplicity and small size. The antenna looks to be only about 20cm in diameter with the other telemetry downlink and command up link antennas to be whisker like omnidirectional. A 50w solar array .5X.5m sq for power says that it is approximately a 3cu sized cubesat. But most likely a custom small-sat than a cubesat design.

Using gravity gradient stabilization you just need a cold gas set of thrusters to handle the initial alignment. With a possible single reaction wheel to control the solar pannel alignment (roll). This ends with a 3 axis stabilized platform whose main antenna is always point at Earth and solar array is nominally pointed at the sun within some reasonable angle.
...

Ah ok, so you are not basing this assumption on specific information, just on the simplicity.

From a simplicity point of view, I'll scrap the idea of using thrusters (even cold-gas) and go for a magnetic-based attitude control for the initial rate damping; it's cheap, extremely simple and avoid all the annoyance of having thrusters (possible mechanical failures, need for filling/pressurization before launch...).
For gravity gradient, it all depends if the other constraints (mass/volume...) allow for such solution with the required pointing accuracy. It means a satellite which has the correct shape (mass distribution) to allow for such gravity gradient stabilization.

A main body that is 20cmX20cmX20cm and a fold out 8 panel solar array, each panel 20cm X20cm (which fold out to a length straight up in relation to earth surface) give the spacecraft a total length of 1.8 tall by only .2m diameter. This shape at 200km would be about right for use of gravity gradient. With the main antenna beam width of 20 degrees there is no significant pointing requirements other than sort of keeping the solar array facing the sun. In a sun synchronous orbit it only requires roll control to maintain the solar influx angle. With 3 full arrays at 120 degree angles, you would not even need the roll control. But the arrays constitute the major weight of the satellite along with batteries.

Weight may not be the concern but simplicity and low number of failure modes.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Zardar on 06/02/2015 10:58 pm
Would they even need >50W?

They are only going to do broadband transmission tests for 10mins every 0.9 days, or ~11mins a day

According to the linked document above, "the power subsystem will allow up to one 12-minute telemetry/video contact per orbit".
At 625 km, that's an orbital period of about 97 minutes, so 12 mins is almost exactly 1/8 of that.

How big/heavy a battery would you need for peak load, and if you just used fixed solar panels and gravity stabilisation, how big an area would you need to ensure you top off the battery over the rest of the orbit?

Other than the antenna and gravity boom, this could be a really small satellite?

Some things suggested by the information:
...
6) total sat power ~50-100w. (Telemetry X band downlink uses 20w)
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.
...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/02/2015 11:22 pm
Would they even need >50W?

They are only going to do broadband transmission tests for 10mins every 0.9 days, or ~11mins a day

The power subsystem will allow up to one 12-minute telemetry/video contact per orbit, (@625km)

How big/heavy a battery would you need for peak load, and if you just used fixed solar panels and gravity stabilisation, how big an area would you need to ensure you top off the battery over the rest of the orbit?

Other than the antenna and gravity boom, this could be a really small satellite?

Some things suggested by the information:
...
6) total sat power ~50-100w. (Telemetry X band downlink uses 20w)
7) possibly gravity gradient stabilized to keep the high gain antenna always point toward the Earth. no or very little attitude control required or used.
...

Yes I did mention a possible 20cm cube size (that's an 8" cube or a half of a concrete block for visual of just how big "small" that is.) One side can be a flat phased array since it does not need to be bigger that 20cmX20cm to be able to have a 25db gain with a beam angle of 20 degrees. The ground segment antennas will be only slightly larger and probably of same design with a 32db gain. A computer the size and capability of a cell phone and batteries and a telemetry transmitter system and a command reciever system and thats it with a 50w (or as you suggest less) solar array with the four sides with arrays no fold outs, it would be able to do 12 watts (with one to two arrays in sunlight). With an 8 to 1 charge to use ratio that would work.

BTW the major power suck is the telemetry downlink at 20w which would be transmitting probably only on command. The computer and command receiver would be on continuously (only a few watts ~<5).

Man every time I revist the design it keeps getting smaller.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 06/03/2015 08:56 am
I wonder if they will be testing sat-sat comms? Laser?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/03/2015 10:11 am
Sat-Sat comms would certainly be a compelling reason to launch two of them at once.  If you're just testing space systems and space-to-ground comms then you'd think a single instance would be enough.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 06/03/2015 05:17 pm
So from what the latest news is saying I would derive the following:

Spacex will be sending up a few micro sats with the Irridium launches from Vandenberg.

These will probably include ones that specifically test the various non communications subsystems needed for the first iteration of the constellation besides iterating on the communications technologies.

I would expect the first 'bus' to be tested in the 2-3 year time frame.

Tesla will not only be providing ground station testing and communications support, Tesla will be developing and testing what will probably be incorporated in every Tesla car that will make it the first clients of the new constellation.

Tesla may legitimately partially invest in the constellation.


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: abaddon on 06/03/2015 05:30 pm
Sat-Sat comms would certainly be a compelling reason to launch two of them at once.  If you're just testing space systems and space-to-ground comms then you'd think a single instance would be enough.
You need two to test hand-off.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: go4mars on 06/03/2015 07:32 pm
Tesla may legitimately partially invest in the constellation.
And Google may become a larger partner if there's a limited scope JV investment opportunity.  Analogous to Gigafactory in that way only.  If you squint your imagination just right.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cuddihy on 06/04/2015 01:15 pm
This is a pretty interesting design -- 200 mile spot beam at 620 km and with cross-linking implies a lot of scalability. Initially you can just downlink it where you have a ground station, then build out ground stations as usage increases...say at Tesla superchargers...also at that altitude you dont need to go through the ITU until you want to go international. You can just turn off transmitters when you go outside CONUS. Eventually with cross linking you can go everywhere, and use the mix of direct downlink and cross linking appropriate to the customer / usage.

Also over time as you need more cross linking (say to go maritime), as the free optical technology continues to advance, you can crosslink with optical, and in favorable climates like Eastern CA/ Nevada downlink optically at near TB rates... Seems like a smart, scalable design.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: QuantumG on 06/05/2015 02:05 am
http://www.nbcnews.com/science/space/how-spacex-plans-test-its-satellite-internet-service-2016-n370196

NBC News reporting Reddit.. what has the world come to?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 06/05/2015 02:42 am
http://www.nbcnews.com/science/space/how-spacex-plans-test-its-satellite-internet-service-2016-n370196

NBC News reporting Reddit.. what has the world come to?

You have to careful on what you believe,  but rather often people who shouldn't talk do in the redditverse.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/05/2015 02:07 pm
http://www.nbcnews.com/science/space/how-spacex-plans-test-its-satellite-internet-service-2016-n370196

NBC News reporting Reddit.. what has the world come to?
I assume you're talking about this quote from the article:
Quote
The documents, dated May 29, came to light in a posting to Reddit's SpaceX forum this week. A spokesman for SpaceX, Phil Larson, confirmed the substance of the filings to NBC News.
I don't see what's surprising about that.  The post to reddit is what a journalist would call "a lead".  Then author of the article did real journalism by following up with SpaceX (and others) to get them to confirm the "substance of the filing" and substantiate some of the speculation (ie that the orbits listed implied launch from Vandenberg was undoubtedly fact checked).

That's exactly how things ought to work.  Read reddit or twitter or NSF L2 if you want "raw news" and can responsibly takes into account that it's got a decent chance of being false.  Wait for the journalists (NBC, NSF L1) if you'd rather read stuff with a higher likelihood of being correct, but at a slightly slower pace.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 06/05/2015 02:16 pm

That's exactly how things ought to work.  Read reddit or twitter or NSF L2 if you want "raw news" and can responsibly takes into account that it's got a decent chance of being false.  Wait for the journalists (NBC, NSF L1) if you'd rather read stuff with a higher likelihood of being correct, but at a slightly slower pace.

Sorry I don't credit mainstream news media with accurate and factual news coverage. I believe you need to be just as responsible filtering their pronouncements as you do what you read on twitter or reddit or any other new media source.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RedLineTrain on 06/05/2015 02:40 pm
Alan Boyle, the author of that article, is a good journalist.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/05/2015 02:58 pm
This is a pretty interesting design -- 200 mile spot beam at 620 km and with cross-linking implies a lot of scalability. Initially you can just downlink it where you have a ground station, then build out ground stations as usage increases...say at Tesla superchargers...also at that altitude you dont need to go through the ITU until you want to go international. You can just turn off transmitters when you go outside CONUS. Eventually with cross linking you can go everywhere, and use the mix of direct downlink and cross linking appropriate to the customer / usage.

Also over time as you need more cross linking (say to go maritime), as the free optical technology continues to advance, you can crosslink with optical, and in favorable climates like Eastern CA/ Nevada downlink optically at near TB rates... Seems like a smart, scalable design.

The beam widths are 20 degrees which means that their size is roughly equal to the orbit height, so the orbit is 625km that makes the beam 625km also. Large enough for 2 sites to communicate through the sat. If the two sats are close enough in the orbital plane then a handoff for uninterrupted communications as both pass over. In a sun sync orbit this would happen twice daily.

This is a test of not only the sat antenna tech but the ground system antenna tech. Software algorithms and info on propagation delays from site to site round trip would be tested also. A ping test getting the time for a round trip packet.

The size of the sat phased array antenna (pure speculation on this being in space) could be larger than 20cm which would be required to create a 20 degree beam. A much larger phased array could create virtual spots limited to only the diameter of the phased array. This virtual spot creation could be tested if the space antenna is the size that would be on the final sat 1st gen version.

Signal strength at ground from sat would be ~ -44db (signal power size relative to 1 watt ERP) distance signal loss  -75db, antenna gain 25db, 4 watt transmitter power 6db: -75db+25db+6db=-44db
Ground antenna gain ~30db so signal at receiver input is -14db. This should give a very strong and solid link.

Edit: Disregard above calc. Using a RF link calculator the link is -66dbm with the noise floor at -84dbm giving a SNR of ~18db.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: SnarlDaisy on 06/05/2015 06:13 pm

Spacex will be sending up a few micro sats with the Irridium launches from Vandenberg.


Please provide source for statement.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: deruch on 06/05/2015 06:27 pm

Spacex will be sending up a few micro sats with the Irridium launches from Vandenberg.


Please provide source for statement.

https://www.reddit.com/r/spacex/comments/387lfy/new_details_of_spacex_satellite_two_kuband/

That was the reddit post that outlined SpaceX's FCC radio frequency filing for the 2 microsats.  Based on the orbital parameters listed in the filing, it's been deduced that they are almost certainly launching piggybacked on one of SpaceX's Irridium launches.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 06/05/2015 06:35 pm

Spacex will be sending up a few micro sats with the Irridium launches from Vandenberg.


Please provide source for statement.

The first line of my post read "So from what the latest news is saying I would derive the following:"  so my source was the latest news and logical steps following from that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/06/2015 07:18 am
It's worth noting that Alan Boyle's article says "Vandenberg launch" but not "on an Iridium launch".
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/06/2015 01:38 pm
7 Iridium
5 other VAFB flights
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/06/2015 02:54 pm
It would be a little rude to launch test satellites for an iridium competitor on an rocket iridium paid for, don't you think?  Especially since there are 5 other options that would be less likely to tick off one of your big customers.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 06/06/2015 07:09 pm
It would be a little rude to launch test satellites for an iridium competitor on an rocket iridium paid for, don't you think?  Especially since there are 5 other options that would be less likely to tick off one of your big customers.

I don't see it as being rude if Iridium was a competitor, but in fact it is not, Iridium is aimed at mobile while this network is not.  But even if it was a competitor, what if Boeing builds 2 702 satellites for two different communications companies that compete in the same market, and those two satellites share and LV?  Or what if two competing television stations are carried by the same carrier? What if the carrier is owned by one of the two television networks?

Wouldn't also be rude for American astronauts to be carried to the ISS on LV's built by a country with which the US has economic sanctions? Wouldn't that be rude?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/06/2015 11:18 pm
@nadreck I'm sure SpaceX's salesperson knows whether or not they should ask for a piggyback slot on iridium's launch.  We'll see what's up when we get more details on SpaceX's schedule.  I think it's telling that Alan Boyle didn't get confirmation of that part of the reddit speculation, that's all, and it seems to me unsurprising.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 06/07/2015 03:55 am
@nadreck I'm sure SpaceX's salesperson knows whether or not they should ask for a piggyback slot on iridium's launch.  We'll see what's up when we get more details on SpaceX's schedule.  I think it's telling that Alan Boyle didn't get confirmation of that part of the reddit speculation, that's all, and it seems to me unsurprising.

7 out of the 14 planned launches out of Vandenberg are going to be Iridium, SpaceX is saying that they will be sending up not just the first two licensed but several more over the next 2 to 3 years. I don't expect these to go up with Jason-3 (they might but, really I would be flabbergasted if they had the hardware ready in time).  So that is now 7 out of 13, the last time frame I had for Iridium fits that 2 to 3 years, and I don't know schedules for all of them, I can't find anything recent on new dates for the Argentinian ones but they were having problems on the old news I dug up.  So of the several flights they will launch with, I will be surprised if Irridium ones are not some of the ones used, but I wouldn't be surprised at all if they weren't most or all of them.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 06/07/2015 06:48 am
Iridium is aimed at mobile while this network is not. 

It may not be aimed at mobile. But once you have decent speed internet you have phone service too. No need to have separate providers.

Edit: But I doubt Iridium would deny the request. That would be acknowledging that the SpaceX network is a dangerous competition.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Galactic Penguin SST on 06/07/2015 06:57 am
Do we know how large the satellites they are thinking of? Does using cubesats as demonstrators indicate that they might have the ability to shrink the satellite bus from today's standard size?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/07/2015 07:07 am
An interesting bit is that Iridium does its own back-haul (unlike other satellite providers) - as is SpaceX's plan.

But in Iridium, the cross links are minimal, or just large enough to support the amount of calls.  SpaceX, IIUC, are building an incredibly capable back-haul "fabric", and so will be looking for "fat" customers, not necessarily mobile.

Which actually makes me wonder - is there room for a different constellation, specializing in mobile hand-set connectivity, that instead of being bent-pipe, will use the SpaceX constellation as its back-haul?

This will leave room for the existing mobile providers to prosper, and give SpaceX a natural customer.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: deruch on 06/07/2015 08:38 am
I assumed that they were going to go on an Iridium launch due to the orbital information (especially the inclination) for the microsats listed in the FCC filing: 625 km x 625 km, 86.6°

According to Gunter's page, the initial/storage orbit for Iridium NEXT (http://space.skyrocket.de/doc_sdat/iridium-next.htm) is 667 km × 667 km, 86.4°. 

With the exception of Jason-3, all the other VAFB launches on SpaceX's manifest are to orbits inclined more than 90°.  Jason-3's is 66°.

The Iridium launch looks like the best match and was also what I thought was the most logical fit based on timing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Mark K on 06/07/2015 02:00 pm
An interesting bit is that Iridium does its own back-haul (unlike other satellite providers) - as is SpaceX's plan.

But in Iridium, the cross links are minimal, or just large enough to support the amount of calls.  SpaceX, IIUC, are building an incredibly capable back-haul "fabric", and so will be looking for "fat" customers, not necessarily mobile.

Which actually makes me wonder - is there room for a different constellation, specializing in mobile hand-set connectivity, that instead of being bent-pipe, will use the SpaceX constellation as its back-haul?

This will leave room for the existing mobile providers to prosper, and give SpaceX a natural customer.

Actually I kind of thought this might be the plan. This whole network feels more like a backbone than a direct to consumer thing, at least as initially deployed and where consumer means individual end user. I could see this network being offered to people like Iridium and other Satellite phone providers as an interconnect. I also see it for things like OnStar or Tesla interconnects, with the cellular network being the ground side last mile.

There are a lot of different visions for this type of thing. I feel like it is similar to Tesla's Powerwall in that the hype was all about individual users but the business case and real market is larger at the institutional level.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/07/2015 04:56 pm
An interesting bit is that Iridium does its own back-haul (unlike other satellite providers) - as is SpaceX's plan.

But in Iridium, the cross links are minimal, or just large enough to support the amount of calls.  SpaceX, IIUC, are building an incredibly capable back-haul "fabric", and so will be looking for "fat" customers, not necessarily mobile.

Which actually makes me wonder - is there room for a different constellation, specializing in mobile hand-set connectivity, that instead of being bent-pipe, will use the SpaceX constellation as its back-haul?

This will leave room for the existing mobile providers to prosper, and give SpaceX a natural customer.

Actually I kind of thought this might be the plan. This whole network feels more like a backbone than a direct to consumer thing, at least as initially deployed and where consumer means individual end user. I could see this network being offered to people like Iridium and other Satellite phone providers as an interconnect. I also see it for things like OnStar or Tesla interconnects, with the cellular network being the ground side last mile.

There are a lot of different visions for this type of thing. I feel like it is similar to Tesla's Powerwall in that the hype was all about individual users but the business case and real market is larger at the institutional level.

I am not sure about the order of priorities but the customers I see are:

- Last-mile aggregators (e.g. cell towers)
- Fixed locations (houses, coffee shops) - basically wireless hot spots
- Cars

And letting go of the mobile hand-held market.

Which is rather brilliant, since trying to get to hand-held sets puts a huge restriction on possible spectrum use, and there are very few usage cases of hand-helds that are not in range of either a stationary tower/hotspot or a car.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/07/2015 05:15 pm
Of course, once SpaceX builds out the satellite constellation and starts reaping profits, they have the option of getting into handsets, too, augmenting high density areas with local towers (could even be drones). Powered by solar and batteries, of course, so no need for any power lines, backup generators, or cable infrastructure.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/07/2015 06:01 pm
Of course, once SpaceX builds out the satellite constellation and starts reaping profits, they have the option of getting into handsets, too, augmenting high density areas with local towers (could even be drones). Powered by solar and batteries, of course, so no need for any power lines, backup generators, or cable infrastructure.

I don't know if they will - they need to get into business in areas where they have an inherent hands-down advantage. 

With towers, they only have an advantage if they deny other cell companies access to their backbone, and that's a war with very little up side.  Also cell towers require installation, maintenance, upgrades - my limited crystal ball says they'll stop with the constellation.  It's been a giant step already from launch provider to satellite builder to satellite operator.

Speaking of rooftops, they do have access to a lot of rooftops in sunny locations...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/07/2015 07:13 pm
Why? That's another trillion or two in revenue potential. And they don't HAVE to pursue it. Google is almost certainly interested in installing their own towers, and they're one of the partners here.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/07/2015 07:20 pm
Why? That's another trillion or two in revenue potential. And they don't HAVE to pursue it. Google is almost certainly interested in installing their own towers, and they're one of the partners here.

I know, but there are dollars everywhere.  The problem is that there's competition everywhere too.

So if you can get your own constellation, you have a huge advantage over everyone else.  If you make your own car/battery, same thing.   Your own RTLS rocket.  The only exception to the pattern is Solar City.

So while I can't rule out cell towers, I think Musk will keep SpaceX doing the things they can rule at, and thus keep high margins for lack of competition.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/07/2015 07:42 pm
There isn't actually a lot of margin in either the space launch or the automotive industries. And in the US at least, there's not a lot of healthy competition in the mobile space, either, but If SpaceX did upgrade its constellation with capability to connect to mobile, then they'd have margin to compete in that space. Ground towers would only come once they had enough of the market that they had problems with saturation in highly dense areas. But mobile is pretty low bandwidth compared to the 1 gigabit stationary connections, so that might not be a huge problem.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/07/2015 07:56 pm
There isn't actually a lot of margin in either the space launch or the automotive industries. And in the US at least, there's not a lot of healthy competition in the mobile space, either, but If SpaceX did upgrade its constellation with capability to connect to mobile, then they'd have margin to compete in that space. Ground towers would only come once they had enough of the market that they had problems with saturation in highly dense areas. But mobile is pretty low bandwidth compared to the 1 gigabit stationary connections, so that might not be a huge problem.

Right - competition is the #1 killer of margin world wide.  So the trick is to find markets where you have some unique advantage and can keep your margins high.  You can get legislative help from friends in high places, or you can pull a technology play that puts you in that position.  Marketing helps too.

So - having an RTLS rocket drops your costs to such an extent that you can both dominate the market and keep margins at a high enough level.  Owning a one-of-a-kind constellation (enabled by said rocket) does the same, but with a LOT more money.  With such a constellation, you can provide a back bone to all the towers in the world.

Building your own towers makes sense ONLY if you restrict the other players from using it.  Because if you don't, what's your advantage in comparison to them?  But if you restrict them, then you have to roll out a world wide tower chain, which is 1000x more of a headache, and actually forces the competitors to use a cable back-bone, giving your competitors more business...  It's not a smart business move.

Instead, make them work for you.  You have a unique back-bone that doesn't require the towers to have land lines.  That's unique.  It allows you to charge more money for your service.  It's a win for everyone, and allows you to focus on the constellation and space stuff.  Otherwise, it never ends.  You can go on to make cell phones, operating systems for cell phones, etc...   I think Musk will stop at the point where he doesn't have a technological ace in the hole. 

We''ll see.  This is far from a sure thing.  I do see them do the rooftop antennas.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/08/2015 12:51 am
A 4000 element satellite constellation would require a ground phased array to be able to create a beam width of <5 degrees. Half that would be better. This would give a phased array size of about 1m square at 14GHz. If this same phased array is used in space that would enable virtual spots of 5 degrees or about 54km at Earth surface. Physical pointing reuse of about 200 times on one antenna or 200 times the bit rate of a single channel 200 X 1Gb is a 200Gb sat throughput. It would be possible to see as many as 200 satellites from the one ground phased array giving something like 200Gb total throughput for a single 60km diameter area. Since the satellites can virtually form any number of or shape or sizes of beams up to the maximum of its bit rate throughput capability then concentration of service where the customers are can be done.

A 1m square ground phased array would definitely make the system not be competitive with a handheld mobile system. But the other item here is that total data rate capability from any one single ground station is from 1Gb data rate to 200Gb data rate where the ground station is communicating through all 200 seen satellites with 200 separately formed virtual beams.

Keeping both the space and ground systems to be very similar (same hardware and software) where quantities are in the 10-100's of thousands or even more would greatly reduce per unit costs.

The system has a tremendous scalability and flexibility.

A 1m phased array can be carried by a car or a plane. First such would be on a plane which would compete with other GEO sat based data services currently being used by airlines.

The other end of the equation is the number of subscribers maximum of the equivalent 20Mb consumer which is ~900 per 1Gb per downlink or 180000 per sat or 720,000,000!!!!! At only $50/m -$600/yr it would take 3.5M to break even and 4M to be profitable which is far below the max capability. A first generation system that can do only 20 spots on the sat would still meet the capability needed for the customers to be desireable and company to be profitable (capability of 72M max 20Mb subscribers). Just think a very local IP in a small town with 5,000-20,000 subscribers with no good other ability to get the needed bandwidth could use a single transceiver using a multiple sat linkage of 5 to 20 satellites out of the 200 visible. If SpaceX charged $30,000/yr for a 1Gb service then a local IP provider would have an annual Internet connection charge of up to $600,000/yr. As he adds subscribers he can contract for more bandwidth without any hardware changes. This makes for a much more economically scalable business case for the small scale IP.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 06/08/2015 01:06 am
The arrays on the satellites are likely to be much larger than 1m^2. And they may use Ka-band instead of Ku-band, allowing more cells and greater throughput per cell.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: meekGee on 06/08/2015 04:07 am
A 4000 element satellite constellation would require a ground phased array to be able to create a beam width of <5 degrees. Half that would be better. This would give a phased array size of about 1m square at 14GHz. If this same phased array is used in space that would enable virtual spots of 5 degrees or about 54km at Earth surface. Physical pointing reuse of about 200 times on one antenna or 200 times the bit rate of a single channel 200 X 1Gb is a 200Gb sat throughput. It would be possible to see as many as 200 satellites from the one ground phased array giving something like 200Gb total throughput for a single 60km diameter area. Since the satellites can virtually form any number of or shape or sizes of beams up to the maximum of its bit rate throughput capability then concentration of service where the customers are can be done.

A 1m square ground phased array would definitely make the system not be competitive with a handheld mobile system. But the other item here is that total data rate capability from any one single ground station is from 1Gb data rate to 200Gb data rate where the ground station is communicating through all 200 seen satellites with 200 separately formed virtual beams.

Keeping both the space and ground systems to be very similar (same hardware and software) where quantities are in the 10-100's of thousands or even more would greatly reduce per unit costs.

The system has a tremendous scalability and flexibility.

A 1m phased array can be carried by a car or a plane. First such would be on a plane which would compete with other GEO sat based data services currently being used by airlines.

The other end of the equation is the number of subscribers maximum of the equivalent 20Mb consumer which is ~900 per 1Gb per downlink or 180000 per sat or 720,000,000!!!!! At only $50/m -$600/yr it would take 3.5M to break even and 4M to be profitable which is far below the max capability. A first generation system that can do only 20 spots on the sat would still meet the capability needed for the customers to be desireable and company to be profitable (capability of 72M max 20Mb subscribers). Just think a very local IP in a small town with 5,000-20,000 subscribers with no good other ability to get the needed bandwidth could use a single transceiver using a multiple sat linkage of 5 to 20 satellites out of the 200 visible. If SpaceX charged $30,000/yr for a 1Gb service then a local IP provider would have an annual Internet connection charge of up to $600,000/yr. As he adds subscribers he can contract for more bandwidth without any hardware changes. This makes for a much more economically scalable business case for the small scale IP.

My eyes got crossed after a while, but I get the point...

This explains why Musk "fundamentally diverged" with Wyler who was going for hand-held spectrum.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/08/2015 04:10 am
The arrays on the satellites are likely to be much larger than 1m^2. And they may use Ka-band instead of Ku-band, allowing more cells and greater throughput per cell.

A 1m array would be plenty large enough for a 1st gen sat with 200 spot capability that uses about 75w for transmitter power to do those 200 five degree spots. -> Total system throughput 800Tb/s

To do more spots for the same area a larger array will be needed. A 2m^2 array could do possibly as many as 800 two degree spots. -> Total system throughput 2.4Pb/s
Note: Would work without any change needed on the 1st gen ground systems. Maximum capability of ground system would remain 200Gb/s because there would only still be 4000 sats at 5 degree separation.

Interesting though is that not much more power is needed for 800 two degree spots than for 200 5 degree spots with the same EIRP.

Their goal for 1st gen system is lightweight but very capable. Not large, heavy and super capable.

Note there is one item not confirmed and that is whether the system will have sat to sat links. It has been speculated but there is no evidence as yet to show that there will or will not be such links. Also the FCC filing may not have this info because non atmosphere penetration frequencies are unregulated.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 06/08/2015 05:43 am
I would think that flying 2 sats together hints at sat-to-sat comms testing, but we'll see.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: deruch on 06/08/2015 05:50 am
The arrays on the satellites are likely to be much larger than 1m^2. And they may use Ka-band instead of Ku-band, allowing more cells and greater throughput per cell.

A 1m array would be plenty large enough for a 1st gen sat with 200 spot capability that uses about 75w for transmitter power to do those 200 five degree spots. -> Total system throughput 800Tb/s

To do more spots for the same area a larger array will be needed. A 2m^2 array could do possibly as many as 800 two degree spots. -> Total system throughput 2.4Pb/s
Note: Would work without any change needed on the 1st gen ground systems. Maximum capability of ground system would remain 200Gb/s because there would only still be 4000 sats at 5 degree separation.

Interesting though is that not much more power is needed for 800 two degree spots than for 200 5 degree spots with the same EIRP.

Their goal for 1st gen system is lightweight but very capable. Not large, heavy and super capable.

Note there is one item not confirmed and that is whether the system will have sat to sat links. It has been speculated but there is no evidence as yet to show that there will or will not be such links. Also the FCC filing may not have this info because non atmosphere penetration frequencies are unregulated.

Keep in mind that the only documentation we've seen has been for the 1st testing mission and as such, it's possible that the actual satellites used in the constellation will have significant differences from those described in the filing.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 06/08/2015 02:44 pm
I would think that flying 2 sats together hints at sat-to-sat comms testing, but we'll see.
This was covered earlier in the thread.  I shared your opinion, and I was corrected.  Flying two sats could also be used for testing handoff between them, it doesn't *necessarily* have to involve direct sat-sat communication.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: starsilk on 06/08/2015 04:00 pm
Note there is one item not confirmed and that is whether the system will have sat to sat links. It has been speculated but there is no evidence as yet to show that there will or will not be such links. Also the FCC filing may not have this info because non atmosphere penetration frequencies are unregulated.

one of the major uses of the constellation was supposed to be backbone fiber replacement - much faster round trip times across the world. you can't do that without sat to sat links.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 06/08/2015 04:06 pm
one of the major uses of the constellation was supposed to be backbone fiber replacement - much faster round trip times across the world. you can't do that without sat to sat links.

Absolutely correct and very important. It does not necessarily follow though that the first two satellite constellation will already have and demonstrate this capability.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: starsilk on 06/08/2015 04:12 pm
one of the major uses of the constellation was supposed to be backbone fiber replacement - much faster round trip times across the world. you can't do that without sat to sat links.

Absolutely correct and very important. It does not necessarily follow though that the first two satellite constellation will already have and demonstrate this capability.

of course.. but if it's a big part of your infrastructure plan, they may want to test it ASAP. hence two satellites...

remains to be seen the amount of cross linking they are planning too - in plane would be easy, fixed antennas presumably pointing 'forward' and 'back'. cross-plane would greatly improve reliability without the need for ground stations, and allow dead satellites to be routed around, but would presumably need sideways pointing steerable dishes or phased arrays again.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 06/08/2015 04:18 pm
I wonder if they will go for laser comm between satellites. Pointing a laser at different angles may be easier than a dish. I don't know enough about laser communication to know if they can use a mirror. A mirror can be extremely lightweight and move fast for tracking.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Semmel on 06/08/2015 04:49 pm
Correct me if I am wrong, but phase arrays dont need physical pointing. Thats why its used for the ground to satellite communications. Why should they go for a different technology for sat--sat communication?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/08/2015 05:10 pm
Correct me if I am wrong, but phase arrays dont need physical pointing. Thats why its used for the ground to satellite communications. Why should they go for a different technology for sat--sat communication?

Ok. Lets say you have a very small sat with only 9 spots using a .2m^2 PA antenna (each spot is a 20 degree beam width in a 3X3 array of spots)[Note: this ignores the spot to spot interference and is used for illustration of concept]. Thats a 9Gb u/d link capability. Now lets say you need a 5Gb sat to sat link for long distance traffic (total sat to sat links would be 20Gb due to the 4 planner directions). Even with the 10 to 1 bit rate to frequency bandwidth modulation techniques you would need a 500Mhz frequency bandwidth for your PA sat to sat link. Lasers offer higher bandwidth than RF systems for a lot less power requirement.

The alternative is for a much larger say 2m^2 PA that can illuminate several sats separately with individual beams in plane and out of plane to give higher total throughput cross link. Suddenly our small sat has become a large and heavy sat that is a 2m^3 such that you could only launch maybe 3 at a time.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Herbie on 06/08/2015 05:49 pm
I wonder if they will go for laser comm between satellites. Pointing a laser at different angles may be easier than a dish. I don't know enough about laser communication to know if they can use a mirror. A mirror can be extremely lightweight and move fast for tracking.

I don't think the mirror would need to move particularly fast. If you calculate the average separation for a 4000 satellite constellation, assuming uniform distribution over the surface of a sphere of radius 7000 miles, you get something in the vicinity of 400 miles (=SQRT(4*PI*7000^2/4000)). Both transmitting and receiving satellites have well known current and future positions. Even though the satellites will be moving at relative speeds of several thousand MPH, there may not be a lot of steering involved due to the dispersion of the laser beam. Of course that would also depend upon the amount of data being transmitted.

Perhaps a circumferential band of outward-pointing optical fibers? Internal optical switches selecting the appropriate fiber, instead of shifting groups of mirrors?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 06/08/2015 06:15 pm
Fast steering/fine pointing mirrors are a proven reliable technology.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: docmordrid on 06/08/2015 06:26 pm
Fast steering/fine pointing mirrors are a proven reliable technology.

And let's not ignore that part of DARPA's SWEEPER program phased array optics were developed for no moving parts LIDAR etc. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 06/08/2015 07:15 pm
Fast steering/fine pointing mirrors are a proven reliable technology.

And let's not ignore that part of DARPA's SWEEPER program phased array optics were developed for no moving parts LIDAR etc.

It's getting better. I was not aware of that.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/08/2015 08:53 pm
Fast steering/fine pointing mirrors are a proven reliable technology.

And let's not ignore that part of DARPA's SWEEPER program phased array optics were developed for no moving parts LIDAR etc.

It's getting better. I was not aware of that.

May be a bit difficult getting access to that tech.

Whereas the gimbaled tracker has already been demonstrated in space from the distance of the Moon to the Earth through the atmosphere and at 500Mb/s. a mere 700 km should be able to do a much higher data rate with a singular colour. Since it will not go through the atmosphere a single tracker with multi colored lasers (each colour a separate data channel at Gb data rates) would make a very high data rate connection between two sats 200Gb/s would be possible with one tracker. Now add multiple trackers on each side (say 4) and that 200Gb/s data rate goes to 800Gb/s for in plane traffic plus 800Gb/s cross plane traffic. with 64 planes thats as much as 51Tb/s from one hemisphere to the other east/west or north /south. Sizes should be small enough to easily fit both physically and power wise on a smallsat.

Edit: Note that each tracker would connect to a different sat so 16 trackers connect to 16 close by sats creating a multiple path web such that a single or even several tracker failures would only degrade the system gracefully.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: BobHk on 06/08/2015 09:15 pm
Fast steering/fine pointing mirrors are a proven reliable technology.

And let's not ignore that part of DARPA's SWEEPER program phased array optics were developed for no moving parts LIDAR etc.

It's getting better. I was not aware of that.

May be a bit difficult getting access to that tech.

Whereas the gimbaled tracker has already been demonstrated in space from the distance of the Moon to the Earth through the atmosphere and at 500Mb/s. a mere 700 km should be able to do a much higher data rate with a singular colour. Since it will not go through the atmosphere a single tracker with multi colored lasers (each colour a separate data channel at Gb data rates) would make a very high data rate connection between two sats 200Gb/s would be possible with one tracker. Now add multiple trackers on each side (say 4) and that 200Gb/s data rate goes to 800Gb/s for in plane traffic plus 800Gb/s cross plane traffic. with 64 planes thats as much as 51Tb/s from one hemisphere to the other east/west or north /south. Sizes should be small enough to easily fit both physically and power wise on a smallsat.

Edit: Note that each tracker would connect to a different sat so 16 trackers connect to 16 close by sats creating a multiple path web such that a single or even several tracker failures would only degrade the system gracefully.

This was 2 years ago.
http://www.nature.com/nature/journal/v493/n7431/abs/nature11727.html (http://www.nature.com/nature/journal/v493/n7431/abs/nature11727.html)

DARPA funded the work.

Seems like someone was there before DARPA...
http://www.esa.int/esapub/bulletin/bullet91/b91lutz.htm (http://www.esa.int/esapub/bulletin/bullet91/b91lutz.htm)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Bargemanos on 06/11/2015 02:48 pm
Just came across this on twitter, don't know if this is recent news, not really following this thread.


http://www.washingtonpost.com/business/economy/spacex-founder-files-with-government-to-provide-internet-service-from-space/2015/06/09/db8d8d02-0eb7-11e5-a0dc-2b6f404ff5cf_story.html (http://www.washingtonpost.com/business/economy/spacex-founder-files-with-government-to-provide-internet-service-from-space/2015/06/09/db8d8d02-0eb7-11e5-a0dc-2b6f404ff5cf_story.html)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 06/15/2015 04:50 pm
Just came across this on twitter, don't know if this is recent news, not really following this thread.


http://www.washingtonpost.com/business/economy/spacex-founder-files-with-government-to-provide-internet-service-from-space/2015/06/09/db8d8d02-0eb7-11e5-a0dc-2b6f404ff5cf_story.html (http://www.washingtonpost.com/business/economy/spacex-founder-files-with-government-to-provide-internet-service-from-space/2015/06/09/db8d8d02-0eb7-11e5-a0dc-2b6f404ff5cf_story.html)

Earlier in the thread is the actual links to the FCC filings.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: AncientU on 07/04/2015 01:53 pm
The 3rd system??

Any input welcome

http://forum.nasaspaceflight.com/index.php?topic=37951.0

An uber large parabolic reflector would excel at localized, extremely faint signal detection; retransmission could also return signal to same location.  Maneuverability would be extremely limited.  These features seem to be the opposite of those needed for large numbers of widely distributed signals (internet users) which need return signals to be redirected to different locations.  Phased array signal processing is fast, variable direction, and can be highly multiplexed -- much more suitable to this market.

Don't think this brute force, uber large parabolic reflector approach has any applicability in the current competition.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Prober on 07/04/2015 08:35 pm
The 3rd system??

Any input welcome

http://forum.nasaspaceflight.com/index.php?topic=37951.0

An uber large parabolic reflector would excel at localized, extremely faint signal detection; retransmission could also return signal to same location.  Maneuverability would be extremely limited.  These features seem to be the opposite of those needed for large numbers of widely distributed signals (internet users) which need return signals to be redirected to different locations.  Phased array signal processing is fast, variable direction, and can be highly multiplexed -- much more suitable to this market.

Don't think this brute force, uber large parabolic reflector approach has any applicability in the current competition.

sleep on it, there's much more to it :)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/07/2015 12:15 pm
This constellation is going to be a HUGE undertaking. Sticking with providing backhaul services in the near-term is probably the overall safest and most realistic bet. For one, it doesn't completely undermine their comm sat launch customers. In fact, leasing bandwidth to them in the future to service their consumers may be more compelling than having to develop and deploy their own sat when it comes time for them to upgrade. 

But even if what can be viewed as a conflict of interest is largely irrelevant (and it probably is), the bigger challenge is the technology requirements needed to operate as a high-performance end-user ISP network. For practical reasons, these sats need to be as "dumb" as possible, which directly communicating with hundreds of millions of nodes on the ground runs counter to. At least in the near future with the first couple generations.

I see no way to offer ISP services that don't require substantial ground infrastructure without going the "Battlestar Galactica" route, and we know that's what Elon is trying to avoid.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Toastmastern on 07/07/2015 01:15 pm
Elon just said on the ISS Conference that they hope have a test satellite in space next year :D

Edit: Spelling
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 07/07/2015 01:45 pm
...I see no way to offer ISP services that don't require substantial ground infrastructure without going the "Battlestar Galactica" route, and we know that's what Elon is trying to avoid.
False, that is EXACTLY what SpaceX is doing and why SpaceX and OneWeb broke up: OneWeb thought Elon's plan was too ambitious and Elon thought OneWeb's was too small.

4000 satellites (greater than cubesat-sized, my guess is 250-1500kg) with interconnection capability. This is much more ambitious than Teledesic.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/07/2015 02:24 pm
This constellation is going to be a HUGE undertaking. Sticking with providing backhaul services in the near-term is probably the overall safest and most realistic bet. For one, it doesn't completely undermine their comm sat launch customers. In fact, leasing bandwidth to them in the future to service their consumers may be more compelling than having to develop and deploy their own sat when it comes time for them to upgrade. 

He talked about pizza box sized customer units. It is going to be an end user service plus backhaul service. He also mentioned explicitly satellite interconnectivity.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/07/2015 03:08 pm
False, that is EXACTLY what SpaceX is doing and why SpaceX and OneWeb broke up: OneWeb thought Elon's plan was too ambitious and Elon thought OneWeb's was too small.


And I'm attempting to fly to the moon by flapping my arms.  ::) They're not 'doing' anything yet except trying to get the ball rolling.

Some of the associated complexity has been touched upon earlier in the thread.

Consider this scenario:

User A wants to directly connect to user B (both using Musk On-Line, or MOL) via IP address. Obviously it would be unfortunate for the connection to not be as direct as possible. But what needs to happen for this to work?

First, the user needs to be authenticated somehow on the network. Avoiding ground stations would require the sats to carry something like authentication keys for every user. Similarly, the sats would need IP routing tables (and we can't get rid of this for compatibility reasons considering not all connections are between MOL nodes) which also need to be mapped to geographical locations. This takes up a decent amount of memory and requires quite a bit of processing overhead, especially if you're talking about hundreds of millions, or even BILLIONS of users. I just don't see this kind of capability being possible in the first, and possibly 2nd gen. After that, who knows?

However, acting as a backhaul service, they could sidestep a lot of these issues. Authentication might not be mandatory if you have responsible customers, and your protocol is kept under wraps, for example. But even if you do have it, your customer base is far smaller, and thus so is your authentication key database. Secondly, routing information can be part of the uplink packet, considering the backhaul customers would know the actual geographic destination of the data.

Any consumer internet service will rely heavily on ground stations, IMO. Besides the obvious benefits of being able to offload authentication and routing information to them, there's also the political implications. There's plenty of countries out there that would be unwilling to let its citizens use an uncensored and unmonitored network. Do you expect that functionality to also fit in a nice small sat? By using ground stations, those governments get what they want, and in turn you have their population as potential customers (not that I like it). 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: symbios on 07/07/2015 05:11 pm
I was wondering when I saw this quote from Elon Musk:

Quote
On deploying global satellite Internet:"We're still in the early stages of (this) idea. We're hopefully going to launch a satellite next year. I don't want to overstate this. You can actually have a more direct path through space. Protons move faster, 40-to-50 percent faster, in a vacuum than they do in fiber optic cables. So if you want to communicate from a server in California to one in South Africa, it's a very long path. We want to be really careful about how we deliver it and not overextend ourselves. I do think this is something that should be built."

Protons means light, light means laser... Is this a confirmation that they are going to use laser communication between satellites?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: kch on 07/07/2015 05:23 pm
I was wondering when I saw this quote from Elon Musk:

Quote
On deploying global satellite Internet:"We're still in the early stages of (this) idea. We're hopefully going to launch a satellite next year. I don't want to overstate this. You can actually have a more direct path through space. Protons move faster, 40-to-50 percent faster, in a vacuum than they do in fiber optic cables. So if you want to communicate from a server in California to one in South Africa, it's a very long path. We want to be really careful about how we deliver it and not overextend ourselves. I do think this is something that should be built."

Protons means light, light means laser... Is this a confirmation that they are going to use laser communication between satellites?

Light would be photons, not protons ... ;)
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/07/2015 05:27 pm
I was wondering when I saw this quote from Elon Musk:

Quote
On deploying global satellite Internet:"We're still in the early stages of (this) idea. We're hopefully going to launch a satellite next year. I don't want to overstate this. You can actually have a more direct path through space. Protons move faster, 40-to-50 percent faster, in a vacuum than they do in fiber optic cables. So if you want to communicate from a server in California to one in South Africa, it's a very long path. We want to be really careful about how we deliver it and not overextend ourselves. I do think this is something that should be built."

Protons means light, light means laser... Is this a confirmation that they are going to use laser communication between satellites?

Protons? You mean photons? :)

"Light" usually implies the visible spectrum (or near visible) of EM radiation, but the term 'photons' doesn't have to - it can be EM radiation at any frequency.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: symbios on 07/07/2015 05:31 pm
Thanks, did not catch that   ;D

That means that the quote would have been taken down wrong since protons would not be involved at all in transmissions in space radio/light or otherwise...

Article I took the quote from: http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html
 (http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html)

Since I now had to read up on photons I now know that it does not only refer to light.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: llanitedave on 07/07/2015 07:25 pm
Thanks, did not catch that   ;D

That means that the quote would have been taken down wrong since protons would not be involved at all in transmissions in space radio/light or otherwise...

Article I took the quote from: http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html
 (http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html)

Since I now had to read up on photons I now know that it does not only refer to light.

I'm sure it was a typo and he was referring to photons.  And since he's comparing the transmission speed in a vacuum with that in a fiber optic cable, it's a pretty safe bet that he's talking about light.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mme on 07/07/2015 09:39 pm
Thanks, did not catch that   ;D

That means that the quote would have been taken down wrong since protons would not be involved at all in transmissions in space radio/light or otherwise...

Article I took the quote from: http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html
 (http://www.bizjournals.com/boston/blog/techflash/2015/07/elon-musk-on-the-future-of-spacex-hyperloop-and.html)

Since I now had to read up on photons I now know that it does not only refer to light.

I'm sure it was a typo and he was referring to photons.  And since he's comparing the transmission speed in a vacuum with that in a fiber optic cable, it's a pretty safe bet that he's talking about light.
Just to be extra pedantic, he is comparing transmission speed of photons in a fibre optic cable to photons in a vacuum.  The former would be in the visible(-ish) spectrum.  The latter could be, but need not be.

Edit:  Typos.  There are always typos...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/08/2015 03:57 am
I don't see what Elon said as confirmation the communication between satellites will be Laser, radio propagates at lightspeed too.

Though laser seems a good option.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Prober on 07/08/2015 03:50 pm
I don't see what Elon said as confirmation the communication between satellites will be Laser, radio propagates at lightspeed too.

Though laser seems a good option.

So an IP in the UK sends to an IP in California transfers will need how many Sats?


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/08/2015 04:11 pm
So an IP in the UK sends to an IP in California transfers will need how many Sats?

No such thing would be handled by the constellation. Internet access would go to the next available access point of the provider.

Do you mean a point to point connection? At a rough guess 4 to 7 but changing as the constellation moves.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MikeAtkinson on 07/08/2015 04:49 pm
From LinkedIn

Steve Pollmann
Director Communication Systems Space Exploration Technologies

Quote
Working on large LEO satellite Constellation focusing on the high speed modem and communication system trade studies. Technical Leader for all spread spectrum radio communication system development.

Edit: also found several RF engineers working for SpaceX in the greater Seattle area.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ludus on 07/09/2015 02:31 am
So an IP in the UK sends to an IP in California transfers will need how many Sats?

No such thing would be handled by the constellation. Internet access would go to the next available access point of the provider.

Do you mean a point to point connection? At a rough guess 4 to 7 but changing as the constellation moves.

If you don't think that would be handled by the constellation, what did Musk mean by more than 50% of long distance traffic and 10% of local? Following a packet from a enduser in the UK to a destination in California you don't think the UK IP would route it (either peer or transit) through the constellation, bypassing transoceanic fiber for most of that distance?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/09/2015 06:01 am
If you don't think that would be handled by the constellation, what did Musk mean by more than 50% of long distance traffic and 10% of local? Following a packet from a enduser in the UK to a destination in California you don't think the UK IP would route it (either peer or transit) through the constellation, bypassing transoceanic fiber for most of that distance?

The path would be:

Customer of a UK provider through the satellite network as enduser connection to an access point of that provider in the UK. That's part of the 10% local enduser.

An intercontinental backbone link UK-USA, possibly directly to the West Coast, not like now a transatlantic cable then a continental connection Eastcoast-Westcoast, likely through several servers on the way. That's part of the backbone business.

Customer of a US-provider through the satellite network as enduser connection to an access point of that provider in the US.

That path would still cut out many detours and Routers on the way.
Watch that Seattle announcement. Elon Musk has laid this out quite clearly. He said something like they would not want to deal with local regulations in every country but work with local providers that take care of this.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/09/2015 10:13 am
At best, I could see something akin to DNS servers that provide the various necessary routing tables to make point-to-point possible for internet end users, but such servers would be at various ground stations (which *could* be stand-alone and not require a network connection other than the constellation), as it's unlikely something you'd be able to stuff in a small satellite.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/09/2015 05:10 pm
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265 (http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265)

Here is something to chew on.

For space to space communications a slightly more powerfull and slightly larger array could easily do a 9 Gbps link between sats. Using several designed for operation at different frequencies can then enable say easily with 8 of them in each of the 4 directions would give 64 Gbps in each direction.

So sat to sat linkages at very high data rates should not be that difficult with current state of the art IC mm wave technology phased arrays.

Edit added: BTW the size of these arrays would only need to be about 5 cm (2") so 16 of them would fit on a side of a sat that is 20cm square form factor. That's a sat to sat link of 128Gbs per side. If these arrays produce multiple beams and track multiple targets then the total data throughput per side can double or triple or ....

Edit Added: Another thought. At 5 cm in size these make perfect super high data rate continuous connection capabilities for cubesats. With these devices being manufactured in the 10's of thousands, they will be cheap compared to other comm solutions for a cubesat. Lightweight, low power (<4W), no moving parts, extremely low volume (a package about 3-5mm thick), and cheap (<$500 each) for space qualified 8Gbps data rate continuous worldwide communications. What kind of high bandwidth cubesat applications would that open up? If the link distance is designed for ~2000km this would cover down to the atmosphere and up to 2500km high LEO orbits. A larger array (10cm array on the cubesat would be able to extend the orbit altitude link capabilities to 4500km).
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: philw1776 on 07/10/2015 06:14 pm
Telecom cubesats, the "vermin of the skies" 21st century!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: TrevorMonty on 07/10/2015 08:09 pm
Telecom cubesats, the "vermin of the skies" 21st century!
Some of large satellite operators call cubesats "space debris with propulsion". Companies like Planet labs might call them "cash cows".
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/10/2015 08:15 pm
Another subject is power usage and sat weight.

1 - 7 low power ION thrusters at 100w  power usage each where at most only 3 are active at one time which weigh 6kg each minus Xenon tank (power available required is max 300w and 36kg total weight + Xenon Tank and propellant) Avg power at 1 % usage to maintain attitude 20w avg (Note one thruster is at CG to perform orbit change and maintenance (raising, and position movement, during this time the transmitters (links are shut down to min))
http://www.researchgate.net/publication/228785869_High_Performance_Low_Power_Ion_Propulsion_System_Design_Concept (http://www.researchgate.net/publication/228785869_High_Performance_Low_Power_Ion_Propulsion_System_Design_Concept)

2- A 512Gbps Ku band Up/Down link and 4 side links 60GHzs mm wave at 512Gbps  each total power max 200w,, min 16w, avg 110w. Weight ~10kg
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265 (http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265)

3 - Processors and routers (4 w per 256Gbps of ports, sat has 512*5 number of Gbps of port capacity or 2560Gbps) Max 60w Avg 30W min 6w total and 10 kg for boards and wiring

4- cooling system 20 kg and 50w

5- other hardware 50kg (includes structure batteries and command transceiver) Max 220w(include battery charging (200w)) avg 1W min(up to 550w battery discharge)

6 - solar array 300w per 1 m^2 and 1kg
https://en.wikipedia.org/wiki/Solar_panels_on_spacecraft (https://en.wikipedia.org/wiki/Solar_panels_on_spacecraft)

7- Xenon tank and 40kg propelant, total 50kg

wattage needed Max 510w Avg 211w Min 73w (2m^2 solar array producing 600w)

Weight sat total 178kg





Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/10/2015 08:28 pm
Another subject is power usage and sat weight.

1 - 7 low power ION thrusters at 100w  power usage each where at most only 3 are active at one time which weigh 6kg each minus Xenon tank (power available required is max 300w and 36kg total weight + Xenon Tank and propellant) Avg power at 1 % usage to maintain attitude 20w avg (Note one thruster is at CG to perform orbit change and maintenance (raising, and position movement, during this time the transmitters (links are shut down to min))
http://www.researchgate.net/publication/228785869_High_Performance_Low_Power_Ion_Propulsion_System_Design_Concept (http://www.researchgate.net/publication/228785869_High_Performance_Low_Power_Ion_Propulsion_System_Design_Concept)

2- A 512Gbps Ku band Up/Down link and 4 side links 60GHzs mm wave at 512Gbps  each total power max 200w,, min 16w, avg 110w. Weight ~10kg
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265 (http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=5746265&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fabs_all.jsp%3Farnumber%3D5746265)

3 - Processors and routers (4 w per 256Gbps of ports, sat has 512*5 number of Gbps of port capacity or 2560Gbps) Max 60w Avg 30W min 6w total and 10 kg for boards and wiring

4- cooling system 20 kg and 50w

5- other hardware 50kg (includes structure batteries and command transceiver) Max 220w(include battery charging (200w)) avg 1W min(up to 550w battery discharge)

6 - solar array 300w per 1 m^2 and 1kg
https://en.wikipedia.org/wiki/Solar_panels_on_spacecraft (https://en.wikipedia.org/wiki/Solar_panels_on_spacecraft)

7- Xenon tank and 40kg propelant, total 50kg

wattage needed Max 510w Avg 211w Min 73w (2m^2 solar array producing 600w)

Weight sat total 178kg

I agree with everything except the thrusters and their fuel, I would bet on Argon not Xenon. I think SpaceX will want the experience with Argon as an SEP fuel and with designing and building Argon engines for application for BLEO activities. Xenon is 250 times more expensive than Argon and it is rare enough that it would be hard to source for bulk use. If SpaceX were launching 800 of these satellites into orbit per year they would be using about 5 times the current worldwide annual production of Xenon.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/10/2015 08:29 pm
RE Xenon - Also about 2Kg of Xenon for station keeping would be more than plenty. But that would still be about 1/4 of the worldwide production of Xenon.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/10/2015 08:49 pm
Thanks for the Xenon info. The prop is also for orbit insertion and position movement. So using Argon would require more prop than Xenon but would as you say reduce costs. The budgeted 40kg of prop using Argon would be able to do a lot of orbital maneuvering to adjust positions as sats fail and the ring is adjusted for better coverage without gaps.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: MP99 on 07/11/2015 10:04 am
Correct me if I am wrong, but phase arrays dont need physical pointing. Thats why its used for the ground to satellite communications. Why should they go for a different technology for sat--sat communication?

Just necro-ing this...

Any sat can see multiple other sats at once. Phased array could allow routing traffic to multiple sats. The ones in the same plane won't really be moving relative to each other, but sats in other planes would be.

Cheers, Martin
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/11/2015 06:47 pm
Correct me if I am wrong, but phase arrays dont need physical pointing. Thats why its used for the ground to satellite communications. Why should they go for a different technology for sat--sat communication?

Just necro-ing this...

Any sat can see multiple other sats at once. Phased array could allow routing traffic to multiple sats. The ones in the same plane won't really be moving relative to each other, but sats in other planes would be.

Cheers, Martin

Eutelsat is going to test a main GEO sat using a Phased array. It will be a software defined comm sat. Able to on demmand create spots of any shape and size as well as bandwidth. This technology would make all comm sats when launched identical (manufactured by the same company) and is redefined when on orbit. Suddenly the payload end becomes a commercial off the shelf sat which is almost immediately available and only needs a LV to put it in space. So time from order of sat to delivery of sat would be shortened to the shortest time for the order of a launch ~18months.

The SpaceX constellation is another type of software defined sat used in LEO. It could be expanded to have other capabilities such as comm support to other LEO sats just by updating the software. Use of the self contained 5cm sized 60GHz arrays which only need the several 8Gbps data connections, power, mounted on a cooling/heating plate, and a low data rate controller connection (this is isolated form the data to improve security of the control of the sat configuration and actions). These arrays come with all the processing power to form the beams and to perform the source tracking as well as the modem function translating the RF signals to a digital Ethernet like data stream. This same array would fit on any small sat and use little power and since exact copies of it are manufactured in greater than 16,000 units for specific frequency, with 16 different frequencies  using 16 optimized designs each of quantity 16,000 units for a total number of similar units of 256,000 units required for the 4000 sats. I am sure some chip manufacturer would be glad to make these for a few hundred each and probably make a $100 profit on each or abot $25M in profits to manufacture them.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/11/2015 07:17 pm
Thanks for the Xenon info. The prop is also for orbit insertion and position movement. So using Argon would require more prop than Xenon but would as you say reduce costs. The budgeted 40kg of prop using Argon would be able to do a lot of orbital maneuvering to adjust positions as sats fail and the ring is adjusted for better coverage without gaps.

Ok for a craft that masses 178kg including 40kg of Argon I calculate a ΔV between 4000 and 8700m/s - for 10kg I get 1000 to 2400m/s.  To me 2400m/s would allow insertion, position change plus a margin for plane change for a spare. Note that when thinking about this I am still only considering a design lifespan of 5 years on the station keeping/attitude management side so the 225 to 500m/s (actually was thinking 440m/s at 3000 ISP and 2kg empty weight for the Argon tank) for simple station keeping. Also of note when I think about this, I see a tender that is used to position each of a planes satellites, carries spares and can collect malfunctioning or otherwise end of life satellites and 'coral' them until they can be deorbitted at the tender's end of life.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/11/2015 07:30 pm
For my general understanding. How would a satellite perform a plane change? Would it fire at 90° to its orbital trajectory or would it involve change of altitude?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/11/2015 07:49 pm
For my general understanding. How would a satellite perform a plane change? Would it fire at 90° to its orbital trajectory or would it involve change of altitude?

Depends how much of a plane change. If you imagine that there are something like 50 planes going up to about 60 degrees latitude then you might only have about 2 degrees of inclination change which would cost you less than 400m/s ΔV in LEO.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: CuddlyRocket on 07/11/2015 08:36 pm
I agree with everything except the thrusters and their fuel, I would bet on Argon not Xenon. I think SpaceX will want the experience with Argon as an SEP fuel and with designing and building Argon engines for application for BLEO activities. Xenon is 250 times more expensive than Argon and it is rare enough that it would be hard to source for bulk use. If SpaceX were launching 800 of these satellites into orbit per year they would be using about 5 times the current worldwide annual production of Xenon.

Plus Argon is readily available from the Martian atmosphere of which is comprises 2%; Elon wants these satellites for use around Mars as well as Earth. Of course that would mean launching them from Mars; interesting question if that would be cheaper than launching them into LMO directly from Earth!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/11/2015 10:06 pm
I agree with everything except the thrusters and their fuel, I would bet on Argon not Xenon. I think SpaceX will want the experience with Argon as an SEP fuel and with designing and building Argon engines for application for BLEO activities. Xenon is 250 times more expensive than Argon and it is rare enough that it would be hard to source for bulk use. If SpaceX were launching 800 of these satellites into orbit per year they would be using about 5 times the current worldwide annual production of Xenon.

Plus Argon is readily available from the Martian atmosphere of which is comprises 2%; Elon wants these satellites for use around Mars as well as Earth. Of course that would mean launching them from Mars; interesting question if that would be cheaper than launching them into LMO directly from Earth!

Why launch them from Mars, why not meet a tender full of them with a tanker launched from mars with a load of Argon. However, far more Argon than the few hundred satellites around Mars would need, would be needed by SEP craft plying back and forth between Mars and Earth. I see a propellant depot with Argon, Lox, Methane and Hydrogen in the future. Argon might be shipped from Mars to Earth.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/11/2015 11:26 pm
Back on the earth sat constellation.

The weight budget has some flexibility, if less prop need, add more batteries, or other hardware such as a larger array.

From the standpoint of volume I see these sats as being .5mX.5mX1m with the solar array folding out of the top at a fixed (not moving except at deployment) . In a somewhat sun synchronous orbit where the arrays can be nominally pointed at the sun they do not need to move. The 2 meter surface area array would be made up of 8 .5mX.5m panels. BTW on the rear of the solar array is the heat rejection panels. On the other end is the .5mX.5m phased array for communications with the ground. On the sides are the up to 16 5cm 60Ghz mm wave phased arrays spaced with a gap of about 5cm between them.

With this size 128 sats will fit in a 4m diameter by 11m tall dispenser and wegh ~32mt perfect for launch on an FHR. Sats are arrangedin the dispenser in 8 sats to a layer that is .75m thick and 4m diameter. Sats are ejected out the side. Ejection removes the restraining clamps on the solar array so the arrays immediately deploy after ejection (no commands or actuators.

A shortened dispenser with 32 sats could be used on a F9R weighing ~8mt.

In the FHR launch case either the sats themselves maneuver to their correct planes or the US performs a plane change after deploying half of the 128 sats for one plane.

Ok costs of launch: since spaceX is launching at costs the values are SpaceX costs and not the prices to an outside customer. An F9R at cost launch at $32M. An FHR at cost launch at $56M.  On an FHR thats a launch cost per sat of $437,500. On an F9R thats $1,000,000. If the sats cost to manufacture are kept as low as $500K, that $1-1.5M total cost per sat #4000 sats is $4B-6B to manufacture and place on orbit the constellation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 07/11/2015 11:33 pm
You may perhaps want to use Krypton instead of Argon. Krypton is far cheaper than Xenon but is nearly as efficient. Argon is lower efficiency than Xenon or Krypton for the same Isp.

http://aero.uc3m.es/ep2/docs/publicaciones/ahed10a.pdf
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ludus on 07/12/2015 12:39 am
If you don't think that would be handled by the constellation, what did Musk mean by more than 50% of long distance traffic and 10% of local? Following a packet from a enduser in the UK to a destination in California you don't think the UK IP would route it (either peer or transit) through the constellation, bypassing transoceanic fiber for most of that distance?

The path would be:

Customer of a UK provider through the satellite network as enduser connection to an access point of that provider in the UK. That's part of the 10% local enduser.

An intercontinental backbone link UK-USA, possibly directly to the West Coast, not like now a transatlantic cable then a continental connection Eastcoast-Westcoast, likely through several servers on the way. That's part of the backbone business.

Customer of a US-provider through the satellite network as enduser connection to an access point of that provider in the US.

That path would still cut out many detours and Routers on the way.
Watch that Seattle announcement. Elon Musk has laid this out quite clearly. He said something like they would not want to deal with local regulations in every country but work with local providers that take care of this.
So you DO think SpaceX would be providing the backbone (satellite to satellite)the ISP's on both ends would use? I've seen the Q&A and read transcripts. The local reference seems to be clear in being about direct access, (also being an ISP), even in crowded competitive markets. I think you're right in the understanding that most of the traffic long distance on SX would be transit or peering connections between other local providers. People who are focused on OneWeb style systems picture SX as primarily trying to sell services directly to customers who are now underserved and it seems clear he's not focused on this.

In the UK/LA example I suggest he means about 10% of the ISP market in both ends would be SX and more than 50% of all traffic over long distance backbone. So 9 of 10 times the packet would originate with another local ISP but more than half the time it's path would go through the SX constellation from sat to sat rather than over fiber cable (except at both ends). That means most of the time a packet would travel as you describe to another local ISP's access point, over the SX constellation and then through a similar path at the destination.

This would as he and you describe be a more direct path with many fewer nodes and also be faster through vacuum than fiber.

This seems like a very indirect and modest way of saying he expects to totally dominate the multi trillion dollar global telecommunications market.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/12/2015 01:29 am

So you DO think SpaceX would be providing the backbone (satellite to satellite)the ISP's on both ends would use? I've seen the Q&A and read transcripts. The local reference seems to be clear in being about direct access, (also being an ISP), even in crowded competitive markets. I think you're right in the understanding that most of the traffic long distance on SX would be transit or peering connections between other local providers. People who are focused on OneWeb style systems picture SX as primarily trying to sell services directly to customers who are now underserved and it seems clear he's not focused on this.

In the UK/LA example I suggest he means about 10% of the ISP market in both ends would be SX and more than 50% of all traffic over long distance backbone. So 9 of 10 times the packet would originate with another local ISP but more than half the time it's path would go through the SX constellation from sat to sat rather than over fiber cable (except at both ends). That means most of the time a packet would travel as you describe to another local ISP's access point, over the SX constellation and then through a similar path at the destination.

This would as he and you describe be a more direct path with many fewer nodes and also be faster through vacuum than fiber.

This seems like a very indirect and modest way of saying he expects to totally dominate the multi trillion dollar global telecommunications market.

This was my assumption, and the least burdensome on the constellation.

As for dominating telecom, perhaps the tier 1 network market and as a replacement in some cases for microwave backhauls. At least in the US, the entrenched major ISPs with lucrative peering agreements might not take kind to this. I'd expect them to go crying to congress and the DOJ if the network ends up gaining substantial market share. The thing is, these major ISPs and other tier 1 providers can just refuse to peer with it in attempt to quash the network. I suspect those that have the most to gain are companies with major global datacenters, and will be the initial primary customers, which would be a good starting point. With so many "cloud" services now a days, this would be nice market on its own. 
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/12/2015 01:43 am
I built a complex spreadsheet for data throughput per sat, the $.32 current transit rate for 1 month of 1Mbps of service, the life of the system on orbit of  7 years, plus the cost of the constellation in both capital equipment for the factories and the per unit costs of manufacturing and launch of the sats.

The conclusion is that on a 50% loading of system capability by transit a 128Gbps throughput per sat of a total revenue of the system ~$7B over the seven years is not enough to pay for the $10B in costs. But add in the 10% of ISP customers at $25 per month for a data cap amount of 2000GB (Bytes not bits) [Note the usual cable data cap is at 350GB for a service fee of $49] [Another note is that in order to download 2000GB of data in a month you would be transferring data on average at a continuous data rate of 6Mbps. But thr real item here is that the 8 million customers using SpaceX as an ISP would be a total of $17B in revenue for the 7 years.

On a 512Gbps throughput per satellite system [note is that there are only a small change in costs per satellite to increase this rate] the transit revenue over 7 years is $27.5B and the revenue over 7 years as a ISP with 33 million direct customers is $70B. A $97.5B in revenue for a total cost of $10B!!!!!!!!!!!

Somewhere around 32-64Gbps throughput per sat is the breakeven point with a 7 year life and a $10B cost. If a lower rate per sat is ultimately what is fielded for the 1st generation then possible higher rates for the ISP customers may make the venture have a ROI.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 07/12/2015 02:21 am
Keep in mind that your figures are likely all too low. Musk mentioned 1Gbps service (you will have bandwidth sharing, but still much more than 1Mbps average) in the announcement, and yes, I'm fairly certain they /intend/ to use a very high per-satellite throughput even on the initial constellation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/12/2015 03:08 am
Keep in mind that your figures are likely all too low. Musk mentioned 1Gbps service (you will have bandwidth sharing, but still much more than 1Mbps average) in the announcement, and yes, I'm fairly certain they /intend/ to use a very high per-satellite throughput even on the initial constellation.

Yes, a high throughput is indicated because at lower data rates use of the system as a transit service doesn't make sense. As seen in the revenue calculations below 128Gbps per sat throughput is marginal for using the system for transit traffic.

There is an interesting corollary here in that the ground transit portals data throughput is limited by the number of sats it can see and not the throughput of the sats. This upper limitation is probably around 128Gbps ( 1Gbps to each sat), being that horizon to horizon only about 220 sats are visible but far less will have good enough SNR to have a good link. This is still the case whether the per sat throughput is 8Gbps or 1024Gbps. The sat throughput only affects the density (how close each transit ground portal can be to its neighbor because of more and smaller sized spots). For a 8Gbps sat throughput system there would be only a couple of dozen sites in the whole US. For a 512Gbps system several thousand sites are possible with a total transit data rate into and out of the US of 1Pbps to anywhere else in the world.

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Llian Rhydderch on 07/13/2015 03:01 am
I built a complex spreadsheet for data throughput per sat, the $.32 current transit rate for 1 month of 1Mbps of service, the life of the system on orbit of  7 years, plus the cost of the constellation in both capital equipment for the factories and the per unit costs of manufacturing and launch of the sats.

The conclusion is that on a 50% loading of system capability by transit a 128Gbps throughput per sat of a total revenue of the system ~$7B over the seven years is not enough to pay for the $10B in costs. But add in the 10% of ISP customers at $25 per month for a data cap amount of 2000GB (Bytes not bits) [Note the usual cable data cap is at 350GB for a service fee of $49] [Another note is that in order to download 2000GB of data in a month you would be transferring data on average at a continuous data rate of 6Mbps. But thr real item here is that the 8 million customers using SpaceX as an ISP would be a total of $17B in revenue for the 7 years.

On a 512Gbps throughput per satellite system [note is that there are only a small change in costs per satellite to increase this rate] the transit revenue over 7 years is $27.5B and the revenue over 7 years as a ISP with 33 million direct customers is $70B. A $97.5B in revenue for a total cost of $10B!!!!!!!!!!!

Somewhere around 32-64Gbps throughput per sat is the breakeven point with a 7 year life and a $10B cost. If a lower rate per sat is ultimately what is fielded for the 1st generation then possible higher rates for the ISP customers may make the venture have a ROI.

Thanks for doing the analytical work that lets all of us benefit in thinking about how such a CommX constellation might make economic sense.

I have one thought to add, relative to " the $.32 current transit rate for 1 month of 1Mbps of service".  Musk has been making the argument that his constellation could conceivably have faster transit times, since they won't follow continental and ground-based  right-of-way fiber optic paths, and because the speed of light in free space is some nn percent faster (I think he said somthing on the order of 40%, but not too relevant for this question right now).

My question is:  might this premium service (faster round trip times, and initiation times one way) be worth something on the demand side of the market?

In other words, Musk isn't planning on selling exactly the same transmission speed service, so perhaps CommX might gain a 5 or 10% (or more???) price premium for the "premium" CommX backhaul and ISP services provision.  This would increase the revenue from the $0.32 figure you used above.

Could this be added to your model spreadsheet?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: cscott on 07/13/2015 04:22 am
There was a little bit of a bubble in fast connections between trading exchanges for a while (see Michael Lewis' book "Flash Boys").  I think that bubble had popped somewhat, now that traders are aware of that particular trick.  But perhaps there are new financial games to play.  Historical data on how much people are willing to spend to shave packet trip times may be misleading.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/13/2015 05:04 am
Thanks for doing the analytical work that lets all of us benefit in thinking about how such a CommX constellation might make economic sense.

I second that.

My question is:  might this premium service (faster round trip times, and initiation times one way) be worth something on the demand side of the market?

Elon Musk said he wants 50% market share on backbone service. That would be mainstream. Everything else being equal the speed advantage may make the difference but it can't be the primary selling point for internet backbone. There may be customers for a point to point service who are willing to pay a premium. All IMO of course.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 07/13/2015 05:11 am
There was a little bit of a bubble in fast connections between trading exchanges for a while (see Michael Lewis' book "Flash Boys").  I think that bubble had popped somewhat, now that traders are aware of that particular trick.  But perhaps there are new financial games to play.  Historical data on how much people are willing to spend to shave packet trip times may be misleading.
Nah, they still do that. It's an arms race. In order to compete, you now need that kind of fast connection.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/13/2015 03:56 pm
You need a better service advantage to get customers to switch to you from their current provider. Else you have to use a lower price. The other item not analyzed is the fact that in 5+ years from now the transit prices will drop but the ISP charges will climb. Expect transit prices to be as low as $.16 per month of 1Mbps service.

A note about transit prices. They are not for bidirectional. They are one direction only so if you want a bidirectional pipe its twice the amount. You can purchase an asymmetrical bidirectional data rate pipe service.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 07/13/2015 05:09 pm
For backbone service the satellite network will have one major advantage over other means of transport. It can provide point to point service between any two points anywhere in the world the customer wants. With the terrestrial routes he will likely need to go through several providers to get from A to B. This will add delays as the datastream goes through routers. It will add complexity as any failure will require determine which hop has gone down. It costs for each hop and the charges add up. Only in very exceptional cases you may be able to have a direct connection A to B.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/13/2015 10:54 pm
Here is a cleaned up spreadsheet of the business case model of the 50%/10% Transit/ISP usage.  The costs of installation and operation of the ground stations for both the Transit portals and the ISP servers is included in the model. The Green and yellow highlights show where the specific portion of the business cases close.

Basicly for the Transit business case to close the sat throughput needs to be greater than 256Gbps.

For the ISP portion to close is accomplished at almost any data rate. even at 16Gbps throughput per sat.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/15/2015 05:18 pm
In this article are hints about the size of the current data market as an satellite ISP for just aircraft connectivity where the service is $1.8M per plane per year for 5000 planes gives a total market size of about $9B per year!!!


http://spacenews.com/2014-top-fixed-satellite-service-operators-gogo-pushes-new-tech-for-aircraft-broadband/ (http://spacenews.com/2014-top-fixed-satellite-service-operators-gogo-pushes-new-tech-for-aircraft-broadband/)

We are thinking the marketfor data connectivity is much smaller than it actually is!!!
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/15/2015 07:34 pm
WOW. I would have never expected that the service would cost that much annually per-plane. It almost comes as a shock that the airlines would offer in-flight wifi at such a great expense when they're hardly willing to give out peanuts anymore.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 07/15/2015 10:30 pm


WOW. I would have never expected that the service would cost that much annually per-plane. It almost comes as a shock that the airlines would offer in-flight wifi at such a great expense when they're hardly willing to give out peanuts anymore.

Apes -------------> Humans
Peanuts ----------> Instagram


Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: acsawdey on 07/16/2015 02:51 pm
WOW. I would have never expected that the service would cost that much annually per-plane. It almost comes as a shock that the airlines would offer in-flight wifi at such a great expense when they're hardly willing to give out peanuts anymore.

Have you seen how much they charge for in-flight wifi lately? It's text messaging pricing for TCP/IP service, basically ...
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/16/2015 04:37 pm
WOW. I would have never expected that the service would cost that much annually per-plane. It almost comes as a shock that the airlines would offer in-flight wifi at such a great expense when they're hardly willing to give out peanuts anymore.

Have you seen how much they charge for in-flight wifi lately? It's text messaging pricing for TCP/IP service, basically ...

The main reason why airlines are willing to spend this amount of money is to get their high data rate continuous flight data in order to do failure predictions, schedule the maintenance and have the appropriate parts on hand when the plane lands, and adjust schedules almost continuously based on flight navigation data. All of which can save the company millions per plane over the year of its operation by having less down time, smoother schedule shuffling of passenger booked seats, etc. It is not to provide passengers with wifi, they just throw that in as an extra to offset the high data connectivity costs. And either charge for it separately or it is a hidden cost for the seat price.

In the end it shows the scope of the market currently. A side comment is that the revenues of companies like Gogo  is sustainable only for the durration until OneWeb becomes operational. Right now the high charge rate is due to the cost of renting all those GEOSAT transponders to get global coverage. With the LEO sat constellations a much lower charge (as high as 1/10 that of the costs of all those GEOSAT transponders) is likely in less than 5 years. Which would drop the per plane /year costs into the 100's of thousands instead of the million range. Soon after that on-board wifi for free may be more common on all airlines.


Then along comes CommX and the price goes down and bandwidth goes up from that of OneWeb. CommX's only viable competitor for this type of service. CommX is unlikely to take all the customers away from OneWeb but they will take a significant number. Changeover from one provider to another requires capital outlay and unless there is a big difference in the service price per plane/year the airlines will be slow in doing the changeovers.

So by the time that CommX is operational this Market  may have shrunk in size to only $900-2000M anually. BTW $900M as a revenue for OneWeb is $4.5B over 5 years which will completely pay for their network. For CommX half that at $450M over 7 years is $3.15B or almost 1/3 of the cost of the CommX network.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/17/2015 06:28 am
@acsawdey

Ah, right. I don't think I've been on a plane since late 2010 since I stopped travelling for work.

@oldAtlas_Eguy

The desire to have real-time networked connectivity makes total sense. Also, interesting analysis on what the future costs may be like. I suspect that you did/do this kind of stuff for a living, correct?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/17/2015 05:07 pm
@acsawdey

Ah, right. I don't think I've been on a plane since late 2010 since I stopped travelling for work.

@oldAtlas_Eguy

The desire to have real-time networked connectivity makes total sense. Also, interesting analysis on what the future costs may be like. I suspect that you did/do this kind of stuff for a living, correct?

I was a Space Systems project officer for the AF and one of the tasks before any contract negotiation was a cost analysis. So yes I did this for space systems as an engineer. Latter I got some exposure to accounting concepts at up to Master levels so I use those now as well in my analysis sometimes of the future predictions. Those are very useful when figuring out market growth and future possible co developments in an industry expansion as such the space industry is going through right now. You might want to read through an intro to business accounting. it will give you significant insights into what is going on.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: philw1776 on 07/18/2015 02:53 am
I'm an EE with an MBA in finance.  Your analyses are great reading.  I'm out of practice on my MBA stuff as I ran Telecom R&D which was mostly Eng with some cost analysis and of course budgeting.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Ronsmytheiii on 07/20/2015 02:06 am
On July 18th SpaceX listed a Linked in Job advertisement for "Avionics Radiation Effect Engineer" for their Seattle, Washington location:

https://www.linkedin.com/jobs2/view/73329691?trk=jserp_job_details_text

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/22/2015 06:31 pm
Noticed this today:

http://qz.com/460916/a-competitor-is-trying-to-force-spacex-to-reveal-its-secret-satellite-internet-plans/ (http://qz.com/460916/a-competitor-is-trying-to-force-spacex-to-reveal-its-secret-satellite-internet-plans/)

Quote
Intelsat has asked US telecom regulators to reveal confidential parts of SpaceX’s application to fly experimental internet satellites, citing fears of orbital gridlock.

or

Quote
SpaceX did not respond to a request for comment, but the company’s filing offered a potential ulterior motive in Intelsat’s objection: “Here it should be noted that Intelsat has invested in and entered into a ‘strategic alliance’ with OneWeb, an announced competitor to SpaceX’s NGSO system.”
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Prober on 07/23/2015 02:59 am
Looks like the FCC was asked to reject the application. 

http://www.wired.com/2015/07/rival-claims-spacex-internet-satellites-block/

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/23/2015 04:10 am
Looks like the FCC turned them down

http://www.wired.com/2015/07/rival-claims-spacex-internet-satellites-block/

Where does it say that, it says that Intelsat's lawyer disagrees with the SpaceX application. It has nothing about what the FCC has decided or even might be considering.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: mfck on 07/23/2015 07:57 am
SpaceNews.com on the same.

http://spacenews.com/intelsat-asks-fcc-to-block-spacex-experimental-satellite-launch/

Intelsat filed a FOIA and SpaceX has apparently replied by disclosing some of the Info.

https://apps.fcc.gov/oetcf/els/reports/ViewExhibitReport.cfm?id_file_num=0356-EX-PL-2015&application_seq=66082

Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Nomadd on 07/23/2015 01:15 pm
 Yeah, right. You can be sure the SpaceX system would interfere with Intelsat. They'd be out of be out of business the next day with their 90s service and insane prices.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: pipo82 on 07/23/2015 04:46 pm
Yeah, right. You can be sure the SpaceX system would interfere with Intelsat. They'd be out of be out of business the next day with their 90s service and insane prices.

Except Intelsat is in a completely different business to the SpaceX satellites. Intelsat are mainly in the transponders agreements, TV broadcasting and government communications sector..
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: D_Dom on 07/24/2015 04:04 pm
There is interference and then there is interference, radiation effects as opposed to price and service.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 07/24/2015 10:12 pm
Not only is Intelsat a GEO Commsat operator but they are also a significant investor of OneWeb.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: Robotbeat on 07/25/2015 03:29 am
Yeah, right. You can be sure the SpaceX system would interfere with Intelsat. They'd be out of be out of business the next day with their 90s service and insane prices.

Except Intelsat is in a completely different business to the SpaceX satellites. Intelsat are mainly in the transponders agreements, TV broadcasting and government communications sector..
...all of those could be addressed (and better) by a sufficiently powerful LEO constellation.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: msat on 07/31/2015 01:46 pm
Yeah, right. You can be sure the SpaceX system would interfere with Intelsat. They'd be out of be out of business the next day with their 90s service and insane prices.

Except Intelsat is in a completely different business to the SpaceX satellites. Intelsat are mainly in the transponders agreements, TV broadcasting and government communications sector..
...all of those could be addressed (and better) by a sufficiently powerful LEO constellation.

Yep. I foresee SpaceX's network eventually dominating the commercial commsat market. With their intent to basically mass produce these sats and launch them at a lower cost than anyone else would be able to, at least in the near-term, it's hard to see how anyone could compete with them. Then users who may have had to deploy their own sats in the past since there was no other option, could now lease low-latency bandwidth at a fraction of the price. Everyone wins except the historical commsat manufacturers, operators, and launch providers. What market is left after that? Specialty sensing and gov/military sats? Since SpaceX also intends to provide modular sat platforms, they may even indirectly seize a chunk of those.  It's weird to think that the sat division could be more lucrative than their launch division. A very logical play by SpaceX. If this actually happens, I suspect some toes being stepped on, and anti-trust concerns being brought up.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: watermod on 07/31/2015 02:19 pm
Is 4,000+ sats (and replacements) a large enough market for some chip fab to be interested in providing cheaper and more modern rad-hardened electronics?  Or, do we expect SpaceX to continue to use voting triple redundancy commercial chips in their sats as they have used in their rockets?
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: nadreck on 07/31/2015 05:23 pm
Is 4,000+ sats (and replacements) a large enough market for some chip fab to be interested in providing cheaper and more modern rad-hardened electronics?  Or, do we expect SpaceX to continue to use voting triple redundancy commercial chips in their sats as they have used in their rockets?

For the commercial market millions of chips are needed to justify something custom at the complexity level of a CPU, single chip cell phone etc. However, that is also because the target price is going to be fairly low. However, allow 10 or 100 times the target price of a laptop CPU or smart phone super chip, then, yeah, probably a few of the devices inside the SpaceX comm sats could be custom. If this does happen, then the whole micro/nano sat industry benefits.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: oldAtlas_Eguy on 08/07/2015 05:55 pm
I found this article interesting. If the AF wants space processors of this capability and the cost of development is only in the 10's of millions of dollars then such Rad Hard custom processors and other chips are within the capability of SpaceX to fund development in a total program that has $10B in costs.

http://www.militaryaerospace.com/articles/2013/03/NextGen-space-microprocessor.html (http://www.militaryaerospace.com/articles/2013/03/NextGen-space-microprocessor.html)

Basically a modern multicore multi GIPS high throughput processor as such needed to do encryption decryption support for data streams in totals reaching a 1000Gbits. And this was posted in 2013. I wonder what the DARPA / AF progress on this is currently?

Also there will not be just one processor or specific custom chip per satellite but probably on the order of 4 or more of each type. So quantities for each chip type are in the 16,000 range. At $1,000 per chip that's an order for each chip type production run of $16M or more. For such huge orders and dollar amounts rad hard custom chip manufactures would definitely be interested.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: guckyfan on 08/07/2015 06:15 pm
Why do you think they will use rad hard components? They don't use them on Falcon and Dragon. Rad hard means slow, expensive, low integration with low capabilities.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: ClaytonBirchenough on 08/07/2015 06:58 pm
And if they really do expect to lower launch costs, it might not make sense cost wise to launch rad hardened chips. Send up multiple, cheap, off-the-shelf chips and use a voting architecture just like they do currently with Falcon and Dragon.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: RonM on 08/07/2015 07:09 pm
Why do you think they will use rad hard components? They don't use them on Falcon and Dragon. Rad hard means slow, expensive, low integration with low capabilities.

SpaceX wouldn't use them on their rockets, but satellites that could stay on orbit for years would benefit from using rad hardened components.
Title: Re: SpaceX - now a satellite manufacturer (Starlink)
Post by: JasonAW3 on 08/07/2015 07:46 pm
Why do you think they will use rad hard components? They don't use them on Falcon and Dragon. Rad hard means slow, expensive, low integration with low capabilities.