SpaceX and DSN

[rockets4life97]

The Deep Space Network needs updating and budget cuts are making ongoing maintenance difficult [1].

SpaceX will need deep space communications for future missions. Will SpaceX build its own dishes around the world? Would they be able to funnel money to DSN through a foundation? How will (or should) SpaceX accomplish this task?


[1] Welcome to the Center of the Universe, Longreads, March 2018.

[docmordrid]

The speculations I've read centered around modding their StarLink birds for use around Mars, and (IIRC) using laser relay satellites stationed at Sun-Mars L1, L4 and L5.

[AncientU]

Or Sun-Earth L4/5.

[S.Paulissen]

The Deep Space Network needs updating and budget cuts are making ongoing maintenance difficult [1].

SpaceX will need deep space communications for future missions. Will SpaceX build its own dishes around the world? Would they be able to funnel money to DSN through a foundation? How will (or should) SpaceX accomplish this task?


[1] Welcome to the Center of the Universe, Longreads, March 2018.

It seems pretty obvious to me that they should build the infrastructure themselves and donate the built infrastructure for 1/5th the price* NASA procurement could generate the same capabilities.

*Wild guess, but likely conservative IMO.

[AncientU]

The Deep Space Network needs updating and budget cuts are making ongoing maintenance difficult [1].

SpaceX will need deep space communications for future missions. Will SpaceX build its own dishes around the world? Would they be able to funnel money to DSN through a foundation? How will (or should) SpaceX accomplish this task?


[1] Welcome to the Center of the Universe, Longreads, March 2018.

It seems pretty obvious to me that they should build the infrastructure themselves and donate the built infrastructure for 1/5th the price* NASA procurement could generate the same capabilities.

*Wild guess, but likely conservative IMO.

The SpaceX needs will be for high bandwidth communications with Mars... this can only be enabled by laser comms with significant aperture telescope collectors.  They have indicated that they will build it, but why 'donate' it -- at whatever price?  They will need it to carry increasing amounts of traffic.  If NASA or anyone else needs to transfer data, buy a subscription like any other commercial customer.

[docmordrid]

I wonder if they'd implement NASA's Delay/Disruption Tolerant Network protocol?

https://www.nasa.gov/content/dtn

[speedevil]

Starlink will have several thousand orbiting satellites sometime 'soon'. (before optimistic departure dates for Mars.)

These contain, according to the original application, and perhaps confirmed by hardware shown in the PAZ launch several 6" telescopes to do LASER comms.

A thousand may also be in a position to do a residual radio beam.

Naive calculations of sqrt(1000) lead to at least the possibility to get more or less nominal radio service out to near GEO with no additional hardware. Much beyond this is very questionable - though depending on design, and dedicating a few satellites, the moon might be doable. (here I'm assuming that the satellites can cooperatively beamform - this is at least a stretch).

LASER may work better.
The 6" telescopes designed for inter-sat comms, at ranges of perhaps a thousand kilometers have such a high bitrate that making reasonable assumptions, it is plausible at least you may get a kilobit signal clear to Mars, to a similar receiver.

However, scaling helps lots.
Launch 20 starlink bus satellites, with a 50cm class scope, not 15cm, make limited changes other than software and put some similar receiver on the other end, and you've gotten several megabits/s to Mars.

There is no particular reason to stick to 50cm other than you can likely put it on the Starlink satellite with minimal modifications as it is of a comparable mass to the original payload. (HIRISE was 68kg).

Going into the future, if you are planning on landing BFS on Mars, and take some satellites along with you, you can have several of these long-range mirror sats, all capable of talking directly to earth at a few megabits/s. (If you drop the satellites after a first aerobraking pass into near LMO, they need only circularise, and you may sacrifice no landing mass)

These may also be able to do double duty as reconnisance satellites, and can let normal Starlink access points work on Mars.

And - well - if you go a little up from 50cm, and make a more custom satellite (co-orbiting with a Starlink sat), you can get gigabits to Mars.

[abaddon]

It seems pretty obvious to me that they should build the infrastructure themselves and donate the built infrastructure for 1/5th the price* NASA procurement could generate the same capabilities.

*Wild guess, but likely conservative IMO.

Is this sarcasm?  I'm guessing not based on the last statement.

Assuming not, what in the world makes you think SpaceX has any idea how to build large Earth-based transmitters/receivers, let alone at a competitive price with manufacturers who specialize in that field?

The idea of space-based relays makes a little more sense, although my gut feeling is using Starlink laser comms designed for small LEO satellites at very short distances wouldn't work at all for planetary distances.  You'd want a specialized satellite that could link in with the LEO constellation but had purpose-designed and built communications linking to Mars.

That still doesn't obviate the ongoing need for a robust DSN, which is IMO often brushed over in this section...

[rockets4life97]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

[docmordrid]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

Not if your Mars-end relay sats are transmitting to similar relay sats in Earth orbit. Those can then hook up to the StarLink birds.

[Roy_H]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

That's the way I read it. And with laser communication to Mars, you don't have to worry about cloud cover. I would assume that the Mars relay satellite(s) would be in orbit somewhat above GEO and only require one or two satellites. Looking farther into the future I would expect 2 relay satellites in an orbit around the sun so communication would be uninterrupted when Earth and Mars were on opposite sides of the sun.

[philw1776]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

That's the way I read it. And with laser communication to Mars, you don't have to worry about cloud cover. I would assume that the Mars relay satellite(s) would be in orbit somewhat above GEO and only require one or two satellites. Looking farther into the future I would expect 2 relay satellites in an orbit around the sun so communication would be uninterrupted when Earth and Mars were on opposite sides of the sun.

Musk has said that very thing when talking about communicating with Mars when one of the Earth-Mars pair is near to behind the sun in relative position.  Relay satellites in solar orbit.  Obvious that he has been thinking about the issue and problems.  Why wouldn't SpaceX set up their own high bandwidth (laser) Mars comm system? Probably one of the easier tasks supporting Mars exploration given the company's expanding skill sets.

[guckyfan]

Of course this won't be a DSN. It will be a Mars specific setup. The DSN can still do its function of connecting to all the deep space probes elsewhere.

[speedevil]

Of course this won't be a DSN. It will be a Mars specific setup. The DSN can still do its function of connecting to all the deep space probes elsewhere.

Well...
Once you have an existing network of receivers around Earth, Mars, and possibly L4, a nice debugged LASER comms system that works to those receivers, deciding to put a copy of that onto your probe to Jupiter or Saturn, and accepting the bandwidth hit may be quite reasonable.
It's going to take upgrades to work to Uranus or beyond.

[ThereIWas3]

Putting the deep space relays on the Moon, linked by laser to the Earth-orbiting cloud, gives you a very stable platform on a body that rotates quite slowly compared to Earth, no pesky atmosphere, and plenty of surface to deploy solar collectors for power.  That, along with a major far-side radio telescope, are major activities that a permanent Moon base could support.

[su27k]

Previous discussion on this: https://forum.nasaspaceflight.com/index.php?topic=41240.0

It looks like one aspect need to be emphasized is DSN is not just for communication, but also for navigation.

[docmordrid]

>
It looks like one aspect need to be emphasized is DSN is not just for communication, but also for navigation.

With XNAV (X-ray pulsar navigation) being developed by NASA (NICER, SEXTANT), China (XPNAV 1) and ESA will that still be necessary in a few years?
 
Paper....(PDF)

[Asteroza]

The new TDRSS sat refresh was considering lasercomm telescopes (currently only one lasercomm per sat though) for a hybrid RF/lasercomm setup.

MIT study was suggesting next gen TDRSS be RF only, and use commercial services for optical, especially in light of lasercomm beyond earth.

[guckyfan]

Well...
Once you have an existing network of receivers around Earth, Mars, and possibly L4, a nice debugged LASER comms system that works to those receivers, deciding to put a copy of that onto your probe to Jupiter or Saturn, and accepting the bandwidth hit may be quite reasonable.
It's going to take upgrades to work to Uranus or beyond.

That assumes a new generation of deep space probes. Assuming they start developing them in a few years with new budgets they can fly in 10-20 years. But even only switching Mars to the new system should provide a lot of relief for the DSN. Much of the data stream comes from Mars.

[AncientU]

Previous discussion on this: https://forum.nasaspaceflight.com/index.php?topic=41240.0

It looks like one aspect need to be emphasized is DSN is not just for communication, but also for navigation.

With laser sats in GSO, L4 or 5, and Mars, solar system navigation will be vastly more precise than with the DSN.

[speedevil]

That assumes a new generation of deep space probes. Assuming they start developing them in a few years with new budgets they can fly in 10-20 years. But even only switching Mars to the new system should provide a lot of relief for the DSN. Much of the data stream comes from Mars.

Assuming that you can't do a deep space probe for $10M.

[guckyfan]

That assumes a new generation of deep space probes. Assuming they start developing them in a few years with new budgets they can fly in 10-20 years. But even only switching Mars to the new system should provide a lot of relief for the DSN. Much of the data stream comes from Mars.

Assuming that you can't do a deep space probe for $10M.

You lost me.

[Jim]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

That's the way I read it. And with laser communication to Mars, you don't have to worry about cloud cover. I would assume that the Mars relay satellite(s) would be in orbit somewhat above GEO and only require one or two satellites. Looking farther into the future I would expect 2 relay satellites in an orbit around the sun so communication would be uninterrupted when Earth and Mars were on opposite sides of the sun.

No, then you still have to have "DSN" to talk to the relay sats. 

Other than direct to earth stations, GSO only other viable point for a receiver.   StarlinK?  nah.  Got to stop thinking space communications is like the internet.  That burned SpaceX on vehicles.  Telemetry (vehicle health) can't be packetized, it is needs to be continuous (until we get to airliner type operations).

[biosehnsucht]

It can be continuous unpacketized (or minimally, depending on the needs of physical layer) transmission from remote device (orbiting Mars or whatever) and more heavily packetized after it arrives at a relay, and then carried over Starlink from there.

Only the first hop needs to be capable of streaming data to get every last bit of data out in order, if that's a priority.

The speed of light means that you're not going to have realtime data no matter what, it's always delayed, so after you hit the first relay (which is probably not going to burst into flames in the middle of some critical data being relayed) you can buffer and store and forward in a more efficient / internet-y method and thus use Starlink as backhaul from there.

Hell you could orbit a few satellites in useful locations (around Earth, Moon, or even the Sun, etc) which have the necessary gear to listen for broadcasts normally sent to the DSN, using a mix of relevant hardware (might require more than one kind of antenna selected as desired, or even all operating at once, plus various other bits optimized to various bandwidths/frequencies) and software defined radio techniques, sample the data (you can do so in a way that you sample a wide range and can decode later) and store and forward back to Earth via Starlink laser comms or anything else you want. This would let you basically put a "DSN dish" anywhere you want. If a $20 USB dongle can tune from ~24MHz to ~1.7GHz with over 2 million samples/sec, then high end flying SDR's for DSN use should be trivial for anyone willing to spend some money on R&D.

[Roy_H]

Let me make sure I understand the commentary above:

If you have satellites in orbit to relay the signal, there isn't a need for the large ground based antennas used by DSN?

That's the way I read it. And with laser communication to Mars, you don't have to worry about cloud cover. I would assume that the Mars relay satellite(s) would be in orbit somewhat above GEO and only require one or two satellites. Looking farther into the future I would expect 2 relay satellites in an orbit around the sun so communication would be uninterrupted when Earth and Mars were on opposite sides of the sun.

No, then you still have to have "DSN" to talk to the relay sats. 

Other than direct to earth stations, GSO only other viable point for a receiver.  StarlinK?  nah.  Got to stop thinking space communications is like the internet.  That burned SpaceX on vehicles.  Telemetry (vehicle health) can't be packetized, it is needs to be continuous (until we get to airliner type operations).

Starlink will have laser communications between satellites. I see no reason that same communications system can't have a dozen or so Starlink satellites able to connect with 2 satellites in earth orbit just above GEO. The link from these Earth to Mars relay satellites does not have to be the same Internet protocol on the Mars link side. BTW 1 satellite would do, I just think 2 for redundancy and always having at least 1 not blocked by Earth between satellite and Mars.

[CuddlyRocket]

StarlinK?  nah.  Got to stop thinking space communications is like the internet.  That burned SpaceX on vehicles.  Telemetry (vehicle health) can't be packetized, it is needs to be continuous (until we get to airliner type operations).

Just like on Earth, some communications are like the Internet and some aren't. Telemetry may need to be continuous, but not all communications are telemetry even in present-day space operations (ISS crew get email, for instance).

[ThereIWas3]

Over such long distances, you can't ask for retransmission if you get a bad checksum.  All a checksum can do is tell you that the data you just received has been garbled, but not how.

Instead you use Forward Error Correction (FEC) which adds additional information so that you can actually correct errors up to some specified limit.  The actual algorithm used depends on what kind on interference you expect.

But FEC does require dividing the data into chunks/frames/packets.

[oldAtlas_Eguy]

If you can offload from the existing DSN the predominance of scientific and image data which is not real time, is highly compressed and heavily FEC encoded, then it will be a very good thing. Not all data from probes is do or die telemetry. Most of is plain packets of data streamed through the transmitter during spacecraft inactivity. If you could be able to ship this data at any time leaving the prime telemetry transmitter for telemetry only then the demands on the DSN Earth assets become less of a problem. At least for a while. But by then the DSN should be represented by a multi-planatary/stations distributed network joined via advanced interplanetary internet system.

What is being discussed is the interim between what exists now and what would exist once settlements are established.

What is suggested is a supplemental system for the predominance of non critical data and even streams a copy of the critical data in case the primary system has drops. This supplemental system, likely a laser system, would have 100X or more the data bandwidth that of the primary telemetry. The difference is that the primary telemetry does not assume a stable craft. This is also true for the command and control uplink.

The scenario is one where the expensive DSN assets are only brought to bare when the high data rate laser link fails, for whatever reason. Most do not realize that often no DSN asset is even listening to a probe. It only when a probe is transitioning states or performing some critical action that in many cases there is more than just one DSN asset listening.

[Giovanni DS]

What is the problem with slight delays in transmission? I assume you can organize the information in packets and add time stamps from an high resolution "system time", that will allow reconstruction of real-time events downstream.

There would be no buffering, one packet is assembled while the previous one is being transmitted.

Lost information would amount to a packet transmission time, assuming the transmitter does not survive long enough.

[Robotbeat]

You can use a radio (which is what the Starlink satellites have) with different higher level protocols. No reason it HAS to be TCP/IP.

[DigitalMan]

You can use a radio (which is what the Starlink satellites have) with different higher level protocols. No reason it HAS to be TCP/IP.

There could be a lot of data that would benefit from a store and forward infrastructure.  Time sensitive data would benefit from having more specific protocols that could alleviate issues, such as multi-path.

Considering the constraints compared to terrestrial internet I am not sure one mechanism is the right path once you introduce time-sensitive data into the mix where remote humans are involved.  Just a thought.

[LouScheffer]

Previous discussion on this: https://forum.nasaspaceflight.com/index.php?topic=41240.0

It looks like one aspect need to be emphasized is DSN is not just for communication, but also for navigation.

DSN does  three main operations:
   a) Receiving data (science and telemetry)
   b) Commanding spacecraft
   c) Navigation - this has two parts - ranging and doppler shift measurement.

Function (a) you could do much more cheaply with an array of smaller antennas.  JPL itself has lots of studies on this, and astronomers use it all the time.  It's well developed technology.
 (b) is harder.  Deep space commanding needs an extremely strong signal, since you have to assume the target spacecraft could be pointed the wrong way, so you  get only the low gain antenna.   JPL uses big transmitters and big dishes combined.  Replacing this with an array is still research - the problem is, that unlike receiving, there is no phase reference.
 (c) is also harder to cheapen.  Both ranging and doppler rely on a strong uplink, as in (b).  Ranging by necessity (you need the round trip time) and doppler by technology (it's easier to have a super-stable oscillator on the ground). 

JPL now opts for the easiest, but most costly, solution.   They build all dishes at all stations capable of all operations.  Then, whenever they track a spacecraft, they do so by running the uplink and using the received data for both doppler and telemetry.  Sometimes they include ranging as well, but that takes modulation on the downlink and reduces telemetry rates.  (There is one exception which is Mars.  Here one spacecraft gets the uplink but multiple can downlink.  But this is the rare case where more than one spacecraft is within the antenna beam.)   

A new facility could do better by specializing.  Perhaps two big antennas at each site, for uplink and ranging.  Neither of these is super time consuming, so two rather than the 4-5 current antennas per site might suffice.  Then supplement this with a larger array of smaller antenna, for receiving only.   For the same expense, this could have 4-5 times more receiving area than the current solutions, with a corresponding larger science return.  The one thing you lose with this arrangement is doppler navigation data obtained simultaneously with downlink.  On some missions, this is a big deal, since this is the data needed for gravity modelling and other radio science.  But these missions could (very soon if not now) carry miniaturized, space rated, high accuracy atomic clocks.  These missions could then do doppler full time without requiring a constant uplink.

Overall, I don't think you could build a good DSN for much less than the current cost.   But you could make new stations several times more effective in terms of science return for the same money, if you are willing to change the use model

[docmordrid]

You can use a radio (which is what the Starlink satellites have) with different higher level protocols. No reason it HAS to be TCP/IP.

There could be a lot of data that would benefit from a store and forward infrastructure.  Time sensitive data would benefit from having more specific protocols that could alleviate issues, such as multi-path.
>

Like NASA's DTN (Delay/Disruption Tolerant Network), which they developed for a solar system internet?

https://www.nasa.gov/content/dtn

[Kansan52]

I think it is optimistic to plan for ST's Enterprise being in the loop (upper right hand corner).

[DigitalMan]

That's interesting, looks like a good place to start.