Starship Artemis Contract & Lunar Starship

[OTV Booster]

Would a large communication satellite at L1 be sufficient to transmit signals to and from the moon?  At least the side always facing Earth.

If the antennas are big enough and the gain high enough, yes. But that's true of any relay in any location. L1, or better yet, a halo orbit around L1 is a handy location.


The Chinese put a relay sat in an L2 halo that was big enough that it cleared the moon and was always visible from earth. The used to relay to their far side lander.

[OTV Booster]

I think I figured out how to use starlink RF to communicate with a moon bound ship.


Just as there is a sun synchronous orbit, out of the many starlink orbital planes there will most likely, at any given moment, be one plane that is 'Artemis synchronus'. From Artemis' pov that orbital plane will be concentric or near concentric to earth.


One sat, to be dedicated to Artemis, would be rotated on an axis passing through the ISL laser such that the laser is still aligned with its orbital plane to perform ISL duties and the sats belly, with phased array antenna(s), is pointing somewhere near Artemis. Gang all the phased array into one beam and start moving data.


Dedicate a second sat if bandwidth demands. Shift to new sats as necessary.

[TheRadicalModerate]

Dragon has demoed the ISLs in space, and they are more like 100 Gbps over 3000 km. Laser power will drop off with square of distance, so that should be 10 Mbps or so at lunar distances. They would just need a discardable cover, or a hatch that can stay open during burns. And validate that they can operate while the engines are firing.

But ISL from lunar distance will still have a much tougher pointing problem.  And isn't that a mechanical system?

Why is pointing tougher? Angular rates should be similar or better compared to pointing at another satellite in LEO, and the beam angles aren't smaller.

There's a difference between pointing (finding the receiver) and tracking (keeping the receiver centered).  Tracking you can do with feedback.  Pointing requires that you find the target in the first place--and then the target finds you in return.

I'm assuming you'd use ISL, where the beam angles are much narrower.  If you use RF, then you're going to have a pretty hefty SNR reduction.

Maybe the better question to ask is how large the beam spot needs to be at the receiver, in order to acquire the beam and signal the sender that they're good to go.  If they're really using a dish instead of beam steering (I'm skeptical about this), you can regain some SNR by reducing the beam angle, so the beam spot is roughly similar to what it is between Earth and LEO.

[TheRadicalModerate]

Just as there is a sun synchronous orbit, out of the many starlink orbital planes there will most likely, at any given moment, be one plane that is 'Artemis synchronus'. From Artemis' pov that orbital plane will be concentric or near concentric to earth.

Leaving aside the fact that there's no program to put Starlinks around the Moon, if they're there, then they can use ISL to relay data to a bird that has line-of-sight to Earth, and problem solved.

Of course, we shouldn't leave that fact aside.  Note that it's pretty hard to build a stable constellation in low lunar orbit.  That stability problem is likely going to make the architecture look very different from the one in LEO.

[OTV Booster]

Just as there is a sun synchronous orbit, out of the many starlink orbital planes there will most likely, at any given moment, be one plane that is 'Artemis synchronus'. From Artemis' pov that orbital plane will be concentric or near concentric to earth.

Leaving aside the fact that there's no program to put Starlinks around the Moon, if they're there, then they can use ISL to relay data to a bird that has line-of-sight to Earth, and problem solved.

Of course, we shouldn't leave that fact aside.  Note that it's pretty hard to build a stable constellation in low lunar orbit.  That stability problem is likely going to make the architecture look very different from the one in LEO.

Um, ah. Where did lunar starlinks come into the discussion? Everything I've been talking about is earth starlinks at one end and moonbound HLS at the other.

[DanClemmensen]

Just as there is a sun synchronous orbit, out of the many starlink orbital planes there will most likely, at any given moment, be one plane that is 'Artemis synchronus'. From Artemis' pov that orbital plane will be concentric or near concentric to earth.

Leaving aside the fact that there's no program to put Starlinks around the Moon, if they're there, then they can use ISL to relay data to a bird that has line-of-sight to Earth, and problem solved.

Of course, we shouldn't leave that fact aside.  Note that it's pretty hard to build a stable constellation in low lunar orbit.  That stability problem is likely going to make the architecture look very different from the one in LEO.

Um, ah. Where did lunar starlinks come into the discussion? Everything I've been talking about is earth starlinks at one end and moonbound HLS at the other.

I mentioned "lunalink" earlier without expanding on it. It's a constellation that allows simple high-bandwidth RF comms that covers the entire lunar surface, just like Starlink does for Earth. It may in fact need much more complicated orbits since LLO is unstable, but orbital details are not important for the architecture. This constellation then feeds all Moon-Earth comms into a single (redundant) Moon-Earth satellite that uses a laser link to a (redundant) satellite in LEO that in turn connects to Starlink via ISLs. This architecture means that the fiddly details of long-haul laser ISL are concentrated in this single (redundant) link. It can use a larger telescope with higher-precision pointing so all the rest of the users on the Moon can use simpler equipment. This is also an excellent prototype for the same type of system for Mars.

This scheme will not be available for the first several Artemis missions, so those spacecraft and landers will continue to need to communicate directly to Earth, But their total bandwidth can be handled in RF.

[Roy_H]

Just as there is a sun synchronous orbit, out of the many starlink orbital planes there will most likely, at any given moment, be one plane that is 'Artemis synchronus'. From Artemis' pov that orbital plane will be concentric or near concentric to earth.

Leaving aside the fact that there's no program to put Starlinks around the Moon, if they're there, then they can use ISL to relay data to a bird that has line-of-sight to Earth, and problem solved.

Of course, we shouldn't leave that fact aside.  Note that it's pretty hard to build a stable constellation in low lunar orbit.  That stability problem is likely going to make the architecture look very different from the one in LEO.

Um, ah. Where did lunar starlinks come into the discussion? Everything I've been talking about is earth starlinks at one end and moonbound HLS at the other.

I think it depends on how long the Artemis missions go on. A permanent moon base is the goal. At some point I think it will be desirable to have a small constellation of Starlink satellites. They are in high production and relatively cheap. I think a high lunar orbit would be desirable as a small number of satellites could provide continuous coverage, maybe 6?

"Note that it's pretty hard to build a stable constellation in low lunar orbit." Why? and how low? It would be advantageous to have them in a fairly high orbit where each satellite could cover at least 20% of the moon surface.

[OTV Booster]

Dragon has demoed the ISLs in space, and they are more like 100 Gbps over 3000 km. Laser power will drop off with square of distance, so that should be 10 Mbps or so at lunar distances. They would just need a discardable cover, or a hatch that can stay open during burns. And validate that they can operate while the engines are firing.

But ISL from lunar distance will still have a much tougher pointing problem.  And isn't that a mechanical system?

Why is pointing tougher? Angular rates should be similar or better compared to pointing at another satellite in LEO, and the beam angles aren't smaller.

There's a difference between pointing (finding the receiver) and tracking (keeping the receiver centered).  Tracking you can do with feedback.  Pointing requires that you find the target in the first place--and then the target finds you in return.

I'm assuming you'd use ISL, where the beam angles are much narrower.  If you use RF, then you're going to have a pretty hefty SNR reduction.

Maybe the better question to ask is how large the beam spot needs to be at the receiver, in order to acquire the beam and signal the sender that they're good to go.  If they're really using a dish instead of beam steering (I'm skeptical about this), you can regain some SNR by reducing the beam angle, so the beam spot is roughly similar to what it is between Earth and LEO.

That render of HLS with dishes promoted my response that they would also work in visible and IR.


After batting this around it looks less practical than at first glance and the reason is poor S/N. The dish on the ship, as rendered, would have something around a 1.0-1.5deg field of view. A big piece of this would include earth itself with all the optical and IR noise it puts out.


If the focal length (which has a major influence on field of view) were longer, less of the earth would be in the FOV to drown out the laser signal. FOV is a functional equivalent to beam width, but optical terminology.


A smaller FOV/beam width would demand tighter target acquisition and tracking and (trust me on this) is 100% off the shelf for non space rated hardware.


The software is 100% commercially available for a fixed position, fixed orientation system tracking sats. Both starlink and HLS will be changing both ways so two constantly changing coordinate systems need to be mapped together. Not trivial but it's not something hard like computational fluid dynamics.

This holds true for visual, IR and RF. If HLS uses RF spread spectrum phased array I suspect it would be easier to get a good S/N than with visual/IR.


Edit: formatting and wording

[TheRadicalModerate]

That render of HLS with dishes promoted my response that they would also work in visible and IR.

About that render.  I think it was based on the attached JPEG, which I downloaded in the published resolution.  It was embedded in this blurb from NASA.  I've also attached a PNG that's blown up 2x, in case I missed something.

I can't see any dishes in this picture.  I can see that there are doors from which solar arrays have deployed, and some... amorphous stuff. 

Is there somewhere else I should be looking?

[TheRadicalModerate]

Um, ah. Where did lunar starlinks come into the discussion? Everything I've been talking about is earth starlinks at one end and moonbound HLS at the other.

Ah, I misunderstood you--maybe.  Were you talking about a Starlink in LEO that was in a sunrise-sunset orbit, for a brief period of time?  I thought you were talking about a bird in a lunar orbit that was perpendicular to the Earth-Moon line.

[OTV Booster]

Um, ah. Where did lunar starlinks come into the discussion? Everything I've been talking about is earth starlinks at one end and moonbound HLS at the other.

Ah, I misunderstood you--maybe.  Were you talking about a Starlink in LEO that was in a sunrise-sunset orbit, for a brief period of time?  I thought you were talking about a bird in a lunar orbit that was perpendicular to the Earth-Moon line.

No, the bird in lunar 'orbit' would be HLS in transit, in lunar orbit or on the surface talking to a StarLink at LEO.


Initially the LEO StarLink was conceived to be cresting over earth's horizon, as seen from HLS. This would allow the ISL to come into play as I doubt the lasers have much up or down angle. ISL seemed preferable to RF because the RF is always on the Earth facing side of the StarLinks.


This idea morphed when I realized that one of the many SL orbital planes would be concentric or near concentric to earth from HLS POV. Then, rotating the SL about the axis through the laser would keep the laser in the proper plane for its intended use and allow the RF antennas to reorient towards HLS.


I think it was Dan that posted a higher res pic showing dishes but at this point in the discussion that is superfluous.


Are we in the same wavelength now? Is it RF or visable?


A note on signal power. The SL RF system is what it is. It's doubtful it will be physically modified for Artemis. That said, there is still a lot of flexibility. The phased array is normally divided into many beams to service multiple customers. This need not be. Each antenna can be bent into one beam. Because signal coherence is not an issue multiple antennas can be ganged.


Transmit and receive capabilities do not have to be symmetrical. Normally the far object has a puny antenna working at low power and the Deep Space Network has mongo antennas working at high power and sensitivity. This can be reversed with StarLinks at LEO being puny and the distant HLS running the mongo system. IIUC, the shipboard Starlink hardware has been currently available OTS equipment. As long as the FCC has no objection this can all be put on steroids with higher power and higher gain antennas.

[StraumliBlight]

https://twitter.com/audrey_decker9/status/1989352112728510935

SpaceX’s new tentative schedule for HLS, per internal document I obtained:
- Prop transfer June 2026

Do we know why the prop transfer has been delayed?

Q: How early in the year are you going to hit orbital refueling with Starship?

Musk: Not that early in the year, I'd say towards towards the end of the year.

[Lee Jay]

https://twitter.com/audrey_decker9/status/1989352112728510935

SpaceX’s new tentative schedule for HLS, per internal document I obtained:
- Prop transfer June 2026

Do we know why the prop transfer has been delayed?

Q: How early in the year are you going to hit orbital refueling with Starship?

Musk: Not that early in the year, I'd say towards towards the end of the year.

Because they blew up another ship?  And remember, Musk estimates are always optimistic, usually by a factor of about pi.

[crandles57]

Because they blew up another ship?  And remember, Musk estimates are always optimistic, usually by a factor of about pi.

Yes they destroyed booster 18 on 21 Nov 2025 which is after that timetable came to light. 1 month 18 days later still not rolled booster 19 to Masseys for pressure and cryotest. There was some PR that this wasn't going to delay flight 12 (much?) but if we assume there is 2 months delay due to booster 18 mishap then June gets delayed to August rather than late 2026 so perhaps it doesn't fully explain the delay?

Perhaps there are other delays like TPS work on ships 39 and 40 that are also adding to delays?

Or maybe June was already optimistic in October given flight 12 suborbital in March, orbital flight in April then build 2 more ships for a refuel demo even if they don't need more boosters or test flights before refuel demo attempt.

[Reynold]

SpaceX’s new tentative schedule for HLS, per internal document I obtained:
- Prop transfer June 2026

Do we know why the prop transfer has been delayed?

Q: How early in the year are you going to hit orbital refueling with Starship?

Musk: Not that early in the year, I'd say towards towards the end of the year.

Just a note from reading this, a propellant transfer demo, which could take place "June 2026", is not the same thing as orbital refueling of a Starship "towards the end of the year".  Refueling of a Starship, presumably from some kind of depot well after the first demonstration of prop transfer, would probably be to then fly the refueled Starship somewhere else.  These timelines don't necessarily contradict.

[meekGee]

SpaceX’s new tentative schedule for HLS, per internal document I obtained:
- Prop transfer June 2026

Do we know why the prop transfer has been delayed?

Q: How early in the year are you going to hit orbital refueling with Starship?

Musk: Not that early in the year, I'd say towards towards the end of the year.

Just a note from reading this, a propellant transfer demo, which could take place "June 2026", is not the same thing as orbital refueling of a Starship "towards the end of the year".  Refueling of a Starship, presumably from some kind of depot well after the first demonstration of prop transfer, would probably be to then fly the refueled Starship somewhere else.  These timelines don't necessarily contradict.

Yes, exactly.
As pointed to above, an actual operational refueling requires a number of pieces to line up, and at this point each step has a significant probability of failing.

[xvel]

In June, at most, the first orbital flight may take place, absolutely nothing more. Orbital refueling this year is not certain. Get back down to earth, folks, where starship is.

[Vultur]

In June, at most, the first orbital flight may take place, absolutely nothing more. Orbital refueling this year is not certain.

*if* Flight 12 goes well, Flight 13 could potentially happen before June.

It's also maybe possible that the first orbital Starship will loiter and become the "target" for the first transfer demo?

I agree it's definitely not certain, but there's a lot of pressure to try as soon as possible.

[OTV Booster]

In June, at most, the first orbital flight may take place, absolutely nothing more. Orbital refueling this year is not certain.

*if* Flight 12 goes well, Flight 13 could potentially happen before June.

It's also maybe possible that the first orbital Starship will loiter and become the "target" for the first transfer demo?

I agree it's definitely not certain, but there's a lot of pressure to try as soon as possible.

I'd like to think this'll happen but question an extended loiter. IIRC, one goal of the first falcon heavy was to demonstrate a 6 hour loiter capability, which suggests that loiter is a whole world unto itself.


What's needed for loiter?  What comes to mind is keeping batteries charged and keeping enough propellant aboard for all expected attitude control and maneuvering plus a reentry burn in case transfer doesn't work. All that and everything has to still work.


Whatever the loiter limits are, the turnaround has to be faster. The new pad is untried. NOBODY knows how fast it can turn around. Could be anything from 12 hours to 12 weeks.


A second operating pad sure would be handy.

[xvel]

In June, at most, the first orbital flight may take place, absolutely nothing more. Orbital refueling this year is not certain.

*if* Flight 12 goes well, Flight 13 could potentially happen before June.

It's also maybe possible that the first orbital Starship will loiter and become the "target" for the first transfer demo?

I agree it's definitely not certain, but there's a lot of pressure to try as soon as possible.

I'd like to think this'll happen but question an extended loiter. IIRC, one goal of the first falcon heavy was to demonstrate a 6 hour loiter capability, which suggests that loiter is a whole world unto itself.


What's needed for loiter?  What comes to mind is keeping batteries charged and keeping enough propellant aboard for all expected attitude control and maneuvering plus a reentry burn in case transfer doesn't work. All that and everything has to still work.


Whatever the loiter limits are, the turnaround has to be faster. The new pad is untried. NOBODY knows how fast it can turn around. Could be anything from 12 hours to 12 weeks.


A second operating pad sure would be handy.

New pad is certainly designed for much faster turnaround than old one, there is no question about that, the question is whether it will work the first time, because it does not look like the assembly and testing are going smoothly, the same applies to vehicles.
As of now there is no evidence that there will be much flying this year nor that the flights will be very successful.