Launching Starlink with Starship

[winkhomewinkhome]

None of this Starlink deployment stuff has anything to do with the current Texas Prototypes. Please take it to the appropriate Starship Engineering thread, or one of the Starlink threads.

Apologies Herb - the question was generated from the render thread, where discussion normally doesn't happen, and most of it shows up here, so...again, sorry.

[OTV Booster]

Edit: now that I think of it, in Starship, the columns of cartridges could be stacked on a turntable to increase the amount of satellites deployed from each cartridge stack.

If you include the ability to tilt the base of each stack's mounting base out 90 degrees then you can do away with ejector springs.

Tilt the stack perpendicular to the ship's longitudinal axis and begin a slow roll of the ship. Release the stack at the proper time and rotate the turntable to the next stack, repeat the tilt and release.

Understand what you are saying, but I think there might be advantages to time to operational position if they ejected satellites already spaced out. I suspect they are not doing that now due to fairing room and debris constraints. But the cartridge ejections are already being used successfully for things like cubesats now. Starship should have the capacity to do things this way which may get things better prepared for raising to final orbits. As it is now, pretty sure they hold back on the raising rate of many satellites to provide a better trajectory to their final position. Staggered ejection should allow for more optimization of these orbit raises.

MKremer and you both have the seeds of an idea. Roll SS so the chomper opens to port or starboard (earth orientation) instead of zenith. Then flat spin parallel to earth surface and elevate the nose end of the stacks. Let each stack of SL's, or even half stacks) dispense separately and over time. Stack only needs to elevate enough to clear everything. Maybe only 30 seat of the pants degrees.

The chomper mechanism and stack elevation could be tied together, and the flat spin would add centrifugal force (yeah, I know) to help or even completely power the chomper opening.

The reason for raising orbit staggered AIUI, is that all orbits precess, but lower ones precess faster. So, raise 20 sats into their plane while holding the other 40 down low. Over time the low sats precess to the next plane and it's time to raise another 20. Repeat for the last 20. I'm still trying to Grok why things precess.

When they start raising the sats they either raise them at a different rate, stagger when they raise them, or both. Low orbit=fast. High orbit=slow. The goal is to time it so by the time all the sats in a plane reach orbit, each is 18deg from its neighbor.

I don't think there's much orbital mechanics advantage to spreading out the deploy but there might be congestion issues with that many sats.

Best of both ideas? Or is this just another half stacked idea?


Edit: ah, erbilo got there first

[TheRadicalModerate]

My thoughts on centrifugal Starlink deployment and the chomper:

We know that the floor of the payload bay is both a tilt and spin table.  Chomper door opens, table tilts, deploys the most exposed payload, spins another one into the open, deploys, etc.

If it tilts a little bit more than shown in the artwork (which in turn means that the chomper opens wider than shown in the artwork), then you could tilt a gaggle of Starlinks out into the clear, tensioned as they are currently on F9 (see my stacking options up-thread), and... start to rotate the whole tilt table?  When you then removed the tensioners, they should fly up and out without fouling the Starship, and the centrifugal force from the spinning tilt table should be enough to separate the blob.

This'll provide weird dynamics for the Starship to damp out, likely with RCS.  They're less weird if the tilt table can tilt all the way to 90º, but that in turn requires the chomper to bend back almost 180º.  I don't see a reason why that wouldn't work, though.

[inaccurate_reality]

My thoughts on centrifugal Starlink deployment and the chomper:

We know that the floor of the payload bay is both a tilt and spin table.  Chomper door opens, table tilts, deploys the most exposed payload, spins another one into the open, deploys, etc.

If it tilts a little bit more than shown in the artwork (which in turn means that the chomper opens wider than shown in the artwork), then you could tilt a gaggle of Starlinks out into the clear, tensioned as they are currently on F9 (see my stacking options up-thread), and... start to rotate the whole tilt table?  When you then removed the tensioners, they should fly up and out without fouling the Starship, and the centrifugal force from the spinning tilt table should be enough to separate the blob.

This'll provide weird dynamics for the Starship to damp out, likely with RCS.  They're less weird if the tilt table can tilt all the way to 90º, but that in turn requires the chomper to bend back almost 180º.  I don't see a reason why that wouldn't work, though.

I like the idea of the rotating payload deployment system, like a lazy suzan for starlinks. I think that, combined with the classic head-over-heels tumble of the starship could scatter a stack of about 50 sats at a time.

The then close up the chomper, alter orbital plane, and deploy the next 50. I haven't done the math, but how many orbital planes could a SS visit, deploying a stack at each plane? I'd assume there's an optimal point between how many sats are deployed and how many planes are visited. In an ideal situation, you'd get a starship to deploy a few full planes of sats, each a plane apart, and have it run out of reaction mass (save for what's spared for landing) right after deploying the last stack.

[ZChris13]

My thoughts on centrifugal Starlink deployment and the chomper:

We know that the floor of the payload bay is both a tilt and spin table.  Chomper door opens, table tilts, deploys the most exposed payload, spins another one into the open, deploys, etc.

If it tilts a little bit more than shown in the artwork (which in turn means that the chomper opens wider than shown in the artwork), then you could tilt a gaggle of Starlinks out into the clear, tensioned as they are currently on F9 (see my stacking options up-thread), and... start to rotate the whole tilt table?  When you then removed the tensioners, they should fly up and out without fouling the Starship, and the centrifugal force from the spinning tilt table should be enough to separate the blob.

This'll provide weird dynamics for the Starship to damp out, likely with RCS.  They're less weird if the tilt table can tilt all the way to 90º, but that in turn requires the chomper to bend back almost 180º.  I don't see a reason why that wouldn't work, though.

I like the idea of the rotating payload deployment system, like a lazy suzan for starlinks. I think that, combined with the classic head-over-heels tumble of the starship could scatter a stack of about 50 sats at a time.

The then close up the chomper, alter orbital plane, and deploy the next 50. I haven't done the math, but how many orbital planes could a SS visit, deploying a stack at each plane? I'd assume there's an optimal point between how many sats are deployed and how many planes are visited. In an ideal situation, you'd get a starship to deploy a few full planes of sats, each a plane apart, and have it run out of reaction mass (save for what's spared for landing) right after deploying the last stack.

Plane changes are extremely expensive in LEO, of the same order as Starship's entire fully refueled delta v budget. Better to just dump the sats, let them sort themselves out on orbit, and get back down. Let the Earth's equatorial bulge and time sort them into planes, like is being done with the launches on Falcon 9.

[TheRadicalModerate]

My thoughts on centrifugal Starlink deployment and the chomper:

We know that the floor of the payload bay is both a tilt and spin table.  Chomper door opens, table tilts, deploys the most exposed payload, spins another one into the open, deploys, etc.

If it tilts a little bit more than shown in the artwork (which in turn means that the chomper opens wider than shown in the artwork), then you could tilt a gaggle of Starlinks out into the clear, tensioned as they are currently on F9 (see my stacking options up-thread), and... start to rotate the whole tilt table?  When you then removed the tensioners, they should fly up and out without fouling the Starship, and the centrifugal force from the spinning tilt table should be enough to separate the blob.

This'll provide weird dynamics for the Starship to damp out, likely with RCS.  They're less weird if the tilt table can tilt all the way to 90º, but that in turn requires the chomper to bend back almost 180º.  I don't see a reason why that wouldn't work, though.

I like the idea of the rotating payload deployment system, like a lazy suzan for starlinks. I think that, combined with the classic head-over-heels tumble of the starship could scatter a stack of about 50 sats at a time.

The then close up the chomper, alter orbital plane, and deploy the next 50. I haven't done the math, but how many orbital planes could a SS visit, deploying a stack at each plane? I'd assume there's an optimal point between how many sats are deployed and how many planes are visited. In an ideal situation, you'd get a starship to deploy a few full planes of sats, each a plane apart, and have it run out of reaction mass (save for what's spared for landing) right after deploying the last stack.

You deploy them all at once, and then different groups hang out for longer or shorter times in low orbit, given them time to have their acending nodes to precess to the proper right ascension before boosting up to the optimal altitude.

It's hard to dribble them out if they use the tension bar scheme.

[Robotbeat]

My thoughts on centrifugal Starlink deployment and the chomper:

We know that the floor of the payload bay is both a tilt and spin table.  Chomper door opens, table tilts, deploys the most exposed payload, spins another one into the open, deploys, etc.

If it tilts a little bit more than shown in the artwork (which in turn means that the chomper opens wider than shown in the artwork), then you could tilt a gaggle of Starlinks out into the clear, tensioned as they are currently on F9 (see my stacking options up-thread), and... start to rotate the whole tilt table?  When you then removed the tensioners, they should fly up and out without fouling the Starship, and the centrifugal force from the spinning tilt table should be enough to separate the blob.

This'll provide weird dynamics for the Starship to damp out, likely with RCS.  They're less weird if the tilt table can tilt all the way to 90º, but that in turn requires the chomper to bend back almost 180º.  I don't see a reason why that wouldn't work, though.

I like the idea of the rotating payload deployment system, like a lazy suzan for starlinks. I think that, combined with the classic head-over-heels tumble of the starship could scatter a stack of about 50 sats at a time.

The then close up the chomper, alter orbital plane, and deploy the next 50. I haven't done the math, but how many orbital planes could a SS visit, deploying a stack at each plane? I'd assume there's an optimal point between how many sats are deployed and how many planes are visited. In an ideal situation, you'd get a starship to deploy a few full planes of sats, each a plane apart, and have it run out of reaction mass (save for what's spared for landing) right after deploying the last stack.

Plane changes are extremely expensive in LEO, of the same order as Starship's entire fully refueled delta v budget. Better to just dump the sats, let them sort themselves out on orbit, and get back down. Let the Earth's equatorial bulge and time sort them into planes, like is being done with the launches on Falcon 9.

Not all inclination changes are equally expensive, though.

[TheRadicalModerate]

Not all inclination changes are equally expensive, though.

Just to be clear:  RAAN changes, which is what you're interested in, cost nothing as long as you're willing to wait for a bit.  You deploy to a low orbit, which precesses faster than things would in the final 550km orbit.  As each target RAAN approaches, you boost only those sats for the target RAAN up to the target orbit.  The difference in precession speeds between the low and 550km orbits lets you do everything with no additional delta-v.

[ZChris13]

Not all inclination changes are equally expensive, though.

Just to be clear:  RAAN changes, which is what you're interested in, cost nothing as long as you're willing to wait for a bit.  You deploy to a low orbit, which precesses faster than things would in the final 550km orbit.  As each target RAAN approaches, you boost only those sats for the target RAAN up to the target orbit.  The difference in precession speeds between the low and 550km orbits lets you do everything with no additional delta-v.

Exactly. As a bonus, this method doesn't require the Starship to do anything except dump the satellites and get out of Dodge.

[guckyfan]

There is the present shell at 53°. There is also a slightly lower proposed shell at 53.3°. That sounds like an inclination change Starship could do. They can launch into those two shells with one launch.

[TheRadicalModerate]

There is the present shell at 53°. There is also a slightly lower proposed shell at 53.3°. That sounds like an inclination change Starship could do. They can launch into those two shells with one launch.

But you'd never do that.  Even if you wanted to build out both shells simultaneously (which isn't currently the case), you'd just launch one Starship full of Starlinks to 53º, then launch a separate one to 53.8º.  No need for an inclination change.

Note that the current license is for 540x540x53º, but there's a modification request pending for 550x550x53.2º.  SpaceX will have to do that using the Starlink electric thrusters.  That same modification takes the second shell to 540x540x53.8º, rather than 53.3º.

[guckyfan]

[quote author=TheRadicalModerate link=topic=51759.msg2151027#msg2151027 date=1604762026
But you'd never do that.  Even if you wanted to build out both shells simultaneously (which isn't currently the case), you'd just launch one Starship full of Starlinks to 53º, then launch a separate one to 53.8º.  No need for an inclination change.
[/quote]

Why wouldn't they? Assuming that they fill both shells at the same time, which they may or may not. If they launch many sats they would take a long time to drift into their plane.

[TheRadicalModerate]

But you'd never do that.  Even if you wanted to build out both shells simultaneously (which isn't currently the case), you'd just launch one Starship full of Starlinks to 53º, then launch a separate one to 53.8º.  No need for an inclination change.

Why wouldn't they? Assuming that they fill both shells at the same time, which they may or may not. If they launch many sats they would take a long time to drift into their plane.

If SpaceX has to fill multiple shells concurrently, then they have a lot of launches to do.  So it makes more sense to maximize the number of birds per launch, which means that you don't want to waste delta-v on inclination changes.

Note that there are two different plane changes we're talking about here:

1) An inclination change.  You can't really do this without expending delta-v.  These are unnecessary as long as all the birds go to the same inclination.  There's no reason for this not to be the case.

2) A RAAN change.  You can do this simply by relying on the natural precession of the RAAN caused by the earth not being perfectly spherical.

So, alternating launches for each inclination you have to fill, you launch as many Starlinks to that inclination as you can, as fast as you can, and release them in the low parking altitude.  Then, independent of the launch cadence, the satellites on orbit segregate into different groups destined for different RAANs.  When each group precesses to its target RAAN, it boosts from the low altitude to the target deployment altitude.  At the target altitude, it will precess slower, so the birds left in the parking orbit will pull ahead of the ones in the deployment altitude, allowing access to all necessary RAANs.

That's it.  It's easy, cheap, and satisfies your requirement to fill multiple shells concurrently.  No inclination delta-v needed.

[Hog]

RAAN=Right Ascension of the Ascending Node

[TheRadicalModerate]

So here's a question:  How will the precession to the proper RAAN trick work with the VLEO constellation(s)?  The difference between the drop-off orbit, assuming they leave it at 260x260ish, isn't that much different from the 340x340ish that the VLEO birds will be in.  The precession trick will still work, but the relative precession between the two altitudes is going to be much slower.

My best answer to this is that the VLEO birds aren't grouped in planes.  Each bird has a separate RAAN.  That means that dropping off a whole Starship's worth of birds covers less distance in terms of RAAN precession per unit time than 60 birds/plane, with the planes' RAANs spaced farther apart would.  So they can boost up to a viable altitude in fairly short order.

[rsdavis9]

I may have missed it ...
What is the point of the VLEO birds at 340km?

[Alvian@IDN]

I may have missed it ...
What is the point of the VLEO birds at 340km?

Faster throughput

[rsdavis9]

I may have missed it ...
What is the point of the VLEO birds at 340km?

Faster throughput

So lower latency.
Who needs less than 20ms?
Stock market traders?
Govt drone operators?

[TheRadicalModerate]

I may have missed it ...
What is the point of the VLEO birds at 340km?

Faster throughput

I think they're more likely there to soak up hot spots.  The ~550km constellation(s) should form a kind of baseload capacity, covering the entire geography fairly uniformly.  My guess is the VLEO birds will be tasked with hopping from one island of high demand to the next.  They can also be tasked to pick up unexpected surges in traffic.

[Asteroza]

I may have missed it ...
What is the point of the VLEO birds at 340km?

Faster throughput

So lower latency.
Who needs less than 20ms?
Stock market traders?
Govt drone operators?

Alleged 5G VR/AR requires low latency, and commercial remote drone driving needs low latency if doing direct driving. There's something to be said for remote construction ops using drone equipment operated by couch surfers. If the overall latency is low enough at long distance, you could outsource night construction drone driving to countries in different timezones.