SpaceX F9 : Starlink v0.9 : May 23, 2019 - DISCUSSION

[ArbitraryConstant]

I have some questions. 

To deliver satellites into multiple planes the Merlin will have to fire for each adjustment.  Right? 

The procedure would be:  Initial burn, release first batch, next burn, release next batch, and so on until de-orbit burn.  Right?

What is the record for Merlin relights? 

My underlying question is how many planes will the initial mission populate?

More likely they'll either have one or maybe two planes per launch or drift the satellites into planes by themselves like Iridium does. This first launch might only be one plane, too, they won't be able to test 24/7 but it'll integrate into the final network.

In this case the planes are phase planes, my understanding is transfers from a parking orbit to a specific phase plane require relatively modest delta-v. The time required to sync up may be considerable so most likely the upper stage would not be used for this. I think most likely the satellites can handle it themselves.

[ThomasGadd]

I have some questions. 

To deliver satellites into multiple planes the Merlin will have to fire for each adjustment.  Right? 

The procedure would be:  Initial burn, release first batch, next burn, release next batch, and so on until de-orbit burn.  Right?

What is the record for Merlin relights? 

My underlying question is how many planes will the initial mission populate?

More likely they'll either have one or maybe two planes per launch or drift the satellites into planes by themselves like Iridium does. This first launch might only be one plane, too, they won't be able to test 24/7 but it'll integrate into the final network.

Drifting takes too long, those Iridium are GSO at 781 km. 
I think they would want the first satellites in place as quickly as possible.

[Lars-J]

To summarize, something this bus needs anyway (reaction wheels) may be able to substitute to some extent for the mechanical complexity of deployable solar tracking arrays etc. This would never normally make sense but they're using a novel design to fit such a large number of satellites without a dispenser and in that fairly unusual circumstance I think it might make sense. They're basically floating 1U servers with solar on top and antennas below.

If you expect that these satellites won't unfold, that is quite a statement to make. I see it as rather obvious that some element(s) of this design will indeed unfold - certainly the solar arrays. But I guess we will see who is right.

And these will all be deployed in one plane and drift to other planes if desired. For *LEO*, F9 upper stages do relights to release at different altitude, not for changing the plane. (not enough delta-V to do that meaningfully) The initial plane will be the same for all payloads of one LEO launch.

[Lars-J]

Do why have any real mass for this next iteration (or pre-production models) of satellites?

Because it certainly looks like this will be the heaviest payload ever launched by a F9. 15-20 tons total?

[DigitalMan]

Do why have any real mass for this next iteration (or pre-production models) of satellites?

Because it certainly looks like this will be the heaviest payload ever launched by a F9. 15-20 tons total?

I haven't seen anything except for the details in the FCC documents that certain materials will be replaced with demisable materials in future iterations.

It is hard to determine from that what the weight difference is but it is probable the materials being replaced can be linked to specific components.  I don't know that anyone has determined what material will be used for the optical components.

[tyrred]

So how long until I can kick Comcast?

[daedalus1]

So how long until I can kick Comcast?

Or Virgin broadband (UK)?

[DaveGee66]

So how long until I can kick Comcast?

Or Virgin broadband (UK)?

This brings up a good question.... Do we have a list of what nations will be serviced from day 1 and/or what nations are most/least likely to grant Starlink access to it's citizens wrt selling its services?

[edkyle99]

So how long until I can kick Comcast?

Or Virgin broadband (UK)?

I going to wait to 1) see if it really works and 2) see how much it costs, because I expect it will be pricey.  Isn't this really aimed for the smaller towns and rural areas, where existing options are already limited and costly?

 - Ed Kyle

[marsbase]

I going to wait to 1) see if it really works and 2) see how much it costs, because I expect it will be pricey.  Isn't this really aimed for the smaller towns and rural areas, where existing options are already limited and costly?
 - Ed Kyle

If Starlink can meet the proposed performance targets (1Gbps and 15ms latency) it will be of great interest everywhere. 

[Herb Schaltegger]

I going to wait to 1) see if it really works and 2) see how much it costs, because I expect it will be pricey.  Isn't this really aimed for the smaller towns and rural areas, where existing options are already limited and costly?
 - Ed Kyle

If Starlink can meet the proposed performance targets (1Gbps and 15ms latency) it will be of great interest everywhere. 

The latency in particular is very optimistic. For comparison, I have fiber-to-the-premises internet at my home and typically get 6ms latency in my tests. For 15 years prior I had cable broadband with 20-30ms latency).

[cebri]

I have some questions. 

To deliver satellites into multiple planes the Merlin will have to fire for each adjustment.  Right? 

The procedure would be:  Initial burn, release first batch, next burn, release next batch, and so on until de-orbit burn.  Right?

What is the record for Merlin relights? 

My underlying question is how many planes will the initial mission populate?

I think all will be delivered to the same plane, plane changing is very costly in terms of delta-v and this payload is already on the heavy side.

[wannamoonbase]

I going to wait to 1) see if it really works and 2) see how much it costs, because I expect it will be pricey.  Isn't this really aimed for the smaller towns and rural areas, where existing options are already limited and costly?
 - Ed Kyle

If Starlink can meet the proposed performance targets (1Gbps and 15ms latency) it will be of great interest everywhere. 

I welcome ISP competition for sure.  It could be a hit but especially in rural areas across America. 

Once it's operational, it's easy to see licensing from other countries happening very quickly.  May as well use the constellation as much as possible and not just over the US.

[CorvusCorax]

How does this "drifting" actually work ? I assume, if you have a bunch of sats in different planes with the same inclination, all sats in all planes  encounter roughly similar precession, so relatively little corrections would be needed to keep the plane and phase distances constant relative to each other. You just need to compensate atmospheric drag.

Changing plane propulsively seems way too expensive . You could do it with multiple inclination changes, or one massive 9ne at the point of intersection, but that seems horribly inefficient.

So what's drifting? Increase apoapsis into a more excentric orbit that encounters less/more precession, then recircularise?

How much more excentricity is needed and how long does that take?

[Semmel]

So what's drifting? Increase apoapsis into a more excentric orbit that encounters less/more precession, then recircularise?

How much more excentricity is needed and how long does that take?

I cant quantify it but I guess thats the plan. It is enough to very slightly change the orbit by changing the apoapsis or periapsis. Then over time, the orbital plane shifts w.r.t the others. For example in a sun-synchronous orbit, the plane shift is 360 degrees per year (hence always facing the sun at the same angle). I dont know how the relative orbital parameters translate to relative drifting speed. But I guess thats done in the time between deployment and final orbit. The sats that need more drift remain longer in their initial orbit or just dont go quite to the destination orbit. Then, when the time is right they raise their orbits to their intended target orbit. That way, you only need a small amount of extra fuel which accounts for the additional drag in the lower orbit.

[jongoff]

One web likely too far along to change their design and business plan. Bezos was waiting to see this before starting. Will the NG 7m fairing fit more sats or bigger ones?

Bezos picked up a lot of ex SpaceX people, so he probably wasn't waiting to see this before starting...

~Jon

[jongoff]

How does this "drifting" actually work ? I assume, if you have a bunch of sats in different planes with the same inclination, all sats in all planes  encounter roughly similar precession, so relatively little corrections would be needed to keep the plane and phase distances constant relative to each other. You just need to compensate atmospheric drag.

Changing plane propulsively seems way too expensive . You could do it with multiple inclination changes, or one massive 9ne at the point of intersection, but that seems horribly inefficient.

So what's drifting? Increase apoapsis into a more excentric orbit that encounters less/more precession, then recircularise?

How much more excentricity is needed and how long does that take?

Drifting works by going into a much lower (or higher) orbit where the nodal precession rates are different from the orbit you want to end up in. This works a lot better if you're going to an orbit much higher than your deployment orbit. But with a 550km deployment orbit, I think they're more likely all going to end up in a single plane.

Just from an analysis we looked at previously for another constellation that wanted to operate at a similar altitude, you'd probably be looking at 1-2yrs to actually spread out over all planes (by going lower or higher by a reasonable amount). So I think it's a lot more likely that these are all going for the same plane or maybe at most 1-3 nearby planes.

~Jon

[ThomasGadd]

How does this "drifting" actually work ? I assume, if you have a bunch of sats in different planes with the same inclination, all sats in all planes  encounter roughly similar precession, so relatively little corrections would be needed to keep the plane and phase distances constant relative to each other. You just need to compensate atmospheric drag.

Changing plane propulsively seems way too expensive . You could do it with multiple inclination changes, or one massive 9ne at the point of intersection, but that seems horribly inefficient.

So what's drifting? Increase apoapsis into a more excentric orbit that encounters less/more precession, then recircularise?

How much more excentricity is needed and how long does that take?

Drifting works by going into a much lower (or higher) orbit where the nodal precession rates are different from the orbit you want to end up in. This works a lot better if you're going to an orbit much higher than your deployment orbit. But with a 550km deployment orbit, I think they're more likely all going to end up in a single plane.

Just from an analysis we looked at previously for another constellation that wanted to operate at a similar altitude, you'd probably be looking at 1-2yrs to actually spread out over all planes (by going lower or higher by a reasonable amount). So I think it's a lot more likely that these are all going for the same plane or maybe at most 1-3 nearby planes.

~Jon

That is my argument... after the constellations are populated they can use the slower, more efficient methods to move satellites into orbit. 

The first satellites will be tested at their injection altitude and then moved up the planned orbit. 
To test, the satellites need to be spread out across the planes.
Launching 60 satellites into a single plane, while impressive, doesn't optimize testing. 

Up to 2 days ago I would have said launching 60 satellites at once is impressive. 

[oldAtlas_Eguy]

How does this "drifting" actually work ? I assume, if you have a bunch of sats in different planes with the same inclination, all sats in all planes  encounter roughly similar precession, so relatively little corrections would be needed to keep the plane and phase distances constant relative to each other. You just need to compensate atmospheric drag.

Changing plane propulsively seems way too expensive . You could do it with multiple inclination changes, or one massive 9ne at the point of intersection, but that seems horribly inefficient.

So what's drifting? Increase apoapsis into a more excentric orbit that encounters less/more precession, then recircularise?

How much more excentricity is needed and how long does that take?

Drifting works by going into a much lower (or higher) orbit where the nodal precession rates are different from the orbit you want to end up in. This works a lot better if you're going to an orbit much higher than your deployment orbit. But with a 550km deployment orbit, I think they're more likely all going to end up in a single plane.

Just from an analysis we looked at previously for another constellation that wanted to operate at a similar altitude, you'd probably be looking at 1-2yrs to actually spread out over all planes (by going lower or higher by a reasonable amount). So I think it's a lot more likely that these are all going for the same plane or maybe at most 1-3 nearby planes.

~Jon

Just a note.
Because of using electric thrusters for the orbit maneuvers, the time frame is month to months to get the sats to final orbit. Such that sats could be dispersed to any position in any plane just as easily as getting them all to a single plane. Only the inclination and initial sunlight to start the sats up and get them going hence the launch window so that sats are deployed in sunlight at the beginning of the sun side transit.

If it had be chemical thrusters optimal would be nearly same plane or close planes. This is because time frame is days to get to final orbit. A few days is insufficient to do drifts to most of the other planes.

[ugordan]

Launch weather forecast has improved to 80% GO but upper level winds have worsened:

Launch day probability of violating launch weather constraints: 20%
Primary concern(s): Cumulus Cloud Rule
 
Delay day probability of violating launch weather constraints: 20%
Primary concern(s): Cumulus Cloud Rule

Upper level winds are not an issue, it's the upper-level wind shear that is an issue, which I don't believe these forecasts capture.