The BBC article SpaceX aims to launch internet from space discusses the SpaceX proposal for a dense array of over 4,000 LEO satellites for global internet coverage, and links to SpaceX non-Geostationary Satellite System Attachment A - Technical Information to Supplement Schedule S.

The initial configuration will be the first group of 1,600 active satellites (plus spares) as 32 orbital planes of 50 satellites each, all at an inclination of 53 degrees and a LEO altitude of 1,150 km. The final configuration adds to that four more sets of planes at four more inclinations and altitudes, for a grand total of 4,425 active satellites (plus spares).

If I understand correctly, this configuration is supposed to provide global coverage such that any place on earth will see at least one satellite above 40 degrees elevation at all times.

My question is - how does this constellation's configuration actually work? Do these inclinations and altitudes satisfy some expected behavior for nodal precession? Is there some ellegant orchestration here? There are five slightly different altitudes, which means each group will have a slightly different period, so I don't think the different groups are phased relative to each other, only within a group or at least within each plane of each group.

Are there five different altitudes simply for collision avoidance - which is more difficult between planes of mixed inclinations than between planes of identical inclination (but different nodes)?

Q: How does this constellation's configuration actually work? - what is the method behind the madness? (idomatically speaking)

Here are some simple plots. I'm working on better ones - hopefully an animation or a 2D map with groundtracks.

enter image description here

enter image description here

enter image description here

enter image description here

above x2: screen captures from SpaceX non-Geostationary Satellite System Attachment A - Technical Information to Supplement Schedule S.

  • 5
    $\begingroup$ It's hard to get what your question really is, and it seems very broad. Could you try to enhance it ? $\endgroup$ – Antzi Nov 18 '16 at 1:56
  • 1
    $\begingroup$ @Antzi I've modified the question a bit. It's not a question that someone can just google and find a Wikipedia article - it's more of a question that someone already familliar with satellite constellations would certainly be able to understand, (how does this work) and possibly be able to answer. Let's give it until Monday. In the mean time I'll keep plugging a way at a real animation and some ground-cover maps. $\endgroup$ – uhoh Nov 18 '16 at 13:06
  • 3
    $\begingroup$ @DavidHammen No, really not at all a duplicate. I've asked here specifically about the particular choice of planes, inclination, and nodal precession. You are muddying the water. Please look at the question again. The other question in its entirety: "Bulk information regarding anything in this world is available in a single touch through Internet. My question is whether it is possible to replace this Internet with a constellation of Cubesats in low Earth orbit for transferring data all over the world more securely than over Internet?" $\endgroup$ – uhoh Nov 19 '16 at 18:13
  • 2
    $\begingroup$ @uhoh I noticed you posted the attachment to the SpaceX report. Do you have the full report to the proposed project? Thanks. $\endgroup$ – Moko Nov 23 '16 at 1:57
  • 2
    $\begingroup$ @Innovine: *4 satellites. But that's irrelevant because a) GPS is simplex (so ground unit power doesn't matter), b) GPS is low-bandwidth, c) GPS suffers no ill effects from high latency. None of those are true at all for telecom satellites. $\endgroup$ – Nathan Tuggy Feb 3 '18 at 4:44

Nodal precession doesn't matter for a plane of satellites like this, they will rotate around in unison, so the coverage will remain the same.

Okay, so why the unusual dual inclination constellation? The inclination of a satellite band tends to let you know what latitude it will work best at. A 0 degree satellite works best at the equator, a 90 at the poles. 53 will work best for customers around 40-60 latitude, north or south.

To have complete coverage, there are two areas that can slip. First of all is the poles, which won't receive any coverage at all. A smaller number of planes could cover the poles quite effectively.

The other is a bit less obvious, but is the equator. The equator will have 32 bands of satellites over it, crossing twice per orbit. In essence then there will be coverage for 64 satellites at a time. Over the higher latitudes, the distance around the globe at that latitude is lower, and thus the density will be higher. A single constellation of equatorial satellites will cover this gap quite effectively.

Looking at SpaceX's plans, they figure the 32 plane constellation is sufficient for the equatorial regions. The 3 inclinations are 70, 74, and 81, with 5, 8, and 6 planes respectively. My guess is the initial deployment is deemed to be good enough for most of the world's population, but they will eventually want to have global coverage. To do that with the fewest satellites possible, they decided to add in the additional 19 planes. It seems this is required because the FCC dictates that coverage must include Alaska to use that frequency band, which makes things particularly difficult.

Thinking this through a bit more, it seems 9 lateral degrees is the limit of coverage. Given 53 for optimal US coverage, 81 covers the poles, there needs to be two sparse groups of satellites to cover the rest if the area. I believe the two were chosen to optimize the satellite count by a simulation of some kind.

Note also that the higher the orbit, the more satellites are in the plane. I suspect this is to make up some of the margin that would otherwise have been lost.

  • 1
    $\begingroup$ Do you think there might be any subtleties in the particular altitudes chosen, or are they likely based on the reality of having to avoid collisions with what's up there already? Actually, that should be a separate question, so I've asked: For constellations of circular LEO satellites, are there allocations of available "slots" in altitudes?. $\endgroup$ – uhoh Nov 20 '16 at 6:11
  • $\begingroup$ The altitudes are basically to maximize rf to the ground. Higher gives a larger field of view, but overall loses some rf by distance. $\endgroup$ – PearsonArtPhoto Nov 20 '16 at 10:25
  • $\begingroup$ Hmmm... they are all within a narrow range. 1,150, 1,110, 1,130, 1,275 and 1,325 km is less than +/- 9%, with high inclinations having both the highest and almost the lowest altitude. However the higher altitudes do have the n=75 plane populations so the trend it right. But my linked question is really about allocations and slots. $\endgroup$ – uhoh Nov 20 '16 at 10:42
  • 2
    $\begingroup$ Geostationary is a poor choice for several reasons. First of all, the distance requires a lot more RF. Secondly, it's actually far enough that speed of light delays get to be annoying. Round trip delay is about 0.24 seconds, which actually adds up considerably when surfing the web. $\endgroup$ – PearsonArtPhoto Nov 20 '16 at 12:20
  • 1
    $\begingroup$ @uhoh I worked for a few years with a satellite communication network. There's a reason to the madness, and I've had my ear to the ground;-) $\endgroup$ – PearsonArtPhoto Nov 20 '16 at 14:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.