The day's successful Starlink launch was a great one for SpaceX – the heaviest payload ever, weighted down with 60 communication satellites that will eventually be part of a single constellation that gives the internet to the world. It's the plan, anyway – and the company pulled the curtain back a little more after its launch, revealing a few details about the birds that just sat in the air.
SpaceX and CEO Elon Musk have been extremely tightly lipped about the Starlink satellites, just dropping some tips here and there before launch. For example, we know that each satellite weighs about 500 pounds, and is a flat screen design that maximizes the amount that fits into each payload. The launch media set also describes a "Startracker" navigation system that allows the satellites to find themselves and orbital roll with precision.
However, on the new, new Starlink website, some new details have been shown along with some images that give the clearest look yet (do, not photographs, but still) of the satellites that will soon come thousands in the ski.
In the CG representation of how the satellites will work, you get a general sense of it:
Thousands of satellites will move along their paths simultaneously, each radiant internet to and from the surface of a given area. It is still not clear exactly how large an area each satellite will cover or how much redundancy is required. But the image gives you the general idea.
The signal comes from and goes to a set of four "phased array" radio antennas. This compact, flat type antenna can transmit in multiple directions and frequencies without moving as you see large radar dishes do. There are also costs, but there is a no-brainer for satellites that need to be small and just need to transfer in a general direction – down.
There is only one sunrise that unfolds upwards like a map (and looks pretty much like you would expect – thus no picture here). The benefits of having just one are mainly related to simplicity and price – having two gives you more power and redundancy if one fails. But if you are going to do a few thousand of these things and replace them every few years, it probably doesn't mean too much. Solar arrays are reliable standard parts now.
The krypton-powered ion thruster sounds like science fiction, but ion thrusters have actually been around for decades. They use a charge difference to shoot ions – charged molecules – out in a particular direction, giving force in the opposite direction. Very like a small electric pea player, which in microgravity pushes the person back with the driving force.
To do this, it needs propellant – usually xenon, which has several (quite difficult to explain) properties that make it useful for these purposes. Krypton is the next Noble gas listed in the table and is similar in some ways, but easier to get. Again, if you use thousands of ion engines, so far, only a handful has actually flown – you want to minimize costs and exotic materials.
Finally, it is the star tracking system and collision hazard. This is not very well explained by SpaceX, so we can only persuade based on what we see. The star tracker tells each satellite its attitude or orientation in space – presumably by looking at the stars and comparing it with known variables such as time of day on earth and so on. This is linked to collision failure, which uses the government's database of known room debris and can adjust courses to avoid it.
How? The image on the Starlink page shows four plates at perpendicular orientation. This suggests that they are reaction wheels that store kinetic energy and can be spun or lowered to give it the power of the boat and turn it as desired. Very smart little devices actually, and quite common in satellites. These would control the attitude and the thruster would give a slight impulse, and the suckers would be avoided. The satellite can return to normal path shortly afterwards.
We still don't know much about the Starlink system. For example, how do the ground stations look? Unlike Ubiquitilink, you cannot receive a Starlink signal directly on your phone. So you need a receiver, which Musk has said lately, is about the size of a pizza box. But small, big or extra large? Where can it be mounted and how much does it cost?
The question of interconnection is also a mystery. Say a Starlink user wants to visit a site hosted in Croatia. Signals go up to Starlink, between satellites and down to the nearest base station? Is it going down to a large interconnection point on the spine that serves the region? Is it going up and then you come down 20 miles from your house on the spot where fiber connects to local spine? It can't be much for ordinary users, but for big services – think Netflix – it can be very important.
And finally, how much does it cost? SpaceX will do this competitively with land-based broadband, which is a bit difficult to believe in taking into account the fiber growth, but it is not so hard to believe because telecoms are pulling their heels to rural areas that still use DSL. Out there, Starlink can be a godsend, while in big cities it can be superfluous.
Chances are we don't know long. The 60 satellites up there right now are just the very first wave and contain nothing more than a test bed for future services. Starlink must prove that these things work as planned, and then send out hundreds more before it can offer even the most rudimentary service. Of course, it's the plan, and can even be completed by the end of the year. Meanwhile, I have asked SpaceX for more details and will update this post if I hear back.