As we discussed earlier how silicon carbide can survive re-entry into Earth’s atmosphere, we can now answer the question of why SpaceX wants a Starlink.
SpaceX have been tight-lipped on many of the details of the satellite, but thanks to that FCC filing we know that the satellites will contain 5 1.5 kilogram silicon carbide components, which indicates that each satellite will contain 5 individual lasers. These lasers, like our fiber optic cables here on earth, will use light pulses to transmit information between satellites.
Transmitting with light in space offers one massive advantage over transmitting with light here on earth, however. The speed of light is not constant in every material, in fact, light travels 47% slower in glass than in a vacuum.
This offers Starlink one huge advantage that will likely be its primary money maker.
It provides the potential of lower latency information over long distances; in simpler terms let’s imagine this as a race between data packets. A user in London wants the new adjusted price of a stock on the NASDAQ from the New York stock exchange. If this information used a typical route, let’s say through the AC-2 cable [RI-3], which has a return journey of about 12,800 kilometers to make through our glass fiber optic cable.
In a vacuum, light travels at a speed of 299,792,458 meters per second. The speed of travel in glass depends on the refractive index and the refractive index depends on wavelength, but we will take the reduction as 1.47 times slower than the speed of light in a vacuum [203940448 m/s].
This means the data packet will take roughly 0.063 seconds to make the round trip, and thus has a latency of 0.063 seconds, or 62.7 milliseconds. With the additional steps that add to this latency like the conversion of light signals to electrical signals on either end of the optical cable, traffic queues, and the transfer to our final computer terminal, this total time comes out at about 76 milliseconds.
Figuring out the latency for Starlink is a lot more difficult, as we have no real-world measurements to go by, but we can make some educated guesses with the help of Mark Handley, a communications professor at University College London.
The first source of latency for Starlink will be during the up and downlink process, where we need to transfer our information to and from the earth. We know this will be done with a phased array antenna, which is a radio antenna that can control the direction of their transmission without moving parts, instead, they use destructive and constructive interference to control the direction of the radio wave.
Each satellite has a cone-beam with an 81-degree range of view. With an orbit of 550 kilometers, each satellite can cover a circular area with a radius of 500 kilometers. At SpaceX’s originally planned orbit this coverage had a radius of 1060 kilometers. Lowering the altitude of a satellite decreases the area it can cover, but also decreases the latency. This is particularly noticeable for typical communications satellites operating in geostationary orbit at an altitude of about 36,000 kilometers. The time it takes data to travel up to the satellite and back down traveling at the speed of light is around 240 milliseconds 369% slower than our subsea cable.
However, since Starlink is intending to operate at a much lower altitude, the up and downlink theoretical latency could be as low as 3.6 ms. This is why SpaceX needs so many satellites in its constellation in order to provide worldwide coverage. Each individual Starlink satellite has four phased array antenna located here, here, here and here.
This directional beam was an essential part of SpaceX’s FCC approval application, as thousands of satellites broadcasting undirected radio waves would cause significant amounts of interference with other communication methods. Once that data is received by one Starlink satellite, it can begin to transmit that information between satellites using lasers. Each time we hop from satellites there will be a small delay as the laser light is converted to an electric signal and back again, but it is too minuscule to consider. Things get tricky here with using lasers, as we need to accurately hit the receiver on neighboring satellites to transmit that data. Let’s look at SpaceX’s proposed constellation to see how this will work.
Space X’s first phase of 1584 satellites will occupy 24 orbital planes, with 66 satellites in each plane inclined at 53 degrees. Communication between neighboring satellites in the same orbital plane is relatively simple, as these satellites will remain in relatively stable positions in relation to each other.
This gives us a solid line of communication along a single orbital plane, but in many cases, a single orbital plane will not connect two locations, so we need to be able to transfer information between these planes too. This requires precise tracking, as the satellites traveling in neighboring orbital planes are traveling incredibly quickly and will come in and out of view. This means the Starlink satellite will need to switch to a new satellite in the network. This can take time, the best figure I could find is about a minute for the European Space Agency’s Data Relay Satellite System, which is currently operating geostationary internet constellation designed to serve European imaging satellites, and other time-critical applications.
Such as serving emergency forces in remote areas, like those fighting forest fires. Starlink may be faster, but it won’t be instantaneous, and thus it has 5 optical communication systems onboard to maintain a steady connection to 4 satellites at all times.
If we now use this system, transmitting from New York to London and back, with the shortest path possible, using the speed of light in a vacuum as our transfer speed, we can achieve latency as low as 43 milliseconds. Even if we took the shortest route possible with an optic fiber, which does not exist, this would take about 55 milliseconds, a 28% decrease in speed. The actual current return trip time for your average Joe is about 76 milliseconds, as we saw earlier. A 77% decrease in speed. This is a huge deal for the two financial markets working out of these cities, with millions of dollars being moved in fractions of a second, having a lower latency would provide a massive advantage in capitalizing on price swings.
In fact, it wouldn’t be the first time a communications company has made a massive investment to specifically serve these groups. The Hibernian Express cable is a privately owned optic cable that is currently the lowest latency connection between the NY4 data center in Secaucus, New Jersey and the LD4 data center in Slough, England at just 59.95 milliseconds, 39.4% slower than our best time with Starlink.
The previous best time was held by the AC-1 cable at 65 milliseconds. At a cost of 300 million dollars this 5 millisecond increase in speed was justified to just connect across the Atlantic.
Imagine how much these time-sensitive industries will be willing to pay for a 17-millisecond increase in speed. It becomes even more valuable when you realize this time differential increases with increased transmission distance. New York to London is a relatively short distance.
The improvements would be even more pronounced for a London to Singapore transmission, for every additional kilometer we travel the potential gains in speed increase rapidly. [RI-2] But SpaceX isn’t just planning on serving this super-fast internet to some customers, they primarily advertise this system as a way to connect every human on this planet to the internet, and they should have plenty of bandwidth left over to serve these people.
Your email address will not be published. Required fields are marked *
17 November, 2019