Cookies   I display ads to cover the expenses. See the privacy policy for more information. You can keep or reject the ads.

Video thumbnail
This episode of Real Engineering is brought to you by Brilliant, a problem solving website
that teaches you to think like an engineer.
On the 24th of May, SpaceX launched a Falcon 9 rocket filled with 60 satellites into space.
This marked the beginning of their ambitious new project called “Starlink” which aims
to provide high quality broadband internet to the most isolated parts of the planet,
while also providing low latency connectivity to already well connected cities.
[1] SpaceX aim to make their broadband as accessible as possible, claiming that anyone
will be able to connect to their network if they buy the pizza box-sized antenna which
SpaceX is developing themselves.
This launch of 60 satellites, was just the first of many.
Spacex has 12,000 satellites [12] planned for launch over the next decade, dramatically
increasing the total amount of spacecraft around Earth’s orbit.
This will cost SpaceX billions of dollars, so they must have a good reason for doing
Let’s see how this network will work, and how it will compete with existing internet
Back in 2015, Elon announced that SpaceX had began working on a communication satellite
network, stating that there is a significant unmet demand for low-cost global broadband
Around that time, SpaceX opened a new facility in Redmond, Washington to develop and manufacture
these new communication satellites.
The initial plan was to launch two prototype satellites into orbit by 2016 and have the
initial satellite consetllation up and running by 2020.
But the company struggled to develop a receiver that could be easily installed by the user
for a low cost, this delayed the program and the initial prototype satellites weren’t
launched until 2018.
After a successful launch of the two prototypes, Tintin-A and B, which allowed SpaceX to test
and refine their satellite design, SpaceX kept pretty quiet about what was next for
the Starlink project, until November 2018 when SpaceX received the approval from the
FCC to deploy 7,500 satellites into orbit, on top of the 4,400 that were already approved.
On May 24th, the first batch of production satellites were launched into orbit and people
around the world quickly started to spot the train of satellites moving across the night
This launch is a sign of things to come, while this initial group of satellites are not fully
functional [8], they will be used to test things like the earth communications systems
and the krypton thrusters which will be used to autonomously avoid debris and de-orbit
the spacecraft once it has reached the end of its lifecycle.
Let’s look at these functionalities first.
We have explored how ion thrusters work in the past, which you can watch for more detail,
but essentially they use electric potential to fire ions out of the spacecraft to provide
Xenon is ideally used, because it has a high atomic mass allowing it to provide more kick
per atom, while being inert and having a high storage density lending itself to long term
storage on a spacecraft.
However SpaceX opted for krypton, as xenon’s rarity makes it a far more expensive propellant.
[2] This ion thruster will initially be used to raise the starlink satellites from their
release orbits at 440 km to their final orbital height of 550 km[1].
They will also be used in conjunction with on board control momentum gyroscopes located
here, and the US Governments’ space debris collision prediction system to allow the satellites
to adjust their orbits to dodge collisions, which we have also spoken about in more detail
in a previous video.
When the satellites have reached the end of their service life they can then use the same
attitude controls and thrusters to de-orbit the satellite.
Space X have included all the necessary hardware to minimise space debris risk.
In their Federal Communications Commission approval application [3], they claim that
95% of the satellite will burn up on re-entry.
With only the ion thruster internal structure and silicon carbide components, standing a
chance of survival.
Those silicon carbide components are likely to survive, as they are essential materials
for the operation of lasers and thus have an extremely high melting point of 2,750°C.
Which brings us to our communications abilities, the primary function of the satellite.
Spacex have been tight lipped on many of the details of the satellite, but thanks to that
FCC filing [3] 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 fibre 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 it’s primary money maker.
It provides the potential of lower latency information over long distance, 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 were use a typical route, let’s say through the AC-2 cable [RI-3],
which has a return journey of about 12,800 kilometres to make through our glass fibre
optic cable.
In a vacuum light travels at a speed of 299,792,458 meters per second [5].
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].
[6] 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 in University College London.
The first source of latency for Starlink will be during the up and down link process, where
we need to transfer our information to and from earth.
We know this will be done with phase array antenna, which are radio antenna that can
control the direction of their transmission without moving parts, instead they use destructive
and constructive interference [7] to control the direction of the radio wave.
Each satellite has a cone beam with a 81 degree range of view.
With an orbit of 550 kilometres each satellite can cover a circular area with a radius of
500 kilometre.
At SpaceX’s originally planned orbit this coverage had a radius of 1060 kilometres.
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 kilometres.
The time it takes data to travel up to the satellite and back down travelling at the
speed of light is around 240 milliseconds [13] 369% slower than our subsea cable.
However, since Starlink is intending to operate at a much lower altitude, the up and down
link theoretical latency could be as low as 3.6 ms.
This is why SpaceX needs so many satellites in their 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 [3],
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 miniscule to consider.
Things get tricky here with using lasers, as we need to accurately hit the receiver
on neighbouring 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.
That will look something like this.
Communication between neighbouring 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 transfer
information between these planes too.
This requires precise tracking, as the satellites travelling in neighbouring orbital planes
are travelling 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 [9] for the European
Space Agency’s Data Relay Satellite System, which is a currently operating geostationary
internet constellation designed to serve european imaging satellites, and other time critical
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 on board 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
a latency as low as 43 milliseconds.
Even if we took the shortest route possible with an optic fibre, 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 centre in Secaucus, New Jersey and the LD4 data centre
in Slough, England at just 59.95 milliseconds, 39.4% slower than our best time with Starlink.
[11] 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 realise 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 kilometre we travel the potential gains in speed increase rapidly [5].
[RI-2] But SpaceX aren’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.
Although the internet has been one of the fastest growing technologies in human history,
by the end of 2019 more than half of the world's population will still be offline (4 billion).
Users will connect to this internet using a Starlink terminal which will cost around
$200 each, this will still be far outside the purchasing power of many third world citizens,
but it’s a start and vastly cheaper then similar currently available receivers like
the Kymeta version at a price of $30,000[15].
Elon Musk says that these will be flat enough to fit onto the roof of a car and other vehicles
like ships and airplanes.
This will allow Starlink to compete with traditional internet providers.
It’s estimated that moving the US from a 4G to a 5G wireless connection will cost around
$150 billion in fiber optic cabling alone over the next 7 years, [16] SpaceX plan to
complete their entire Stralink project for as little as $10 billion.
Each Starlink satellite cost around $300,000 which is already a massive cut in cost for
communication satellites.
SpaceX are also saving on launch costs, as they are launching on their own Falcon 9 rocket,
something that no other satellite manufacturer has.
If everything goes to plan, Starlink is estimated to generate $30 billion to $50 billion in
revenue each year on the back of premium stock exchange memberships [14], demolishing their
current annual revenue of around $3 billion.
And this is a vital part of Elon Musk's long term goals.
The money generated from Starlink will mean SpaceX will have vastly more funding than
Which could go on to fund research and development of new rockets and the technology needed to
monetise lunar and martian colonies.
For now the project is simply connecting the world even more and potentially opening avenues
for education for third world countries that do not have adequate connections to the internet.
Imagine if Leonal Messi, the greatest footballer alive, was never found and brought to Barcelona
to develop his talents.
Where he received vital medical treatment for growth hormone deficiency.
We would have been deprived of his talents, and this happens regular for remarkably talented
and intelligent children in countries where simply do not have access to education.
Widely available internet will help solve this problem, and platforms like Brilliant
have taken great steps to remedy this through their high quality interactive math and science
Brilliant has helped thousands of users realize their potential in math and science, and it’s
stories like this why I am proud to promote Brilliant every month.
They have a huge number of courses that will allow you to educate yourself.
From foundational maths courses to courses taylored for you to ace the American Mathematics
Competition, with others geared for computer science and physics.
Brilliant recently introduced a new feature, called “Daily Challenges”, which will
present with you with interesting scientific and mathematical problems to test your brain
every day.
If I have inspired you and you want to educate yourself, then go to
and sign up for free.And the first 500 people that go to that link will get 20% off the
annual Premium subscription, so you can get full access to all their courses as well as
the entire daily challenges archive.
As always, thanks for watching and thank you to all my Patreon supporters.
If you would like to see more from me the links to my instagram, twitter, subreddit
and discord server are below.