Written by Angus Keenan
Ahh the internet, you’ve probably heard of it… Bringer of fact and falsity alike. Purveyor of everything from cat video’s to conspiracy theories. From TikTok to a mail order glock the internet can provide near enough everything a person may ever want or need at only the click of a button. In only 30 years it has gone from a fairly niche tool, primarily reserved for businessmen and enthusiasts, to a cornerstone of our society. In fact I’d be willing to wager that if you’re reading this post, the last time you met someone who had never accessed the internet was over 2 decades ago. Which might make it seem like the internet is now a mature technology, as ubiquitous as the internal combustion engine or penicillin. But that’s just not the case, the internet still has a long way to go both in terms of usage and technological development, as of July 2022, only 61% of the world had access to the internet. That’s a large number sure but ubiquitous, definitely not.
So what’s the billionaire owner of a space launch company to do when he sees these kinds of technological shortcomings? Well capitalise on them of course and in doing so take the next step in ensuring that no human is ever deprived of the latest megaprojects post again. In today’s megaprojects we will talk about the next step in human communication technology, a step that could very well break the final barrier to connecting every human in the world or perhaps cause a catastrophic chain of events that ultimately results in humanity being confined to this planet forevermore.
That’s enough beating around the bush though, what exactly is this so-called next step? Well the Starlink project is Elon Musk’s way of breaking into a new industry: He uses his vast research budgets to create a product that exceeds all competitors, effectively rendering every other member of that industry obsolete. In this case the advancement is to create the largest and most refined satellite communication network as yet seen by mankind. Currently the primary use case is to provide consistent and reliable internet access to every inhabited part of the world.
So where did all of this come from? Well, the idea of large satellite constellations can be dated back to… Yep, you guessed it, the cold war. Back then the fear of ICBM attacks were great enough to warrant the US government investigating the possibility of a swarm or constellation of satellites that could detect and potentially neutralise nuclear ICBMs that had been launched during the eventuality of an attack. This was the Strategic Defence program, which we have covered in a Sideprojects post if you’d like to check that one out, AFTER, this post.
Anyway, the project was not to be, but the idea of a constellation of satellites stuck around and as communications technology became more relevant and more advanced, companies began to see the profitable side of making world wide wireless communication a thing. As of right now, the system you are using to view this post is most likely, not wireless. Behind every megabyte of data you consume there is several thousand miles of underwater copper wiring carrying that data. In more recent times, this web of wiring has been replaced by fibre-optic cables but the concept is in essence the same. In very simple terms the internet works like this:
You tap some buttons on a device of your choosing, let’s call this device the client. After tapping those buttons, the client generates a packet of information, let’s say you are looking for a youtube video. This packet contains information about you, such as an identifier for your device and a request for the information that you are looking for. Once you have made the request the packet is sent to your internet service provider, which would be your wifi router. This in turn, relays the information from server to server until it reaches the youtube server. This server is pretty much just a storage unit for all the videos on youtube. Once the packet is received and the relevant data is found, the server, in this case youtube’s server, sends the requested data back in its own packet of information.
So let’s say you live in London and you are looking to listen to some music. You type in the URL of your chosen Rick Astley song and press enter. That signal will be sent from your device to either your Internet Service Provider or ISP or to the nearest cell tower, if you are using mobile data. This then starts making its way along the old telephone wires in the direction of the coast. This will likely make it to the Highbridge landing station, on the British West coast, where it will jump onto the TGN-Atlantic cable. This carries the packet across the Atlantic onto the landing station in Wall Township, New Jersey, United States. From there it is transferred to one of Google’s many storage server locations where the information is found and then sent back to your computer along the same route. And hey presto, I’m never gonna give you up.
The point here is to make it clear that even though the process of using the internet may feel wireless from where you are sitting, there are in fact a great many wires required to send and receive the information.
This kind of system requires the very expensive and complicated process of laying cables, building landing stations, building overland wires, then the cell towers and everything else required. Fortunately most of that expensive work has already been done, every continent except Antarctica has been connected into this web of wires.
Unfortunately, unless you live in a rich country or are rich in a poor country you won’t actually reap the benefits of this infrastructure. According to the World Bank, the cost of connecting Africa to the internet, to the same extent that Europe and America are connected, would cost in the realm of $100 billion. As of 2018 24% of Africa have access to the internet, that is compared to the 91% in Europe as of 2022.
And Africa is just one of the many poorer regions yet to be fully connected to the internet, so why not skip the expensive and arduous process of physically connecting the internet and instead do it from space. The whole idea of communicating through space first appeared in a short story called the Brick Moon by Edward Everette Hale published in 1869. In it he describes a ball, 60 metres in diameter and constructed of bricks. It would assist mariners in navigation and could relay morse code back to earth. Just shy of a century after Hale’s short story was published, NASA launched Echo 1 into orbit. This tiny aluminium ball was a far cry from what Hale had envisioned but it was able to relay a signal from one location to another.
However, this was little more than the reflection of a signal and there was still a long way to go before anyone would be streaming Rick Astley. Jumping ahead to the 1990’s communication technology was beginning to meet that threshold. In the 1990’s several companies, including Celestri, Teledesic, Iridium and Globalstar were established in an attempt to be the first provider of a global satellite communication network. Of these, only Iridium and Globalstar were able to get a network up and running by 2000. A considerable achievement at the time, however these systems were patchy, the service was very expensive and required a specific kind of handset with an antenna powerful enough to connect with the satellites many hundreds of kilometres away.
In part the technology just wasn’t good enough to handle any significant volume of data which meant that the services they could provide were restricted to phone calls and very basic internet access. Then there was the fact that, during the 90’s, launching a satellite into orbit was only affordable to companies that had the budget of a small nation. This restricted the number of satellites that could be launched, compounding the issue of coverage. Ultimately this first attempt was a flop, with both of the companies folding by 2003 and then returning in a reduced capacity after being bought out.
Which brings us to 2014, when SpaceX filed an application with Norway’s telecom regulator to become a member of the international communications union. Elon Musk didn’t confirm his intentions to begin working on a new satellite communications network until January of 2015, at the unveiling of his satellite development facility in Redmond Washington.
Then in 2016 SpaceX filed an application to licence the name Starlink with the FCC, here we see the first use of the name Starlink in an official capacity. In this application, they established themselves as a (prepare yourself) ‘non-geostationary orbit satellite system in the fixed-satellite service using Ku minus and Ku positive frequency bands’. Which was essentially a way of reserving a specific frequency for their satellites to operate on… I think.
So far, the team at spaceX hadn’t revealed too much about exactly what their plan was with their satellite communications array. It was around this time that the cult of Elon Musk was beginning to form but he was still mostly known in the circles of space and electric car enthusiasts. As such, there were far fewer people scrutinising his companies filings and patents. So when SpaceX informed the FCC of their plans, in 2017, still very few people took notice. The ones that did, saw that he was planning on creating the largest satellite array on… Well not on Earth but you get the jist of what I’m saying, it was a lot of satellites. Just shy of 12,000 satellites were to be launched in the coming years.
To put that in perspective, there are about 2,900 total, publically known satellites in orbit around the earth as of 2022, not including the ones associated with SpaceX. To give you a little more perspective on that, the entire world’s GPS system operates on only 31 satellites. And to give you one last bit of perspective, when Iridium established their network in the 90’s it was using a total of 75 satellites. 75 to Starlink’s 12,000, which was only the jumping off point, SpaceX have already been filling plans to place a total 40,000 satellites into orbit. So perhaps it’s a good idea to know exactly what it is that Elon Musk is planning on putting up there.
Development of the Starlink satellites seems to have begun as early as 2004. In June of that year, shortly after Elon Musk established SpaceX, the company bought a stake in another company called Surrey Satellite Technology, itself a subsidiary of Surrey University. This company specialised in the development and manufacturing of small form satellites. This stake in the company was later sold off to Airbus in 2008 and the Starlink idea was put on the back burner.
In 2012, a tech entrepreneur called Greg Wyler, established a company called WorldVu which sought to fill much the same niche as Starlink. Shortly after establishing the company, Elon Musk began involving himself but by 2014 the pair split after a disagreement in the direction of the company. As of right now, WorldVu is still in operation and establishing their own satellite communication network. Their first launch was in February 2019 though shortly after, in March of 2020 the company filed for bankruptcy and restructured.
The vision of this company is to launch a far more modest 650-700 satellites into orbit, after restructuring, the company re-emerged under the name OneWeb and have successfully launched some 288 satellites into orbit as of 2022. Which makes them the company with the 2nd most satellites in the world, second only to SpaceX of course… Well not in the world, above the world.
However following their separation in 2014, Elon Musk embarked on his own, solo venture. Using the enormous profits of SpaceX to fund the development, manufacturing and launching of the Starlink system. While SpaceX was filing their permits for their own satellite communications network, the team at their newly purchased research and development facility, in Redmond Washington, were getting to work developing the framework for their satellite array.
Their first decision was to make something in Low Earth Orbit or LEO. The concept of LEO is pretty self-explanatory, it is a satellite that orbits fairly closely to the earth, anywhere between 500 and 2500 km from the surface of the earth, giving it an orbital period of around 90 minutes give or take.
There are a variety of bands of orbital heights that we could talk about but the only ones you need to know about here are LEO and Geostationary orbit. Geostationary orbit, is again, fairly self-explanatory, it’s a satellite that orbits at a much greater distance from the surface, around 35,000 km away. Which, as I’m sure you noticed, is a much greater distance than LEO, so far in fact, that its orbital period is exactly 24 hours. Meaning that the time it takes to orbit the Earth once is the exact same amount of time it takes for the earth to do one full rotation. So if you were to look up into the sky and observe a satellite in Geostationary orbit, it would appear not to move, hence the name geostationary.
The good thing about Geostationary orbit is that, you can create a very stable and consistent network, each satellite will cover an allocated area without the concern of unforeseen breakages in coverage. Then there is the fact that a single geostationary satellite has a view of about 43.4% of the earth’s surface, meaning you only require about 3 or 4 satellites to get a near complete coverage of the Earth, with some outages at the poles. But that’s where the advantages end, at such distances latency for a system like that would be at around 280 ms and that’s with highly favourable conditions, in reality it would probably sit around 480ms and likely more. Plus the cost of reaching that kind of altitude is significantly larger than reaching LEO.
Which makes it good for systems that require stability and aren’t bothered by high latency, such as GPS and TV broadcasting. For communications and broadband satellites, LEO works far better. The reduced cost of launching means that you can use the space otherwise spent on fuel, as space for additional payload, which will make up for the increased number of satellites. For these reasons and many others the decision to use an LEO system was a foregone conclusion.
After this the next step was developing a satellite that could feasibly stand up to the enormous broadband demands of the modern human. This meant low latency and high volume of information. Not unachievable, especially when the sheer number of satellites would provide a lot of slack and spread the load across the system. But even this, in itself, presented quite a few challenges. With so many planned to go up, they needed to fit at least 40 satellites in a single launch payload, in order for the process to be cost effective. Again this was not unachievable, and achieve it they did: By 2015 the team in Redmond had created two prototype satellites called microsat-1a and microsat-1b respectively.
Ever the ambitious businessman, in 2015 Musk announced plans to have these 2 test pieces in orbit by 2016. However, it was decided that these should remain within the atmosphere for ground testing. Two further test satellites were built called microsat-2a and 2b but it wasn’t until 2018 that they would finally make it into space, in which time they had been renamed to Tintin A and Tintin B. They were launched on the 22nd of February 2018 in the payload of a Falcon 9. Their permitted altitude was set at 512km as per their FCC filings with an orbit that would give them 9-12 months before the test pieces would re-enter the atmosphere and burn up.
With testing proving very promising Starlink pushed ahead. Between 2015 and 2018 the development team had cleared the final hurdles of production, the main one being the issue of sustainability. You see, satellites are rather complex bits of engineering and it can often take months and even years to build a single satellite. What the team in Redmond had to do was take that precision and complexity and mass produce it. Furthermore many people began sharing concerns that the satellites might add to the space debris issue and so they had to be designed using only components that would burn up in the atmosphere, oh and they also had to be cheaper to produce than any other type of satellite and the way they achieved this is the most ingenious part of the whole project, which I will explain in a moment.
For now, suffice to say that it’s all in a day’s work for an engineer and the production line was up and running by 2018, now all they had to do was put them in space. Fortunately someone had already figured that bit out. Not everybody knows this but SpaceX doesn’t just make satellites, they also have a side hustle building rockets, little bit of inside knowledge for you there. By May of 2019 60 test satellites had been released into orbit, 57 of which reached a fully operational status. In the following 4 months these were used for testing, primarily making sure that they were able to communicate with each other and ground stations. By September these initial 60 had been deorbited and the full scale roll out of the system began.
Launches continued at a rate of 1 launch a month, with each launch adding 60 new satellites to the constellation. Invariably some of the satellites in the launch would be a dud and were subsequently deorbited. If you have seen a picture of the orbital patterns of the Starlink satellites, you’ll know that it kind of looks like the planet is wearing an excellent set of intergalactic fishnets. That net pattern is designed to solve a great many requirements. First it provides the easiest conditions for each individual satellite to communicate with one another. Second it means that there will be no breakages in the coverage of the satellite. And third, it significantly reduces the chances of any two satellites colliding.
SpaceX have, for obvious reasons, been fairly unyielding with information on the precise makeup of each satellite, but there is still a fair bit that we do know about their design. Each satellite weighs about 260kg and has a kind of flat pack design. There are essentially two primary parts to the satellite, the main body, which carries the communication equipment, and the solar array. The main body houses a star tracer navigation system, which allows the satellite to orientate itself for contact with other satellites. Each satellite is also equipped with a krypton fueled hall effect thruster for orbital adjustments that need to be made through its lifespan and then for deorbiting at the end of its service life.
Finally there is the communication equipment, in total there are 4 individual phased array antennas located on the body of the craft as well as several high powered laser emitters. These are the parts that allow the satellite system to theoretically beat the current wire based system. There are a great many resources online that explain this system in far more depth than we have time for but a simple explanation would be that the satellites operate by receiving a signal from one of the many ground stations and then passing said signal from satellite to satellite, using directional lasers and receivers.
So if you were to compare this to the current system of internet operation, the ground station would be the server. When the ground station beams the information up to the satellite, one of the phased array antennas will receive the packet, which will contain the intended destination, that being your computer. From there it will path out a route to you based upon the nearest satellites, this would be the equivalent of the landing stations on either side of the atlantic. Using the star tracker navigation system the satellite will be aware of its location relative to the nearest satellites around it. Using its directional lasers it will beam the data from satellite to satellite until it reaches one within range of your receiver dish at which point it will be sent down to you.
Each satellite has an 81 degree field of view, so at the satellite’s altitude of 450 kilometres this will give it a 500 square km area that it will be able to receive and send information to. When a batch of satellites is sent up to orbit, they are done so with a 53 degree orbital inclination, which just refers to the angle that the satellite orbits relative to the equator. Which brings us to one of the most interesting parts of the whole project, that being how they managed to get these satellites into their orbital patterns.
The whole process takes roughly 4 months to complete and to start with they aren’t delivered directly to their intended altitude of 450 km but are generally detached around 275 km. Now, you’d expect that the majority of the time is taken up by the satellites strategically disconnecting from one another as with the James Webb space telescope, but that would actually take far too much fuel, instead they are just detached and then allowed to freely float away for several weeks. They will very often bump into each other as they do this but because they all have slightly different relative velocities to one another, they will begin to drift apart into a kind of buckshot pattern.
The controllers on the ground simply wait until the satellites are far enough apart that each satellite’s signal is distinguishable from the other, after which they are organised into a line. Once each satellite has been accounted for the inoperational satellites will be deorbited while the functioning ones use their Hall-effect thrusters to boost them into a higher orbit of around 380 km. This is called a parking orbit and it’s essentially a spot that is low enough to not interfere with other operational satellites but high enough that atmospheric drag is negligible.
This is where it starts to get a little complex but it’s well worth knowing. So as we all know, the Earth is constantly spinning, this spinning creates a centrifugal force which causes parts of the planet to bulge out near the equator and for physics reasons this bulging causes an orbiting satellite to precess around the earth. Think of it like a coin, if the edge of the coin is the orbit of the satellite, then placing the coin on its edge and spinning it would be the precession of said satellite.
Next think about the way a gyroscope moves as its speed changes, as it slows down it will begin to precesse, the slower the gyroscope spins, the more it precesses around. And what happens when you make the gyroscope spin faster? Well it precesses less. So bringing this all back around, once our satellites are in their parking orbit they will begin to precess around the earth, the engineers at SpaceX realised that they could use this effect to place their satellites in a variety of orbits without having to spend millions of dollars on fuel.
Let’s say they want to split their 60 satellites into 3 groups of 20, with each group having a separate angle to their orbit. All they need to do is raise the first group from their parking orbit to their operation orbit of 450 km. As they rise up they are, in effect, spinning around the earth faster than before, which means they precess less than the other satellites. In time their orbital procession will offset and the second group will then go up into their operational orbit, then the third group will repeat this and Robert’s your mothers brother, you just got 3 orbital groups for the price of one launch.
As with… Well pretty much everything these days, the Starlink project has received its fair share of criticism. If we start by discarding the ‘Elon is trying to spy on you’ crowd, we can see that there are some very reasonable objections brought against the project. And they all focus on the fact that this project increases the total number of satellites orbiting the Earth by a factor of 14, with such a huge jump it’s obviously going to have a lot of potential adverse effects.
The most likely, is an effect on Astronomy, the obvious concern being that with so many satellites floating around out there, they may start to get in the way of ground based observatories. There is a real risk that the number of satellites in orbit will begin to obscure and outnumber the visible stars in the night sky. As of September 2022, the starlink satellites number just over 2,300 and have already had a detectable impact upon the readings that observatories across the world have been receiving.
This is not just a matter of them blocking the stars out though, observatories have many methods of gathering data from the night sky and only a fraction of them rely on data from the visual spectrum. The ones that do, take long exposure photographs while focusing on specific areas of the sky allowing dim stars to more easily be spotted. Even though each satellite passes through the picture in a matter of seconds, their proximity means that even a short exposure is enough to interfere with a photo, such as with the Bianco Telescope at the CTI observatory. On the 20th November 2019 a photo of the night sky showed long white streaks where the satellite’s orbital path had crossed. There has also been disturbance to Radio telescope readings but it hasn’t been as severe.
For their part SpaceX at first insisted the satellites would not create significant disturbance to astronomy and any issues could be mitigated by image stacking and pixel masking. When this proved to be incorrect and ineffective, they began experimenting with various coatings for the satellites that would reduce their albedo. Their first attempt was on the 7th of January 2020 with the ascension of an experimental satellite dubbed ‘darksat’, this reduced the albedo by a magnitude of 0.8 or 55% which sounds promising but only created a very minor improvement to the disruptions and the idea was scrapped. Their second and most recent attempt to mitigate the issue was to equip 200 satellites with sunshades but the results were only marginally better than darksat.
The other primary issue that has been brought up about this program is the significantly raised chances of a satellite collision cascade, otherwise known as Kessler Syndrome. You’ve probably heard about this issue in recent years but the concept is that, with so many more satellites orbiting Earth, the likelihood of an unforeseen collision becomes exponentially higher. Should those unlikely events come to pass, it will trigger a domino effect of collisions that ultimately results in humans being unable to exit the Earth’s atmosphere.
This is because the resulting debris from the initial collision, will spread and destroy other nearby satellites, which in turn will create more debris, which will destroy more satellites, which will create more debris, which will destroy more satellites and so on until the entire world is encapsulated in a shell of debris, travelling around 22 times the speed of sound. Not only would this lead to a total loss of all satellites in Low Earth Orbit, destroying many of the systems that people rely upon for day-to-day life, but the pieces of debris would be so numerous and small that any hope of tracking or removing them would be, as of yet, impossible.
This is a very appreciable danger, but in spite of what some headlines might have you think, we are not doomed. For one thing we already have protocols to avoid such an eventuality. And for another thing, adding 40,000 satellites isn’t going to cause a space debris problem because a space debris problem already exists. The European Space Agency estimates roughly 670,000 pieces of space debris larger than 1cm, currently orbiting the earth. Now granted, anything smaller than perhaps 15cm isn’t going to obliterate a satellite but space junk is something that we are actively tracking and avoiding. So to say that these satellites are causing a space debris problem is a little like saying Alex Jones is causing a misinformation problem, he’s definitely a big part of it but he’s not exactly causing it.
That being said, in 2019 a potential disaster was averted when one of the earliest starlink launches came within a 1 in 1000 chance of colliding with an ESA satellite. ESA considers a 1 in 10,000 chance to be grounds for immediate course correction. This came about as a result of an issue with SpaceX’s pager system, which begs the question, why is SpaceX even using a pager system? Whatever the reason this issue caused various important emails from the European Space Agency to be missed.
So what can you say about this project? Well, to start with, the results of the Starlink Program could really go in any direction, it could be catastrophic, it could be a flop, it could be a catastrophic flop, and it could also become humanity’s next step into the future. As with most things, the results are likely going to be nuanced. When the world is as large as it is, consequences of every kind can and will arise from technological change. On the one hand you will be providing the internet to everyone, everywhere and at all times, on the other you will be hindering potential discoveries that could have been made in the field of astronomy.
Advancement has always carried risks and living with those risks is the price of living at this point in history. On the whole so far, the costs of our mistakes have been lesser than the gains of our advancements and the general goal is to keep it that way. Realistically, the starlink project is only going to serve those that can afford it and if you can afford it then you probably don’t need it. And the ones that need, can’t afford it, so all the talk of connecting the world, is really just that, talk. It would be nice to think that an impoverished family in Ethiopia or Somalia would benefit from this internet connection, providing them with free education or allowing them to video chat with a doctor that was a 2 day journey away, but statistically it’s unlikely they would have access to the electricity required to hold that connection, let alone afford the $600 upfront fee and subsequent $110 monthly payments.
By the time this kind of thing becomes affordable and feasible to every single person on the planet, it’s perfectly possible that we will have already surpassed this method of sharing information. So if it isn’t really about connecting the world, what is it about? Well it’s a passion project, to provide more convenience in the most convenient time in history thus far. That’s not a detraction by any means, the desired outcome is still a positive one but at the end of the day, these are just satellites and something doesn’t have to be wholly virtuous in order to be good. Nor does something have to be virtuous in order to be featured on this channel, it just has to be interesting and you can definitely say that about this project, it’s very interesting.