Spacing Out: OneWeb Rescue, Starlink Base Stations, And Rocket Tests

Another couple of weeks, and a fresh crop of space news to run through as a quick briefing of the latest in the skies above us.

OneWeb's most recent launch, from Baikonur on the 21st of March 2020.
OneWeb’s most recent launch, from Baikonur on the 21st of March 2020. (OneWeb)

The global positioning orbits are getting pretty crowded, with GPS, Russia’s GLONASS, the EU’s Galileo, Japan’s QZSS, and now with the launch of the final satellite in their constellation, China’s BeiDou. As if five were not enough the chance that they might be joined by a sixth constellation from the United Kingdom resurfaced this week, as the UK government is expressing interest in supporting a rescue package for the troubled satellite broadband provider OneWeb. The idea of an independent GPS competitor from a post-Brexit UK has been bouncing around for a couple of years now, and on the face of it until this opportune chance to purchase an “oven ready” satellite constellation might deliver a route to incorporating a positioning payload into their design. The Guardian has its doubts, lining up a bevvy of scientists to point out the rather obvious fact that a low-earth-orbit satellite broadband platform is a very different prospect to a much-higher-orbiting global positioning platform. Despite the country possessing the expertise through its work on Galileo then it remains to be seen whether a OneWeb purchase would be a stroke of genius or a white elephant. Readers with long memories will know that British government investment in space has had its upsets before.

Happily for Brits, not all space endeavours from their islands end in ignominious retreat. Skyrora have scored another milestone, launching the first ever rocket skywards from the Shetland Islands. The Skylark Nano is a relatively tiny craft at only 2m high, and gathered research data during its flight to an altitude of 6km. We’ve followed their work before, including their testing in May of a Skylark L rocket on the Scottish mainland with a view to achieving launch capability in 2023.

A Starlink phased array end user antenna, spotted in Winsconsin. (darkpenguin22)
A Starlink phased array end user antenna, spotted in Winsconsin. (darkpenguin22)

SpaceX’s Starlink is never far away from the news, with a fresh set of launches delayed for extra pre-launch tests, and the prospect of signing up to be considered for the space broadband firm’s beta test. Of more interest for Hackaday readers though are a few shots of prototype Starlink ground stations and user terminals that have made it online, on the roof of a Tesla Gigafactory and at a SpaceX facility in Wisconsin. What can be seen are roughly 1.5m radomes for the ground stations and much smaller dinner-plate-sized enclosed arrays for the user terminals. The latter are particularly fascinating as they conceal computer-controlled phased arrays for tracking the constellation as it passes overhead. This is a technology more at home in billion-dollar military radars than consumer devices, so getting it to work on a budget that can put it on a roof anywhere in the world must be a challenge for the Starlink engineers. We can’t wait to see the inevitable eventual teardown when it comes.

Elsewhere, the Virgin Galactic SpaceShip Two completed its second glide test over its Mojave Spaceport home since being grounded in 2019 for extensive refitting, and is now said to be ready for powered tests leading to eventual commercial service giving the extremely well-heeled the chance to float in the zero gravity of suborbital spaceflight. And finally, comes the news that NASA are naming their Washington DC headquarters building for Mary W. Jackson, their first African American female engineer, whose story some of you may be familiar with from the book and film Hidden Figures. The previously unnamed building sits on a section of street named Hidden Figures Way.

23 thoughts on “Spacing Out: OneWeb Rescue, Starlink Base Stations, And Rocket Tests

  1. A shallow analysis of the LEO sat mesh gives you the “hurr durr it’s not geostationary” knee jerk, but those sats need to know where they are. How are they achieving this? I would imagine they specifically need to know where they are in relation to their major uplink sites, so they are triangulating from those fixed points on the surface of the earth to know whey they are at any instant. Therefore fast and adaptive enough calculations can triangulate from them where one is at this instant on the planet. How fast and accurate can it be? Yes that will take work, but work with 2020 tech, not 1980s tech which GPS is based on so there is a decent chance of doing it well with reasonable cost.

      1. Yes, but you are limited by the dish design on how wide a field of view you can sweep the beam over. It all gets a bit optical, higher magnification, lower field of view etc. You also need a bigger dish, than a mechanically tracked one, since it’s the equivalent of tiling a lot of dishes pointing in slightly different directions.

    1. GPS birds are not geostationary. Not even geosynchronous. Not LEO, but quite a bit lower than geo.
      “calculations can triangulate from them where one is at this instant on the planet.” is exactly how it works too.

      A LEO constellation would be great for positioning, but it would take many more satellites to get continuous global coverage. Gosh, if only there was a large constellation of satellites in LEO available for this…

      It all depends on whether their clocks are good enough to generate the precision timing needed. And I’ll bet they are good enough for the usual civilian uses, but the market probably is not worth the effort of making dedicated receivers for inferior position data.

      1. Seeing as how difficult space-qualified atomic clocks with appropriate performance and long mission life-times have turned out to be in Galileo, you can safely assume the clocks on the currently flying OneWeb satellites are by far not useable for GNSS at the accuracies needed for eg turn-by-turn navigation. So procuring the currently flying constellation gives no advantage for a British GNSS.

        You may be able to use the platform of the OneWeb sats for a newly developed GNSS payload. The UK does have the necessary heritage with SSTL, but it’s not a given the OneWeb platform can fit a GNSS payload without extensive modifications. Also, the actual manufacturing of OneWeb is not at all in Britain but in the US. So it’s not really clear what advantage you’d actually gain by basing a British GNSS on OneWeb.

        Sounds suspiciously like a publicity stunt to me :/

          1. Yes, chip-scale atomic clocks are really nice, but you do realise that their Allen deviation is two orders of magnitude (or more) away from eg the maser clocks on Galileo? Let alone long-term stability and ageing.
            These thing are really useful USOs for all sorts of telecoms applications in LEO and beyond, but they are not sufficient for the next gen GNSS products. Think level 4 or 5 autonomous driving.

            And you do realise GNSS on Earth in urban environments riddled with multipath issues at sub-meter accuracy (in practice, in theory down to cm-level) is a completely different problem than lunar GNSS with 10m accuracy (claimed, compare to the cm-scale of modern GNSS)? Not that the latter is easy! It’s just … very different.

            Of course you can do _something_ based off of a telecoms platform with COTS hardware, the question is whether it’s more than a stunt.
            I look forward to be proven wrong, but so far I’m not convinced.
            (I’ve been proven wrong by SpaceX several times and loved it :)

  2. I know the SpaceX user terminals use a phased array antenna, but it also appears that they may use some automated mechanical positioning. Does anyone know whether they just orient themselves to a certain azimuth and elevation and then stay there, or if those parameters are being continuously adjusted?

      1. Luke,

        A LEO constellation with only three orbital planes would be of marginal utility: impossible to maintain a connection as you’d only see satellites for an hour or so (depending on latitude) about six times per day.

        No, Starlink plans 24 orbital planes in LEO, so some are always in view.
        Even now, with just a 500-odd satellites up, they occupy about a dozen orbital planes already.

        Very cool visualization of the planes getting populated at

    1. Don’t know, but in general it’s easier to do an array that more or less covers a plane, so you’d want to align your plane with the path of what you want to talk/listen to. Then sweep the plane, as opposed to having to figure out 3D phasing.

      1. The phased array panels used for Ku and Ka band geostationary satellites are based on the technology used in TV screens where the RF feed delay for each of the radiating elements of the panel is selected in the same way the images are on a TV screen. The big phased array panels (i.e. 1 meter which is not really big) are very expensive ( $10,000 range) and very inefficient (requiring BPSK 1/2 where a conventional metal parabolic dish could use 16APSK 7/8).
        It appears that they have launched a fleet of birds without a reasonably priced subscriber terminal. Chinese dishes for direct TV or viasat cost about $3.00 in bulk.

        Oh by the way. Every time any satellite innovator approaches the FCC and ITU for a license they provide a very flowery exhibit carefully detailing how this will be used to provide services to grossly underserved communities. Which gives us a satellite to provide service to the “other 3 billion” people who aren’t served by a satellite footprint.
        Honestly, excepting the poles, every single populated area and all maritime shipping routes are covered by generally 3 or more footprints. And have been for a long time.
        All public filings aside, every time you ever see a satellite launch you are seeing military birds.
        If you had 100 times the annual revenue from every farmhouse in America you couldn’t fuel the first flight.
        The primary reason to launch these things is to provide high-speed battlefield comms.
        Keep in mind a predator typically uses (X) Gbps and may burn through $ tens of thousands per hour in on demand bandwidth.
        Which is why they will probably go ahead and deploy with a multi thousand subscriber terminal. Drones use stabilized dishes for their terminals and money is no object.

        1. Hmmm, I might have a design in mind they could do for $3 a pop and $2 of that is my cut. Not dropping a single clue though because it will be a “Anyone could have thought of that…” but they haven’t.

          1. Sure, and one of three things will happen:
            1. Your idea is bad, so nothing will come of it
            2. Your idea is good, and someone else will think of it and implement it
            3. Your idea is good, and noone else will think of it because of your unmatched genius, and it will wither away forever

            If it never gets used, your “idea” is completely useless, whether it’s a good idea or not. Either do something about it (exploit it yourself, patent it and sell it, or release it for free), or accept that it’s a bad idea and move on.

        2. I think I read that the current price (or mfg cost?) of user terminals is in the realm of $1k. Obviously it would be in their interest to reduce that, but as-is I wouldn’t consider it untenable, especially if they provide a rental or payment option. After all, most people in the US rent their cable modem from their ISP, and make payments on their $500-1000 phones rolled into their cellphone bill. That is, if SpaceX can afford to swallow that up-front cost. I suspect that they’d be able to secure some cash from somewhere if they had to.

          I don’t think that SpaceX ever claimed that their primary target was rural households, but of course they’re portraying their service as a viable option for that, presumably to grab some of those billions of dollars the US gov put on the table for that purpose. There are tons of uses for their infrastructure, including comms between military assets as you mentioned. SpaceX wants Starlink to be their big moneymaker, and if it works as hoped, it probably will be.

          I can’t really comment on the technical and performance limitations of phased array antennas you mentioned, but wouldn’t poor efficiency really only be of a concern for the satellites themselves, and be at least somewhat mitigated by their relatively low altitude? User terminals tend to have the benefit of greater power budgets, so as long as we’re not talking multi-kW, most customers won’t be significantly impacted by it on their electricity bill.

          1. Not the electric bill cost. The cost of BUCs goes exponential with power. In order to provide consumer satellite service you will have to be in the same price point range as wildblue and viasat which dictates very low terminal costs.
            I’m fully convinced from what I’ve seen that all of the leos who are presented as commercial are fronts.

          2. You make it sound as if the user terminal is the only expense, and that Starlink will have identical performance to current sat comm services, which we know at the very least, will in fact have much lower latency (while bandwidth will be dictated in part by the number of available satellites). Currently, consumer internet via geostationary sats is expensive relative to most other forms of broadband, while having poorer performance. I think many people would pay for a relatively expensive terminal if it provided greatly improved service over anything else available to them.

            What makes you think they are fronts? And for whom? The military? If so, I have no doubts that they could be a big customer, but claiming anything beyond that sounds a bit paranoid. AFAIK, they’ve made no financial investment in the network. Not only that, but there’s no logical reason why they’d even be secretive about it. There are PLENTY of uses for such a service outside of the military. It’ll probably make any future ground-based LoS microwave backhauls obsolete. Plenty of money to be made off these LEO sat networks.

          3. Not paranoid, just realistic based on knowledge of existing networks. If anybody thinks the leos will provide low latency gigabit internet to consumers for $35 bucks a month, then I have a condo in NYC I’d like to sell them. We have a nearly identical example with O3b low latency meo which was initially marketed to provide cellular backhaul for small villages in Africa, farmers, the other 3 billion people who didnt have satellite, etc. Primary actual use is multi hundred thousand dollars per month contracts for luxury cruise ships, extremely remote oil and gas plants and mines.
            Without exception, satcoms is far and away the single most expensive and least reliable method of providing high speed data communications in the world. Very vitally necessary where no other option is available, but…. Maybe the new leos will change this, but history generally points otherwise.
            Leos replacing end to end secure carrier grade 5 nines point to point links which don’t have monthly recurring bandwidth or usage charges with a satellite link subject to rain fades, variable bandwidth, shared, possibly unsecure service and with MRC usage fees? Maybe some places as a last resort, but unlikely otherwise.

            “no logical reason why they’d even be secretive”

            Oh by the way, information on contracts, footprints, throughput, specific satellites, etc. for US military usage is not only locked down, but very locked down in a way that most can’t imagine.

  3. A LEO satellite is two orders of magnitude closer to earth than a typical GPS satellite, so if everything else is equal*, you have 20 dB higher signal levels at ground level.

    You can either throw 2,00km (LEO) and 20,200km (MEO) into the standard Free Space Path Loss equation, or you could do the basic dB calculation in your head 10dB would be 10x, so 20dB is 100x.

    *Apart from the ridiculous number of satellites required, for the same coverage you need a hell of lot more at a lower orbits (OneWeb planned to initially launch 648 satellites, and then add an additional 1972 satellites for a total of 2620 satellites). Contrast that with the US Global Positioning System navigation satellites which has a target of 33, but currently there are only 31 of which are operational, with 9 in reserve at an orbital altitude of 20,180 km.

  4. can’t wait for this to go public. I hope it’s cheap,. I’m tired of the fake 4G in all the country, and they actually give you 3,5G (which they claim it’s a minimum of 5,6 Mbps) that actually works at 400 Kbps and the service in my zone goes offline every now and then… Virgin needs to wake up and make their net faster… Or lower the prices to the reality of their speeds and data limits.

    My country sucks xD

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.