SpaceX Confident About Its Starlink Constellation for Satellite Internet; Others, Not So Much

Despite technical glitches, SpaceX plans to launch the first of nearly 12,000 satellites in 2019

graphic link to special report landing page

SpaceX has, of course, been ferrying quite a bit of stuff into space lately. [See “Musk vs. Bezos: The Battle of the Space Billionaires Heats Up,” in this issue.] But last February, SpaceX launched two small satellites of its own. They were for an initial test of gear intended for use in a globe-spanning broadband data network, called Starlink, made up of thousands of small satellites. SpaceX CEO Elon Musk nicknamed the two test satellites Tintin A and Tintin B, after the beloved Belgian cartoon character known for his adventures. And just as their fictional namesake often did, the satellites ran into unexpected troubles.

After launch, Tintin A and B were supposed to propel themselves from their initial orbital altitude of 511 kilometers to their final operational orbit of 1,125 km. But the satellites remained in their initial orbits; SpaceX has never been clear about why. (SpaceX declined to comment for this story.)

“The satellite problems clearly date back quite a number of months,” says Tim Farrar, president of the satellite consulting firm TMF Associates. “The propulsion system is one you check out pretty quick after launch. One of the satellites wasn’t able to move at all. The other one has tried to maneuver without much success.”

Even so, Musk tweeted about the satellites’ strong, low-latency signals. He has acknowledged that the Starlink concept still has some challenges to overcome. But SpaceX nevertheless plans to launch the first wave of satellites in 2019, according to a statement by Patricia Cooper [PDF], the company’s vice president of satellite and government affairs. The company has not clarified the number of satellites or the launch schedule beyond mid-2019, though.

Outside observers aren’t as optimistic about SpaceX’s chances. There is a consensus that SpaceX’s business model, even more than the technical challenges it faces, could doom the constellation of satellites it plans to deploy. And it’s a big constellation: The U.S. Federal Communications Commission has currently approved SpaceX to launch 4,425 of these communications satellites into low Earth orbit (LEO) and 7,518 more in very low Earth orbit (VLEO), for a total of nearly 12,000 satellites.

According to filings with the FCC, the LEO satellites will broadcast in the Ku (12- to 18-gigahertz) and Ka (26.5- to 40-GHz) spectral bands, which are typical bands for communications satellites. The VLEO satellites, however, will make use of the V band, a higher frequency band ranging from 40 to 75 GHz.

The V band has been used for short, line-of-sight terrestrial applications, because the frequencies are too high to penetrate walls and because moisture in the atmosphere tends to absorb the signal over longer distances. SpaceX and other companies planning LEO and VLEO constellations, including Kepler Communications, LeoSat, and OneWeb, are betting that V band can nevertheless serve as a high-data-rate option for their near-Earth satellites.

An untested spectral band isn’t the only technical hurdle. LEO and VLEO constellations require a radical rethinking of the way satellites communicate with one another and with stations on the ground.

A traditional geostationary satellite can single-handedly provide continuous coverage for up to a third of the globe. The trade-off for that expansive coverage is high latency and some degradation as the signal spreads out on its way down to Earth.

An individual satellite in a LEO or VLEO constellation can’t be reached from a specific location most of the time, simply because Earth gets in the way—just as it does for distant ships, which appear to sink below the horizon when they travel far enough away. “With these satellites, they’re maybe over the horizon for 5 minutes,” says Zac Manchester, an assistant professor of aeronautics and astronautics at Stanford University. “You need to launch enough satellites and space them out such that you can guarantee there’s always one above the horizon.”

While that’s hard enough, a space-based network like SpaceX’s requires two more capabilities to be successful. First, the satellites must be able to communicate with one another directly. Traditional geostationary satellites, or geosats, work by receiving a signal from one location on Earth and directly beaming it somewhere else within its coverage area. But satellites orbiting close to Earth have such small coverage areas—ones that are constantly moving—that the signal received by one satellite will need to be bounced across the constellation and then back down to reach the right destination.

The other challenge is on the ground. The ground stations communicating with these satellites will have to be more flexible than those that communicate with geosats. “With geosats, you point the antenna forever in the direction of the satellite,” says Manchester. Ground stations communicating with constellation satellites will have to track smaller satellites moving across the horizon quickly, and the stations will also have to seamlessly begin communicating with new satellites as they move into their field of view.

Supposing that SpaceX can handle these technical challenges, the question that Farrar, Manchester, and others still have is: How will Starlink make money?

The new satellite broadband companies depict themselves as squaring off against one another in a race to build their constellations. But at the end of the day, SpaceX isn’t competing against OneWeb or Kepler. “They’re competing with Verizon,” says Manchester. “And that’s really tough.”