Hacker News Reader

Hacker News

Comment by exabrial | Hacker News Reader

Parent Full thread

exabrial

The problem is the speed of light is too slow :( Latency just orbit and back is 50ms+

View on HN

Says they're intended at 1100 kilometers: https://en.wikipedia.org/wiki/SpaceX_satellite_constellation

On a straight shot with no switching time that's up and back in 7.3ms.

This article [0] says they expect latency to be between 25 and 35ms with the 4,425 satellite constellation.

[0] https://arstechnica.com/information-technology/2016/11/space...

There are lots of different orbits. The ISS is 250 miles away, which shouldn't cause much of a latency problem. Of course, it's a compromise between the number of satellites and the latency.

There's also issues with how fast the satellites move across the sky. At low orbits they fly by and you either need ridiculously sensitive wide angle antennas or solid-state steerable high-gain antennas that track them as they go by.

See ex: https://www.kymetacorp.com/

It's a solved problem, the question is how expensive the ground stations are.

Iridium phones have always worked with fast-moving satellites. The 9555 phone is $939 on Amazon, and the Iridium Go hotspot is $689. Ideally, an internet access terminal would be closer to $100, but if you're sharing the connection with a bunch of neighbors using mesh WiFi, $689 is doable.

Iridium GO has the bandwidth of a dialup modem.

But there's no reason why higher bandwidth equipment is going to be more (edit: I meant hugely more....) expensive: same pointing problem to solve.

BTW there's going to be higher bandwidth (edit: adding Ka band to the previous L) from Iridium once Iridium NEXT is fully up.

Higher bandwidth means more noise, which means you need either higher gain or more sensitive detection (until you hit thermal limits). They're very interrelated issues.

The overall discussion in this sub-thread is the cost of being able to point at fast-moving satellites.

For straight bandwidth, how the the price of phone modems go up from the first analog phones to 5G? Seems to have not been much of a burden to phone prices. (Edit: shaped beams in the cell towers make a huge difference.)

I'm addressing your point that it's a solved problem. The issues around making high bandwidth data interconnects are not just "straight bandwidth". I'm explaining that there are fundamental interconnections between bandwidth and SNR that make large bandwidths harder to build at sufficient SNR:

https://en.wikipedia.org/wiki/Johnson%E2%80%93Nyquist_noise

The iridium satellite receivers use low-gain antennas, so reception is limited to small bandwidths. This means they don't care where the satellites are. New technology will enable high-gain antennas that can track the motion of satellites, enabling much higher channel rates.

I think you misunderstand what I meant by "it". I added a few notes to my earlier comments; you're assuming that the new satellites don't have higher power or tighter beams? Not to mention that low-end WiFi equipment now has shaped beams, so it's not exactly rocket science to have that in an Iridium receiver. As a former radio astronomer who did some VLBI, solid state phased arrays are easy to understand, I'm mostly glad that I don't have to be the person making them affordable in a few-hundred-$ ground station!

Yeah but aren't they also ancient in internet terms.... weren't they placed up there in the mid 90s?

The trick that most of these folks have gone with is a active phase array antenna. These antennas aren't small (smallest I've seen is 1x.25 meters, but they "just work" -- you leave them outside and suddenly connectivity.

Reply on news.ycombinator.com