undefined | Better HN

0 pointstopspin2y ago0 comments

> that you are not receiving from the same direction as you are sending

It's a thrill to think about that. Starlink is really out there.

I bet this is lost on a lot of people. Not to patronize anyone, but what Tuna-Fish is pointing out is that due to the speed of light, the distance between satellites and their relatives velocities, when one satellite is beaming data to another satellite it must aim where the receiving satellite will be, as opposed to where it is now, when the light arrives. Further, the receiver must be looking at where the transmitter was back when the signal was sent, as opposed to where the transmitter is now. And because this is all bidirectional, each satellite must send and receive in different, continuously changing directions at the same time.

0 comments

londons_explore2y ago

I understand what you're saying... But lets check.

Assuming the lenses are about 2 inches across (from photos) and they are 1310nm IR lasers:

https://www.wolframalpha.com/input?i=single+slit+diffraction...

So we have a minimum beam width of 0.0014 degrees.

And the speed of light round trip distance is say 3.3 milliseconds.

So the question is, does the angle between the satellites change faster than 0.0014/0.0033 = 0.42 degrees/second?

Well the worst case is one satellite heading north at 7.4 Km/s and another heading south at 7.4 km/s. Lets assume the satellites are 550 km apart (the distance between planes at the equator), and use the small angle approximation... Comes out as 1.4 degrees per second.

So yes, these satellites do need the ability to aim transmit and receive in different directions! (although they might be able to just defocus the beam a little when angles are changing fast to trade off throughput for design complexity)

j / k navigate · click thread line to collapse