update 3: Guest Post: Modelling of Starlink trail brightness and comparison to observations

update 2: Teslarati's SpaceX’s Starlink “VisorSat” launch plans revealed by Elon Musk explains Open Book and Shark Fin configurations (hint hint)!

update 1: The BBC's Musk says SpaceX is 'fixing' brightness from satellites says:

SpaceX is working on a "sunshade" that will reduce reflection of satellites sent in future launches.


Currently, the spacecraft are in a parked orbit, but over the next few months, the craft will use on-board engines to move slighter further from the Earth and rotate their solar panels towards the Sun. That will make them less visible to the naked eye but could mean they cause light pollution for astronomers trying to take pictures of the farther reaches of space.

I'd like to simulate the visibility of the tens of thousands of satellites on Earth based on descriptions of the distribution of orbits and their segments illuminated by the Sun visible to locations on Earth where dawn/dusk astronomical observations are possible (e.g. see answer(s) to What is the LSST's plan to address frequent satellite trails in data?).

This will be a rough simulation but it still requires some model of the Starlink satellites themselves, since they contain both diffuse and specular surfaces. The same way that the nearly flat antennas of the Iridium satellites caused "Iridium flares", the solar panels of the starlinks could cause (much less spectacular) flares and the diffuse surfaces would also be sources of light; I'll just start with a Phong reflection model for those.

From the images below I see that the satellites look like giant L-brackets with the satellite and its solar panels both thin, flat, and oriented at right angles.

The antennas are phased arrays and so can steer their beams over some range electronically, so the "bottom part" might not necessarily have to be oriented facing exactly nadir. Each satellite will be power-hungry, using electrical power continuously for transmitting tweeted images of cats to thousands of customers and during daylit periods recharging its batteries for the next eclipse, so it will want to orient itself to capture plenty of sunlight.

Question: What attitudes will Starlink satellites have in different orbits? Will the attitude be fairly constant throughout each period, or will it do a "dance" during each orbit in order to keep its bottom facing nadir to Earth while its panel optimally facing the Sun?

There are some depictions of Starlink satellites below of unknown accuracy.

Screen shots from Starlink.com (click for full size):

artists concept of a Starlink satellite artists concept of a Starlink satellite

below left: From Teslarati.com's SpaceX wants to offer Starlink internet to consumers after just six launches, right: an earlier incarnation of SpaceX constellation from SpaceX's 4,425 satellite constellation - what's the method to the madness?

artists concept of a Starlink satellite SpaceX's earlier 4,425 satellite constellation

  • 1
    $\begingroup$ This Yahoo Finance has an "Artist's Rendition" attributed to SpaceX showing them in orbit. Might be of questionable accuracy, as the solar panels aren't adjusted to point to the sun. finance.yahoo.com/news/… $\endgroup$ – Greg Miller Jan 17 at 4:40
  • 1
    $\begingroup$ What kind of platform you plan to use for your simulation? Considering the huge amount of satellites to be simulated at the same time, performance issues might kick in (at least it is what I experienced before). $\endgroup$ – prop-a-gator Feb 7 at 10:42
  • $\begingroup$ @prop-a-gator I only use Python. I'll just put them in circular motion. I won't propagate any orbits, but if I decided to it wouldn't be that hard 1, 2. The difference is small. It's a very simple problem geometrically, the hard part is to estimate the brightness as a function of illumination and viewing angles. Thus the question about the attitude. $\endgroup$ – uhoh Feb 7 at 14:02
  • 1
    $\begingroup$ Alright, I see. In my work I am considering a short-scale, simplified dynamic scenario (without J2-perturbation), where I propagate constellation in the separate orbital propagation software and then just report the position and velocity data to Matlab, where I also deal with visibility, look angles and other related calcs. To add a humble opinion on your topic, I would assume SpaceX would like to keep them pointing strictly nadir when the whole constellation will be on-orbit. However for initial deployments they for sure might change satellites' attitude to serve specific user groups. $\endgroup$ – prop-a-gator Feb 7 at 14:17
  • 1
    $\begingroup$ @uhoh I think these reddit threads might be useful. Here are the links: reddit.com/r/Starlink/comments/g1uou5/… reddit.com/r/Starlink/comments/g5odmg/… $\endgroup$ – prop-a-gator Apr 23 at 8:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.