Hacker News Reader

Comment by teruakohatu | Hacker News Reader

I took a screenshot of the default view on my web browser. The satellites appear to be around 100km in diameter, but they are in reality 7m wide, so they are 14,000x larger on the map than in reality. The entire constellation would fit within 0.2% of one of the white dots as they appear in the map.

Then you need to remember that the space is 3D, unlike road surfaces (ignoring bridges etc.)

View on HN

Space is not really 3d in the intuitive sense from a packing perspective, because satellites take up orbits rather than stationary points and orbits don't pack densely. It's more like an embedded 2d space, the intuition for that being that a ring, think Saturn's, is 2d and intersects any orbit.

Yes, the key is that all circular orbits at the same altitude must cross, and if the orbits cross there is a potential collision. Satellites at the same altitude can only avoid collisions by controlling the position in the orbit.

Also, mean free path in a particle cloud depends on number density, cross sectional area and speed. The cross sectional area is small, but relative speed is also very high, which makes collisions more likely than one would think just based on their number density.

Though note: SpaceX does exactly that. They passively position the positions of the satellites along their orbits so that Starlink satellites don't need to avoid other Starlink satellites in orbital operation.

Bullshit. Space is definitely 3D and has 3D packing. Starlink is at about 550km. That has an orbital period of 1.5917hrs. Satellites are 7m (as suggested above, but this sounds large). So let's space the next group by 10km. Now the orbital period is 1.5952hrs. We have to go to 580km (+30km) to get to 1.6hrs, which is a difference of 37 seconds. So you got 60km (520km-580km) where you have an error rate of under 1% (0.65%). That drift per day is easily accounted for and you're not going to have a destructive resonance event during the lifetime of the satellites.

For comparison, the ISS (~420km orbit) has a natural decay of 2km per year. It's safe to say that you can pack these satellites <1km in orbital height without realistic risk of them running into one another. The real problem is tracking everything and we often have large margins of safety. Sure, you won't pack them like you would oranges, but even several kms apart that can create a pretty dense coverage map. Gravitational orbits aren't electron orbits and don't need to be quantized (at least on the tens of km level).

The problem with your model is all orbits at a given altitude intersect. You can orbit at the equator, or orbit above and below the equator thus crossing the equator. There’s nothing special about the equator here. All polar orbits meet at the poles, and any non polar obit is forced to cross all polar orbits etc.

In practice this is not be a significant problem because you can pack a great deal of satellites at a given altitude and moving to a slightly lower or higher orbit works just fine.

> The problem with your model is all orbits at a given altitude intersect.

> In practice this is not be a significant problem

Which is really why I didn't bring it up. We didn't even talk about eccentricity. My model is simple but this is HN, so I added magnitudes of error.

I'm not considering such concerns at all; my comment holds even considering perfect orbits followed by frictionless point objects of zero mass. You can of course pack into the height-dimension trivially when using circular orbits; the issue then is how many orbits you can pack into an orbital plane. I contend it scales with radius, not radius squared.

Also, it is kind of like an extreme version of the risks of flying cars. One hubcap/collision can mean catastrophic downstream effects.

Huh. I wasn’t worried before reading this thread. Now I’m a little more worried.

1. We don’t need to worry about Kessler syndrome, as the orbits of Starlink will decay in five years (from 550km). 2. ISS is at 420km.

Has anyone else ever worked with something dangerous dangling over their heads? It’s kind of nerve wracking, even if you know that there’s very little chance of it falling and hitting you.

(Main body) Size matters a lot. That 7m estimate above is with a solar array unfurled and is much larger than the main body, which is about 3.2m x 1.6m x 0.2m[0]. These will burn up in the atmosphere, and actually are designed to do so.

Are you afraid of airplanes flying overhead? Probably not. These don't burn up in the atmosphere and when they fall they do a ton more damage than satellites that don't. A big reason you're not afraid is because the chance of you (or any person) being hit is pretty small because the object is so small compared to the potential falling area. Now we're talking about objects that are significantly smaller than airplanes and with a potential "falling area" (if somehow something doesn't burn up) is magnitudes larger and into an area that is basically empty (humans don't even occupy 15% of the surface).

You shouldn't be worried. Even if parts of these things were hitting the ground they wouldn't be that dangerous just due to the statistics of being hit. Odds get even better when we talk about typical orbits.

[0] https://lilibots.blogspot.com/2020/04/starlink-satellite-dim...

The point isn't that I should be worried. The point is that folks in ISS should be more worried than if they were, say, at 600km...

The satellites can't just fall out of the sky suddenly like something hanging above your head. That's not how orbital mechanics works. ISS would have warning of any threat long in advance, plenty of time to make an avoidance maneuver or return to Earth in their lifeboat.

Is ISS between satellites that are expected to de-orbit and the Earth, or is it not?

I'm well aware of tracking and avoidance mechanisms, but, up until now, I hadn't considered the orbital planes in use.

The satellites are expected to control their deorbit, just as they controlled their ascent, and neither controlled ascent nor controlled descent pose any threat to ISS. If a satellite failed in orbit and was unable to control its deorbit, it is still very unlikely that it would come anywhere close to colliding with ISS on the way down. And in the very unlikely event that it did, again there would be plenty of advance notice for ISS which could do a maneuver to avoid it.

There's no real chance of starlink satellites falling and hitting you.

They're designed to fall out of orbit at the end of their lifespan (or whenever something goes wrong), and fully burn up in the atmosphere before reaching the ground.

I'm talking about them hitting ISS.

The moon seems pretty stable up there. Doesn't rattle my nerves on iota that it is overhead.

The moon is slowly getting farther away. Starlink satellites are expected to relatively quickly return to earth.

The moon is gaining velocity from earth’s rotation which is a finite energy source, given a long enough timescales it would collide with earth. However, the sun is going to consume the earth before that happens.

I'll put $10 up against that bet. In a few billion years, one of us is going to look pretty stupid!

Comparing our terrestrial notion of 2 or 3d space with orbital positions doesn't solve much. Low orbit is a very different area requiring we drop traditional notions of size and speed.

You could just have left out the word bullshit, you know?

Put another way, it is like there are 10,000 fish equally spaced around the ocean. One could swim 100 lifetimes and never see another fish.

As the surface area of a theoretical 3 dimensional spherical shell at earth diameter plus 550 km is considerably larger than the oceans, more like having 10,000 tiny anchovy sized fish, swimming a hundred lifetimes and never seeing one.

I found this helpful, to illustrate how massive the orbital 'spheres' are

https://i.imgur.com/zGaLWDg.jpg

Reply on news.ycombinator.com