Hacker News Reader

Comment by demondemidi | Hacker News Reader

How many satellites can we put up there before we get one giant game of snooker?

If we give each a cubic kilometre keep-out zone, then we can put up 500 billion between 300 and 1300 kilometre altitude.

A cubic kilometre keep out zone is pretty generous, we could probably put in trillions if needed.

How do you keep each satellite in its own cubic kilometer? How do you know that is "pretty generous"? Why not .5cukm or 2cukm?

If all those satellites were static instead of zipping around at several tens of thousands of miles per hour, that would work out...

Except sats have a very predictable path so its not actually that difficult, stationary or not.

The system of airplane management is 1000x more complex.

Satellite paths are only predictable in the short term. Predicting satellite paths beyond about a week is not possible unless you're comfortable with huge amounts of uncertainty.

Newtonian physics says that ideal two-body problems are predictable to any point in the future. But ideal two-body problems exist nowhere in the universe, especially in near-earth orbit.

Good thing we have radar.

If you actually know about this stuff, i would be very interested to hear the reasoning behind a cubic kilometre being a generous keep-out zone (and cubic kilometres even being a sensible unit for a keep-out zone). Since this is HN, my guess is you don't, but i would love to be wrong for once.

Less keep-out is needed for altitude compared to plane

Source: I watched a Scott Manley video once

>Since this is HN, my guess is you don't, but i would love to be wrong for once.

Why are you on HN if you truly believe this? To talk shit on HN?

HN is the exact opposite. I can also say with 100% certainty, tons of SpaceX/Tesla employees browse HN to take breaks.

I'm on HN because the links are interesting. If there are real SpaceX employees who know about this, let's hear from them. The huge majority of comments on this piece right now are from people who know nothing about the subject - the same as on every post.

You need a lot more than a cubic kilometer per satellite, because you need an entire orbit around the planet.

You can't have one row of satellites orbit at the equator, and another row orbit at 40 degrees north latitude. A row is already difficult, they catch up with each other quickly just due to random perturbations and elevation differences.

Instead, each wrap of your ball of yarn crosses over the others. Each row intersects most other rows.

You mean a Kessler syndrome, or event https://en.wikipedia.org/wiki/Kessler_syndrome And to answer your question : we don't know exactly yet, but I guess we'll know pretty soon

> Kessler syndrome soon

No, certainly not in LEO. Debris in LEO falls back to Earth in only a few years so even if a shooting war started and shattered every satellite in LEO into a mist of deadly orbital speed bullets, a few years after the war ends LEO would be clear again. All of the other effects of such a large-scale shooting war would be far more concerning to everybody on the ground, space debris would be the least of the concerns for any survivors.

Seriously. 30,000! They should just go ahead and call it the Kessler Network

I never understand why people are so obsessed crying about Starlink and Kessler Syndrome when any rational thought about it instantly suggest that this is practically impossible to happen.

Do you also worry Supervolcano anytime anybody digs a hole?

OK. Do you have any actual counterargument, or just your sense of disbelief?

> Do you also worry Supervolcano anytime anybody digs a hole?

No. The risk is not remotely comparable.

Put 29 988 people spread out on the globe (slightly larger since we are further out in the radius, but lets ignore that).

Do you still wonder? Probably you could put a few million up there before collision is a worry.

Space debris and blocking observatories is a much larger problem.

> Put 29 988 people spread out on the globe...

... and then have them run at 17,000 miles per hour.

... on highly predictable and known paths.

Where only a percentage of objects can change their path if needed.

If a collision does happen 2 objects become 20,000 objects in an instance.

If a 300kg satellite at a 300 mile orbit becomes 20,000 pieces, those are pretty darn small pieces. Objects in space regularly get hit by <1cm micro-meteoroids. They may become disabled but they don't shatter. And by being so small their mass/area ratio becomes really small and they deorbit within 90 minutes from that altitude.

The orbital tracks at https://en.wikipedia.org/wiki/2009_satellite_collision clearly show how a high-speed collision can raise the orbits of the resulting debris, having imparted significant acceleration to them. (For example: <https://en.wikipedia.org/wiki/2009_satellite_collision#/medi...>)

That smash "created at least 1,000 pieces of debris larger than 10 cm (4 in)".

https://en.wikipedia.org/wiki/Satellite_collision lists six unintentional high-speed collisions involving satellites that have already occurred. That's with 2-8k or so in orbit at the time.

https://en.wikipedia.org/wiki/2009_satellite_collision alone made two thousand cataloguable fragments to track on new paths we didn't get to pick.

5 of those 6 were with debris. There is >500K pieces of debris in orbit bigger than 1cm in size, many of those in unknown orbits Adding 30K satellites in known orbits where debris cannot stay in orbit and required by law to have disposal plans before launch will not meaningfully change those odds.

Iridium 33 was active and capable of avoidance maneuvers. Both orbits were known. Despite this, they collided. It is clearly possible for this to get fucked up, and with 30k satellites in a shell, adding thousands of pieces of debris to the mix is clearly a non-zero risk. The exact magnitude of that risk may be debatable, but its existence cannot be waved away.

SpaceX's existing constellation (also in such a self-cleaning orbit) made 1,300 maneuvers in six months to avoid debris from a satellite explosion a year earlier. https://techcrunch.com/2023/07/10/starlink-satellites-are-do...

The risk of a collison is non-zero. The risk of Kessler from 30K satellites in a 300 mile orbit is zero.

Sorry, but that's simply false.

https://en.wikipedia.org/wiki/ASM-135_ASAT#1985_test destroyed a satellite at 345 miles. Debris from it persisted until 2004.

The low orbit reduces the long-term impact of such a chain-reaction - we wouldn't be locked out of space for a thousand years - but it would still be a serious disaster.

Only 8 pieces from that collision remained in orbit long term. All 8 were trackable and avoidable.

Just like Kosmos 2251 was.

Something still hit it.

Again, non-zero. You can argue it's low; you can't argue it's zero.

In the real world, infinitesimal is zero.

In the real world, we've had six satellite crashes, including at least one with basically the perfect mitigatable scenario; active satellite capable of avoidance, known orbits, large easily trackable objects.

There have been several notable collisions in the last decade. I think a full blown Kessler scenario is unlikely but it will definitely be a problem for satellite operators. Let alone the impacts on terrestrial astronomy.

are orbits (LEO, MEO, GEO) path-specific, like lanes, or it's willy-nilly where only altitude is mandatory-ish?

"Snooker" would be so much simpler, unfortunately it just turns into super high speed shrapnel.

Reply on news.ycombinator.com